Re: [Ocfs2-users] issues with my ocfs2 cluster
Hi Jim, Have you tried to update the kernel as suggested by Changwei? The messages seem to indicate kernel 4.4.0, a google search shows this ubuntu version should be able to use kernel 4.10. Best Regards,Luis On Wednesday, January 10, 2018 4:12 PM, Jim Okkenwrote: hello again list, We seem to be having issues on more servers where according to the linux developers here: "the kernel is stuck in a spin lock during a disk operation." The call traces are below, I see a lot of ocfs in the call traces, but I don't know how to read them, please tell me does the issue come from ocfs?thanks --Jim 2018-01-06T17:10:02.194362+00:00 node-115 kernel: [87885.155288] Modules linked in: vhost_net vhost macvtap macvlan ip6table_raw xt_mac xt_tcpudp xt_physdev br_netfilter veth ebtable_filter ebtables openvswitch ocfs2 quota_tree ocfs2_dlmfs ocfs2_stack_o2cb ocfs2_dlm ocfs2_nodemanager ocfs2_stackglue configfs ip6table_filter ip6_tables xt_multiport xt_conntrack iptable_filter xt_comment xt_CT iptable_raw ip_tables x_tables xfs bridge 8021q garp mrp stp llc intel_rapl x86_pkg_temp_thermal intel_powerclamp coretemp crct10dif_pclmul ipmi_ssif crc32_pclmul ghash_clmulni_intel kvm_intel aesni_intel aes_x86_64 kvm lrw gf128mul glue_helper ablk_helper irqbypass cryptd hpilo 8250_fintek serio_raw ioatdma ipmi_si sb_edac edac_core ipmi_msghandler shpchp dca acpi_power_meter lpc_ich mac_hid ib_iser rdma_cm iw_cm ib_cm ib_sa ib_mad ib_core ib_addr iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi nf_conntrack_proto_gre nf_conntrack_ipv6 nf_defrag_ipv6 nf_conntrack_ipv4 nf_defrag_ipv4 nf_conntrack autofs4 raid10 raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx xor dm_round_robin ses enclosure scsi_transport_sas raid6_pq libcrc32c raid1 raid0 multipath linear uas usb_storage psmouse lpfc be2net vxlan ip6_udp_tunnel scsi_transport_fc udp_tunnel wmi fjes scsi_dh_emc scsi_dh_rdac scsi_dh_alua dm_multipath2018-01-06T17:10:02.194364+00:00 node-115 kernel: [87885.157143] CPU: 15 PID: 11936 Comm: qemu-system-x86 Not tainted 4.4.0-98-generic #121-Ubuntu2018-01-06T17:10:02.194366+00:00 node-115 kernel: [87885.157144] Hardware name: HP ProLiant BL460c Gen9, BIOS I36 02/17/20172018-01-06T17:10:02.194367+00:00 node-115 kernel: [87885.157280] task: 882036ff ti: 881f80ca task.ti: 881f80ca2018-01-06T17:10:02.194400+00:00 node-115 kernel: [87885.157281] RIP: 0010:[] [] native_queued_spin_lock_slowpath+0x15c/0x1702018-01-06T17:10:02.194414+00:00 node-115 kernel: [87885.157566] RSP: 0018:88203f143c30 EFLAGS: 02022018-01-06T17:10:02.194416+00:00 node-115 kernel: [87885.157567] RAX: 0101 RBX: 8820046c83f0 RCX: 00012018-01-06T17:10:02.194418+00:00 node-115 kernel: [87885.157705] RDX: 0101 RSI: 0001 RDI: 8820046c83ec2018-01-06T17:10:02.194440+00:00 node-115 kernel: [87885.157705] RBP: 88203f143c30 R08: 0101 R09: 811924a72018-01-06T17:10:02.194442+00:00 node-115 kernel: [87885.157706] R10: ea0040d6d680 R11: 0800 R12: 8820046c83ec2018-01-06T17:10:02.194443+00:00 node-115 kernel: [87885.157707] R13: 0800 R14: 4c63ee00 R15: 08002018-01-06T17:10:02.19+00:00 node-115 kernel: [87885.157708] FS: 7fbcbb7eec00() GS:88203f14() knlGS:2018-01-06T17:10:02.19+00:00 node-115 kernel: [87885.157709] CS: 0010 DS: ES: CR0: 800500332018-01-06T17:10:02.194445+00:00 node-115 kernel: [87885.157710] CR2: 7f54266a8000 CR3: 000fcc2f2000 CR4: 001426e02018-01-06T17:10:02.194446+00:00 node-115 kernel: [87885.157711] Stack:2018-01-06T17:10:02.194448+00:00 node-115 kernel: [87885.157712] 88203f143c40 81844421 88203f143c60 818425352018-01-06T17:10:02.194449+00:00 node-115 kernel: [87885.157714] 881e88a9ca80 8820046c84b0 88203f143c70 8184257b2018-01-06T17:10:02.194450+00:00 node-115 kernel: [87885.157716] 88203f143ca0 c074158d 881e5d3beb80 08002018-01-06T17:10:02.194450+00:00 node-115 kernel: [87885.157717] Call Trace:2018-01-06T17:10:02.194451+00:00 node-115 kernel: [87885.157718] 2018-01-06T17:10:02.194453+00:00 node-115 kernel: [87885.157725] [] _raw_spin_lock+0x21/0x302018-01-06T17:10:02.194454+00:00 node-115 kernel: [87885.157727] [] __mutex_unlock_slowpath+0x25/0x502018-01-06T17:10:02.194456+00:00 node-115 kernel: [87885.157729] [] mutex_unlock+0x1b/0x202018-01-06T17:10:02.194457+00:00 node-115 kernel: [87885.157766] [] ocfs2_dio_end_io+0x6d/0x80 [ocfs2]2018-01-06T17:10:02.194458+00:00 node-115 kernel: [87885.157770] [] dio_complete+0x11c/0x1c02018-01-06T17:10:02.194460+00:00 node-115 kernel: [87885.157771] [] dio_bio_end_aio+0x73/0x1002018-01-06T17:10:02.194461+00:00 node-115 kernel: [87885.157774] []
Re: [Ocfs2-users] Backup issues
Dirk, Also, you should have noatime enabled on this filesystem, check your mount options. Or else the rsync will end up causing access time to be updated. Regards, Luis From: Eduardo Diaz - Gmail ediaz...@gmail.com To: Dirk Bonenkamp - ProActive d...@proactive.nl Cc: ocfs2-users@oss.oracle.com Sent: Tuesday, April 3, 2012 11:26 AM Subject: Re: [Ocfs2-users] Backup issues For backup I use a local copy using tar, but I use rsync too. I don't compare but if you send any stadistics about rsync we can said more information. use rsync --stats I don't note difernt speed, but did you make a full rsync, and show --progress for see that file it is going at what speed? regards On Mon, Apr 2, 2012 at 9:04 PM, Dirk Bonenkamp - ProActive d...@proactive.nl wrote: Hi All, I'm currently testing a OCFS2 set-up, and I'm having issues with creating backups. I have a 2 node cluster, running OCFS2 on a dual primary DRBD device. The file system is 3.7Tb of which 211 Gb is used: about 1.5 million files in 95 directories. Everything works fine, except for the backups, which are taking way more time than on 'regular' file systems. I'm using rsync for my backups. When I rsync the file system above, this takes more than an hour, without any modifications to the file system. Network / disk speed is good. I can rsync a 10 Gb file from the OCFS2 filesystem to the same backup server with just under 100 Mb/s. I know there is some penalty to be expected from a clustered file system, but this is al lot. Rsyncing an ext3 file system double the size (in Mb's and files) of this file system takes about 600 seconds... Has anybody some advice on a backup strategy for me? Or some tuning tips? Thanks in advance, Dirk -- http://www.proactive.nl T 023 - 5422299 F 023 - 5422728 www.proactive.nl http://www.proactive.nl ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] partition offset/alignment on SAN devices.
Thomas and James, Usually the partition is aligned to the Cylinder boundary. In the case of a LUN, the cylinder bondary might have no sense at all? Funny, I never tought about this before. There are some tools that can overwrite a few sectors on the start of the disk (lilo, grub, DOS fdisk), and expect the disk to have a partition table, where these sectors are not used for this reason. I don't know if OCFS2 has provisions for leaving this space unused. Regards, Luis From: thomas.zimol...@bmi.bund.de thomas.zimol...@bmi.bund.de To: james.mas...@tradefair.com; ocfs2-users@oss.oracle.com Sent: Fri, July 9, 2010 11:24:38 AM Subject: Re: [Ocfs2-users] partition offset/alignment on SAN devices. mkfs.ocfs2 -L SOMELABEL /dev/dm-10 This 64k (32/128k whatever) issue is usually only a problem if you've used a fdisk to create a partition to put your data on. For hysterical reasons the first partition is created an arbitrary amount of kb into the disk. This almost never lines up with the LUN raid-stripe. Then the I/O going from app FS LUN disks overlaps block sizes - causing unnecessary I/O through the chain and sub-optimal performance. Hi folks, we had that issue too with our CX4-480 and the guys from EMC telling us not to forget the alignment. So we first wondered how to automate this, because we'd have had to align the partitions on quite a few LUNs (more than 40). Since you'd have to use the expert options of fdisk, we did'nt find a quick solution with sfdisk, though with some further investigation there will surely be one. But in the end this was not necessary anyway. We specifically requested that at EMC and they confirmed, that alignment is only a point when using partitions at all. We don’t use partitions, so no todo. IMHO: Concerning the risk of destroying whatever is on the LUN: If someone thinks the LUN is unused, just because he/she sees no partition on it, isn't that more of a lack of diligence? We have a similar situation with ASM-devices: Even if you have a partition on it, it's not mountable (as there's some ASM data inside it and no FS). So you'd always have to check more than that to be sure, that the device is unused. Maybe you can clarify that with the vendor of your shared disks. Greetings, Thomas ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Combining OCFS2 with Linux software RAID-0?
Brian, Hmm, I was not aware of this. Seems Novel uses other volume manager, called EVMS, not CLVM (?). From: http://wiki.novell.com/index.php/Linux_Data_Management Some Open Source OCFS2 Features Oracle Linux Certification matrix OCFS2 project web site OCFS2 Development Roadmap Oracle Cluster File System v2 (OCFS2) is an open source cluster management and No exclusive write lock capability yet (now every lock request returns: successful). This feature is candidate for SLE10 SP2 (Q1 2008). OCFS2 on top of a software mirror is not supported yet Can be managed by EVMS OCFS2 offers integration with heartbeat2. Heartbeat2 offers a Resource Agent 'md group take over'. (which enables fail-over of host based mirroring of SAN volumes), but OCFS2 on top of a software mirror is not supported. ... Best Regards, Luis Freitas --- On Fri, 12/11/09, Brian Kroth bpkr...@gmail.com wrote: From: Brian Kroth bpkr...@gmail.com Subject: Re: [Ocfs2-users] Combining OCFS2 with Linux software RAID-0? To: Luis Freitas lfreita...@yahoo.com Cc: Patrick J. LoPresti lopre...@gmail.com, ocfs2-users@oss.oracle.com Date: Friday, December 11, 2009, 2:09 PM Luis Freitas lfreita...@yahoo.com 2009-12-11 05:40: Patrick, Depending on what you are using, you could use the volume manager to do the striping, but you need to use CLVM. So if you can, go for Heartbeat2+CLVM+OCFS2, all integrated. Not sure but I think Heartbeat2+OCFS2 is only available on the vanilla kernels, not on the enterprise ones. Maybe Suse has support, I don't know, you will have to check. Best Regards, Luis Freitas Just to elaborate on these comments. Last time I checked CLVM required the openais/cman cluster stack, which neither heartbeat nor ocfs2 use (by default). The userspace stack option for ocfs2 in recent mainline kernels added support for the openais stack and pacemaker is required to make heartbeat work with that rather than use it's own cluster stack. Now, you can do an basic LVM linear span, concatenation, or whatever you want to call it without any cluster stack, so long as it's not striped and so long as you heed Sunil's warning about fat fingering changes to the thing while more than one host is using it. That means that if you want to add another LUN to the span you can't do it on the fly. You have to do something like this: # On all nodes: umount /ocfs2 # On all nodes but one: vgchange -an ocfs2span # Or, to be extra safe: halt -p # On the remaining node: vgextend ocfs2span /dev/newlun lvextend -l+100%FREE /dev/mapper/ocfs2span-lv tunefs.ocfs2 -S /dev/mapper/ocfs2span-lv # You might actually need the fs mounted for that last bit, I forget. # Probably a fsck somewhere in there would be wise as well. # Bring the other nodes back up. Brian --- On Wed, 12/9/09, Patrick J. LoPresti lopre...@gmail.com wrote: From: Patrick J. LoPresti lopre...@gmail.com Subject: [Ocfs2-users] Combining OCFS2 with Linux software RAID-0? To: ocfs2-users@oss.oracle.com, linux-r...@vger.kernel.org Date: Wednesday, December 9, 2009, 9:03 PM Is it possible to run an OCFS2 file system on top of Linux software RAID? Here is my situation. I have four identical disk chassis that perform hardware RAID internally. Each chassis has a pair of fiber channel ports, and I can assign the same LUN to both ports. I want to connect all of these chassis to two Linux systems. I want the two Linux systems to share a file system that is striped across all four chassis for performance. I know I can use software RAID (mdadm) to do RAID-0 striping across the four chassis on a single machine; I have tried this, it works fine, and the performance is tremendous. I also know I can use OCFS2 to create a single filesystem on a single chassis that is shared between my two Linux systems. What I want is to combine these two things. Suse's documentation ([1]http://www.novell.com/documentation/sles11/stor_admin/?page=/documentation/sles11/stor_admin/data/raidyast.html) says: IMPORTANT:Software RAID is not supported underneath clustered file systems such as OCFS2, because RAID does not support concurrent activation. If you want RAID for OCFS2, you need the RAID to be handled by the storage subsystem. Because my disk chassis already perform hardware RAID-5, I only need Linux to do the striping (RAID-0) in software. So for me, there is no issue about which node should rebuild the RAID etc. I understand that Linux md stores meta-data on the partitions and is not cluster aware, but will this create problems for OCFS2 even if it is just RAID 0? Has anybody tried something like this? Are there alternative RAID-0 solutions for Linux that would be expected to work? Thank you. - Pat ___ Ocfs2-users mailing list [2]ocfs2-us...@oss.oracle.com [3]http
Re: [Ocfs2-users] Combining OCFS2 with Linux software RAID-0?
Sunil, I am getting too old for all these cluster stacks! Seems that heartbeat2 is deprecated... Can ocfs2 be integrated with pacemaker in the same way as it was possible with heartbeat2 on Suse 10? I know that the RedHat cluster stack cannot, so I used to consider this as an additional feature for Suse Linux. Best Regards, Luis Freitas --- On Mon, 12/14/09, Sunil Mushran sunil.mush...@oracle.com wrote: From: Sunil Mushran sunil.mush...@oracle.com Subject: Re: [Ocfs2-users] Combining OCFS2 with Linux software RAID-0? To: Luis Freitas lfreita...@yahoo.com Cc: Brian Kroth bpkr...@gmail.com, ocfs2-users@oss.oracle.com Date: Monday, December 14, 2009, 4:20 PM That's old. sles11 has added the pacemaker cluster stack that works with clvm. Luis Freitas wrote: Brian, Hmm, I was not aware of this. Seems Novel uses other volume manager, called EVMS, not CLVM (?). From: http://wiki.novell.com/index.php/Linux_Data_Management Some Open Source OCFS2 Features Oracle Linux Certification matrix http://www.novell.com/products/server/oracle/matrix.html OCFS2 project web site http://oss.oracle.com/projects/ocfs2/ OCFS2 Development Roadmap http://oss.oracle.com/osswiki/OCFS2/Roadmap Oracle Cluster File System v2 (OCFS2) is an open source cluster management and * No exclusive write lock capability yet (now every lock request returns: successful). This feature is candidate for SLE10 SP2 (Q1 2008). * OCFS2 on top of a software mirror is not supported yet * *Can be managed by EVMS * ** * OCFS2 offers integration with heartbeat2. Heartbeat2 offers a Resource Agent 'md group take over'. (which enables fail-over of host based mirroring of SAN volumes), but OCFS2 on top of a software mirror is not supported. ... Best Regards, Luis Freitas --- On *Fri, 12/11/09, Brian Kroth /bpkr...@gmail.com/* wrote: From: Brian Kroth bpkr...@gmail.com Subject: Re: [Ocfs2-users] Combining OCFS2 with Linux software RAID-0? To: Luis Freitas lfreita...@yahoo.com Cc: Patrick J. LoPresti lopre...@gmail.com, ocfs2-us...@oss.oracle.com Date: Friday, December 11, 2009, 2:09 PM Luis Freitas lfreita...@yahoo.com http://us.mc514.mail.yahoo.com/mc/compose?to=lfreita...@yahoo.com 2009-12-11 05:40: Patrick, Depending on what you are using, you could use the volume manager to do the striping, but you need to use CLVM. So if you can, go for Heartbeat2+CLVM+OCFS2, all integrated. Not sure but I think Heartbeat2+OCFS2 is only available on the vanilla kernels, not on the enterprise ones. Maybe Suse has support, I don't know, you will have to check. Best Regards, Luis Freitas Just to elaborate on these comments. Last time I checked CLVM required the openais/cman cluster stack, which neither heartbeat nor ocfs2 use (by default). The userspace stack option for ocfs2 in recent mainline kernels added support for the openais stack and pacemaker is required to make heartbeat work with that rather than use it's own cluster stack. Now, you can do an basic LVM linear span, concatenation, or whatever you want to call it without any cluster stack, so long as it's not striped and so long as you heed Sunil's warning about fat fingering changes to the thing while more than one host is using it. That means that if you want to add another LUN to the span you can't do it on the fly. You have to do something like this: # On all nodes: umount /ocfs2 # On all nodes but one: vgchange -an ocfs2span # Or, to be extra safe: halt -p # On the remaining node: vgextend ocfs2span /dev/newlun lvextend -l+100%FREE /dev/mapper/ocfs2span-lv tunefs.ocfs2 -S /dev/mapper/ocfs2span-lv # You might actually need the fs mounted for that last bit, I forget. # Probably a fsck somewhere in there would be wise as well. # Bring the other nodes back up. Brian --- On Wed, 12/9/09, Patrick J. LoPresti lopre...@gmail.com http://us.mc514.mail.yahoo.com/mc/compose?to=lopre...@gmail.com wrote: From: Patrick J. LoPresti lopre...@gmail.com http://us.mc514.mail.yahoo.com/mc/compose?to=lopre...@gmail.com Subject: [Ocfs2-users] Combining OCFS2 with Linux software RAID-0? To: ocfs2-users@oss.oracle.com http://us.mc514.mail.yahoo.com/mc/compose?to=ocfs2-us...@oss.oracle.com, linux-r...@vger.kernel.org http://us.mc514.mail.yahoo.com/mc/compose?to=linux-r...@vger.kernel.org Date: Wednesday, December 9, 2009, 9:03 PM Is it possible to run an OCFS2 file system on top of Linux software RAID? Here is my situation. I have four identical disk chassis that perform hardware RAID internally
Re: [Ocfs2-users] ocfs2 heartbeat and drbd in dual primary
Unni, Well, with this setup you can't properly deal with a split brain scenario. DRDB will have it's heartbeat algorithm, I don't know how it works, but it certainly has one, and OCFS2 has one too. If they are not integrated you could have a scenario where OCFS2 kills one node and DRDB kills the other, and you end up without any of them. OCFS2 chooses the first node in the case of a network failure. So you should either use OCFS2 and DRDB integrated with heartbeat2 (Available only for non-enterprise distros) or setup DRDB to choose the first node too. Btw, someone actually using DRBD can comment better on this. I use OCFS2 for RAC, so I have to use shared storage. Best Regards, Luis Freitas --- On Tue, 12/8/09, unni krishnan unnikrishna...@gmail.com wrote: From: unni krishnan unnikrishna...@gmail.com Subject: Re: [Ocfs2-users] ocfs2 heartbeat and drbd in dual primary To: ocfs2-users@oss.oracle.com Date: Tuesday, December 8, 2009, 3:13 AM Any idea ? On Mon, Dec 7, 2009 at 10:13 PM, unni krishnan unnikrishna...@gmail.com wrote: Hi, I am new to the world of cluster file system and also with ocfs2. I am setting up a cluster as a research project and it looks like : http://picasaweb.google.com/lh/photo/A7CUF3b_SkzJ0NZKbkyyng?feat=directlink The heartbeat will not support ocfs2. Also since in my case the drbd is in dual primary, for me it is not necessary to add it in heartbeat or pacemaker ( Its for me and not sure, if that is correct. ) In my case the crm resource are the two VPSs in the picture. I am using the built in heartbeat provided by ocfs2 and the ocfs2 dlm. My questions are : Since the ocfs2 or drbd is not added in crm, is there any worst thing or any disaster going to happen with this setup ? Please suggest the best method to deal with the ocfs2 + drbd in dual primary + heartbeat + pacemaker I am using : OpenVZ for VPS CentOS 5 as OS -- Regards, Unni -- Regards, Unni ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] 2 node OCFS2 clusters
Srinivas, If this is true then I would sugest OCFS2 is not taking the best decision in this scenario. The node that still has network connectivity should survive instead of the lowest node number. Oracle CRS has heuristics to detect if the network is down and in this scenario the node that lost network conectivy is evicted. That is why it is required to use a switch between the two nodes, instead of a cross cable. OCFS2 should do the same. Best Regards, Luis Freitas --- On Mon, 11/16/09, Srinivas Eeda srinivas.e...@oracle.com wrote: From: Srinivas Eeda srinivas.e...@oracle.com Subject: Re: [Ocfs2-users] 2 node OCFS2 clusters To: Thompson, Mark mark.thomp...@uk.experian.com Cc: ocfs2-users@oss.oracle.com Date: Monday, November 16, 2009, 2:05 PM Thompson, Mark wrote: 2 node OCFS2 clusters Hi Srini, Thanks for the response. So are the following statements correct: If I stop the networking on node 1, node 0 will continue to allow OCFS2 filesystems to work and not reboot itself. If I stop the networking on node 0, node 1 (now being the lowest node?) will continue to allow OCFS2 filesystems to work and not reboot itself. In both the cases node 0 will survive, because that's the node that has lowest node number (defined in cluster.conf). This applies to the scenario where interconnect went down but nodes are healthy and are heartbeating to the disk. I guess I just need to know if it’s possible to have a 2 node OCFS2 cluster that will cope with either one of the nodes dying, and have the remaining node still provide service. If node 0 itself panics, reboots then node 1 will survive. Regards, Mark From: Srinivas Eeda [mailto:srinivas.e...@oracle.com] Sent: 16 November 2009 14:57 To: Thompson, Mark Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] 2 node OCFS2 clusters In a cluster with more than 2 nodes, if a network on one node goes down, that node will evict itself but other nodes will survive. But in a two node cluster, the node with lowest node number will survive no mater on which node network went down. thanks, --Srini Thompson, Mark wrote: Hi, This is my first post here so please be gentle with me. My question is, can you have a 2 node OCFS2 cluster, disconnect one node from the network, and have the remaining node continue to function normally? Currently we have a 2 node cluster and if we stop the NIC that has the OCFS2 o2cb net connection running on it, the other node will reboot itself. I have researched having a 2 node OCFS2 cluster but so far I have been unable to find a clear solution. I have looked at the FAQ regarding quorum, and my OCFS2 init scripts are enabled etc. Is this possible, or should we look at alternative solutions? Regards, Mark This e-mail has come from Experian, the only business to have been twice named the UK's 'Business of the Year’ === Information in this e-mail and any attachments is confidential, and may not be copied or used by anyone other than the addressee, nor disclosed to any third party without our permission. There is no intention to create any legally binding contract or other binding commitment through the use of this electronic communication unless it is issued in accordance with the Experian Limited standard terms and conditions of purchase or other express written agreement between Experian Limited and the recipient. Although Experian has taken reasonable steps to ensure that this communication and any attachments are free from computer virus, you are advised to take your own steps to ensure that they are actually virus free. Companies Act information: Registered name: Experian Limited Registered office: Landmark House, Experian Way, NG2 Business Park, Nottingham, NG80 1ZZ, United Kingdom Place of registration: England and Wales Registered number: 653331 ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users -Inline Attachment Follows- ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Heartbeat Timeout Threshold
Bret, These are my two cents on this subject. If I am not mistaken, there are two heartbeats, one on the network and one on the disk. Failure on either one of them will cause a node to be evicted. If you have network bonding, depending on your configuration and the network topology, when one path fails the switches might have to reconfigure the path to that MAC address, and that could take some time. This can be reduced forcing arp broadcasts on the new path, so the network equipment between your servers can reconfigure itself faster. For the disk heartbeat, assuming that: - You have dual FC cards on each server - You have dual FC switches connected to each other - You have a storage with two or more FC ports, connected to the switches. You have a FC card timeout (Probably set on the HBA firmware or on the driver), the multipath timeout, and a storage controller timeout. Each of these need to be smaller than your cluster heartbeat timeout in order for all the nodes to survive a component failure. For example, an EMC storage has two internal controllers (SP) and a SP failover in the order of two minutes. During this time the LUNs that are routed through the failed SP will be unresponsive, and the FC card on your server will not report this to the O/S until its timeout is reached. After the failover inside the EMC storage, the multipath software on your server will also need to establish the new path to the LUN using the surviving SP. Best Regards, Luis Freitas --- On Fri, 8/7/09, Brett Worth br...@worth.id.au wrote: From: Brett Worth br...@worth.id.au Subject: [Ocfs2-users] Heartbeat Timeout Threshold To: ocfs2-users@oss.oracle.com Date: Friday, August 7, 2009, 9:29 PM I've been using OCFS2 on a 3 way Centos 5.2 Xen cluster for a while now using it to share the VM disk images. In this way I can have live and transparent VM migration. I'd been having intermittent (every 2-3 weeks) incidents where a server would self fence. After configuring netconsole I managed to see that the fencing was due to a heartbeat threshold timeout so I have now increased all three servers to have a threshold of 61 i.e. 2 minutes from the default 31 i.e. 1 minute. So far there have been no panics. I do have a couple of questions though: 1. To get this timeout applied I had to have a complete cluster outage so that I could make all the changes simultaneously. Making the change to single node prevented it from joining in the fun. Do all parameters really need to match before you can join? The timeout threshold seems to be one that could differ from node to node. 2. Even though this appears to have fixed the problem, 2 minutes is a long time to wait for a heartbeat. Even one minute seems like a very long time. I assume that missing a heartbeat would be a symptom of a very busy filesystem but for a packet to take over a minute to get over the wire is odd. Or is it that the heartbeats are actually being lost for an extended period? Is this a network problem? All my nodes communicate heartbeat on a dedicated VLAN. Regards Brett PS: If anyone is planning to do Xen like this my main piece of advice is that you must put a ceiling on how much RAM the Dom0 domain can use. If you don't it will expand to use all non-vm memory for buffer cache so that when you try to do a migration to it there is no ram left. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] OCFS2 hosting and running binaries
Saul, OCFS2 1.2 doesnt has support for indexed directories. So you will need to have a cleaning procedure for the database dump directories to keep the quantity of log files reasonable. Unless you dont mind a ls hanging when you try to find a trace. I am not sure if OCFS 1. has support for indexed directories on the enterprise tree. Also when sharing database binaries there is a need to use a special type of file that is different on each node for certain configuration files . The OCFS2 1.2 enterprise has support for this. This was necessary on 10g, not sure about 11g. There are some files that need to be outside of OCFS2 1.2 so you will need to create some symlinks for each database too. And the mount options for database files and binaries/traces are different, so you will need to plan your filesystems in a way to keep datafiles, logfiles, archivelogs separated from your binaries and trace/dump locations. One important point is that when you do this you will be unable to do a rolling patch application. Depending on your availability requirements this could be important. Also you introduce a single point of failure. If someone accidentaly damage or delete a important database or configuration file the entire cluster will fail. When you have non-shared binaries filesystems one would expect the other nodes to continue operating if this happens. Regards, Luis --- On Tue, 6/9/09, Sunil Mushran sunil.mush...@oracle.com wrote: From: Sunil Mushran sunil.mush...@oracle.com Subject: Re: [Ocfs2-users] OCFS2 hosting and running binaries To: Saul Gabay sa...@herbalife.com Cc: Server Ops_Linux serverops_li...@herbalife.com, ocfs2-users@oss.oracle.com Date: Tuesday, June 9, 2009, 5:20 PM Sure. One can use ocfs2 to host almost anything. The one exception is the crs_home. crs_home needs to be on a local volume. OCFS 1.2/1.4 has two limits. Like ext3, the number of sub-directories in _a_ directory cannot exceed 32000. (There is no limit to the number of subdirs in a volume.) The other limit is the volume size. Currently max is 16T. There is no limit to the number of files in a volume. (The two limits have been relaxed in mainline for few kernel versions.) As far as performance goes, I have yet to see a benchmark that shows ocfs2 slower than gfs/gfs2. For certification, please check metalink. Sunil Saul Gabay wrote: We are currently using OCFS2 to host multiple Oracle 10g RAC databases on Itanium servers running Redhat AS 4.7, we are running this OCFS2 version so far with no issues ocfs2-2.6.9-78.0.13.EL-1.2.9-1.el4 We would like to use OCFS2 to host binaries files for the database and / or application. This will be 4 active nodes mounting an OCFS2 formatted LUN through iSCSI. What are the issues, caveats or things we need to be aware if we take this approach. Like, is there a limit on the number of files or directories hosted on OCFS2? Are there a performance issue / degradation in comparison with GFS hosting binaries files? What are the good, bad and ugly of OCFS2 in comparison with GFS hosting binaries files? Is OCFS2 certified by Oracle to run database/application binaries? Please advice what is your experience on this topic, it will be greatly appreciated. /*/ /*/ /*/Saul/*/ ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] [Fwd: Re: Unable to fix corrupt directories with fsck.ocfs2]
Robin, To me, anyone else includes the kernel of the current node. Well, if it is unclear the man page should be revised. Also a big warning message on ocfs2.fsck would be nice, after all we all make mistakes. But this is only my two cents. Running fsck on any journaled filesystem will replay the journal. This will cause corruption if the filesystem is mounted read/write, even if the filesystem was not corrupted on the first place. You could mount it read only, but you risk getting a kernel panic when the filesystem suddenly changes if fsck corrects something. I am not aware of any filesystem that can withstand a online fsck. Sun ZFS can do online correction, but it doesnt have a fsck tool. Regards, Luis --- On Tue, 5/19/09, Robin Garner robin.gar...@scu.edu.au wrote: From: Robin Garner robin.gar...@scu.edu.au Subject: Re: [Ocfs2-users] [Fwd: Re: Unable to fix corrupt directories with fsck.ocfs2] To: ocfs2-users@oss.oracle.com Date: Tuesday, May 19, 2009, 11:05 PM Joel Becker wrote: On Tue, May 19, 2009 at 02:49:31PM +1000, Robin Garner wrote: Robin Garner wrote: Yes. This is a 24/7 application (at least during semester), and arranging extended downtime is a challenge. Ok, you ran fsck against a live filesystem and skipped the cluster locking with the '-F' option. So now you have two problems. 1) The original directory problem. 2) The duplicate blocks created by your fsck of a mounted filesystem. Do you have backups? Joel OK, now I'm confused: The man page for fsck.ocfs2 says -F Usually fsck.ocfs2 will check with cluster services and the DLM to make sure that no one else in the cluster is actively using the device before proceeding. -F skips this check and should only be used when it can be guaranteed that there can be no other users of the device while fsck.ocfs2 is running. To me my colleagues no one else in the cluster is actively using the device means that the filesystem must be mounted on *at most* one node in the cluster (the node doing the fsck). That's what we did. This filesystem is normally mounted by both nodes of a 2-node cluster. We had cleanly unmounted the filesystem on the other node. fsck.ocfs2 without '-F' gave errors, but then mounted.ocfs2 claimed the disk was mounted on both nodes. Eventually we shut down the other node, and mounted.ocfs2 still thought it had it mounted. At this point we used '-F'. I can't see any reference in the man page about not doing an fsck on a mounted disk. e2fsck for example says this: WARNING!!! Running e2fsck on a mounted file system may cause SEVERE filesystem damage. Do you really want to continue (y/n)? when you try to fsck a mounted filesystem. May I suggest that fsck.ocfs2 do something similar ? Perhaps 'everyone knows' you can't run fsck on a mounted filesystem, but we were assuming that ocfs2 being a modern cluster filesystem might be a little more advanced. Apparently not. We'll try to salvage the data another way (we believe the directory corruption is some way down the directory tree), and pull missing data back from backups. Robin ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Encountered disk I/O error 19502
Diane, Are you using ASM and OCFS2? Some of the log messages point to a disk group. Can you post a copy of your /etc/fstab with the mount options? Regards, Luis --- On Mon, 4/6/09, Diane Petersen diane_peter...@yahoo.com wrote: From: Diane Petersen diane_peter...@yahoo.com Subject: Re: [Ocfs2-users] Encountered disk I/O error 19502 To: Karim Alkhayer kkha...@gmail.com, ocfs2-users@oss.oracle.com Date: Monday, April 6, 2009, 1:42 PM Hi, We already have TAF implemented, unfortunately that doesn't help. It suppose TAF might help if the instance was terminated, but that's not happening instead it terminates these individual sessions directly. This happens on both nodes during writes to the OCFS2 partition at random times but never at the same time. There is nothing else in the db alert log or crs logs other than what I've included below. Thanks, Diane Petersen ServerCare, Inc. From: Karim Alkhayer kkha...@gmail.com To: Diane Petersen diane_peter...@yahoo.com; ocfs2-users@oss.oracle.com Sent: Monday, April 6, 2009 9:11:06 AM Subject: RE: [Ocfs2-users] Encountered disk I/O error 19502 Hello Diane, I believe that implementing TAF could help a bit in this case, at least to become transparent to the end users, unless of course, the following points are blocking in your case: 1. ALTER SESSION statements are lost: Statements such as ALTER SESSION ... are not automatically re-issued to the server following a failover. This can have a significant effect on application behavior. For example: ALTER SESSION SET NLS_DATE_FORMAT='-MM-DD'; select sysdate from dual; Result 2009-01-31 Fail over the connection select sysdate from dual; Result 31-JAN-09 2. In-progress transactions must be rolled back 3. Continuing work on existing cursors may raise an error (eg: ORA-25401 cannot continue fetches) 4. Failed over selects may take time to re-position (when FAILOVER_TYPE=SELECT) 5. Client awareness of a Failover Can we have an overview of the database setup, nature of transactions, and parameters? It would also help to examine the troublesome node behavior and recovery measures. Best regards, Karim Alkhayer From:ocfs2-users-boun...@oss.oracle.com [mailto:ocfs2-users-boun...@oss.oracle.com] On Behalf Of Diane Petersen Sent: Monday, April 06, 2009 4:06 PM To: ocfs2-users@oss.oracle.com Subject: [Ocfs2-users] Encountered disk I/O error 19502 Hi, We have a 2-node 11g RAC database running OCFS2 1.4.1-1.el5 with Linux kernel 2.6.18-92.1.17.el5 64-bit. Lately we've been seeing errors on both nodes almost ever other day. The system administrator has checked the SAN array and said there are no issues being reported. Another part of the problem, it appears the instances alter the service_names parameter not allowing new connections to the node with the reported error, but also terminate sessions already connected using the RAC service. The errors all start with - Encountered disk I/O error 19502 - and contain the following: ARC2: Encountered disk I/O error 19502 (ifxdb2) Errors in file /u01/app/oracle/diag/rdbms/ifxdb/ifxdb2/trace/ifxdb2_arc2_15414.trc: ORA-19502: write error on file /u03/arch/2_1917_656008464.dbf, block number 155649 (block size=512) ORA-27072: File I/O error Linux-x86_64 Error: 5: Input/output error Additional information: 4 Additional information: 155649 Additional information: -1 ORA-19502: write error on file /u03/arch/2_1917_656008464.dbf, block number 155649 (block size=512) Errors in file /u01/app/oracle/diag/rdbms/ifxdb/ifxdb2/trace/ifxdb2_arc2_15414.trc: ORA-19502: write error on file /u03/arch/2_1917_656008464.dbf, block number 155649 (block size=512) ORA-27072: File I/O error Linux-x86_64 Error: 5: Input/output error Additional information: 4 Additional information: 155649 Additional information: -1 ORA-19502: write error on file /u03/arch/2_1917_656008464.dbf, block number 155649 (block size=512) ARC2: I/O error 19502 archiving log 10 to '/u03/arch/2_1917_656008464.dbf' ARCH: Archival stopped, error occurred. Will continue retrying ORACLE Instance ifxdb2 - Archival Error ORA-16038: log 10 sequence# 1917 cannot be archived ORA-19502: write error on file , block number (block size=) ORA-00312: online log 10 thread 2: '+REDO1/ifxdb/onlinelog/group_10.265.656605479' Errors in file /u01/app/oracle/diag/rdbms/ifxdb/ifxdb2/trace/ifxdb2_arc2_15414.trc: ORA-16038: log 10 sequence# 1917 cannot be archived ORA-19502: write error on file , block number (block size=) ORA-00312: online log 10 thread 2: '+REDO1/ifxdb/onlinelog/group_10.265.656605479' Sun Apr 05 15:05:16 2009 ALTER SYSTEM SET service_names='ifxdb.gointranet.com' SCOPE=MEMORY
Re: [Ocfs2-users] OCFS2 FS with BACKUP Tools/Vendors
We use Veritas Netbackup 6.0 (With MP7) to backup some export files on a OCFS2 filesystem, didnt have any issues so far but we didnt test it too much either. OCFS2 can be regarded as a regular file system, unlike OCFS1 which could not be accessed without a special version of cp and tar. But I doubt Veritas would directly support any issues we eventually find when backing up (or restoring) files from OCFS2. If you really want support you might need to copy the files from OCFS2 with operating system commands to a ext3 filesystem and then copy from there. Regards, Luis --- On Thu, 4/2/09, Daniel Keisling daniel.keisl...@ppdi.com wrote: From: Daniel Keisling daniel.keisl...@ppdi.com Subject: Re: [Ocfs2-users] OCFS2 FS with BACKUP Tools/Vendors To: Bumpass, Brian brian.bump...@wachovia.com, ocfs2-users@oss.oracle.com Date: Thursday, April 2, 2009, 5:22 PM I use HP Data Protector. OCFS2 is supported in v6.0. From: ocfs2-users-boun...@oss.oracle.com [mailto:ocfs2-users-boun...@oss.oracle.com] On Behalf Of Bumpass, Brian Sent: Thursday, April 02, 2009 1:31 PM To: ocfs2-users@oss.oracle.com Subject: [Ocfs2-users] OCFS2 FS with BACKUP Tools/Vendors My apology up front if this has been discussed already. I've reviewed the archives back to Nov. 2005 and found little of anything. I need some information concerning support for OCFS2 by backup products. Currently we use IBM/Tivoli's TSM tool. They don't support OCFS2 filesystems. And it looks like they have no intent to supporting the FS in next releases.Note... They do support their own NAS FS, GPFS. But this costs extra. Additionally, the small testing I have done is that a file under an OCFS2 FS backs up and recovers quite nicely.I have not tested using ACL-lists. But don't really care about those. This issue comes down to support. So... I guess what I am looking for is some indication of what the user community with OCFS2 and doing backups has been along similar issues. Sorry... The environment being supported is SLES 10 SP2 64-bit on DELL HP hardware. Thanks in advance, -B __ This email transmission and any documents, files or previous email messages attached to it may contain information that is confidential or legally privileged. If you are not the intended recipient or a person responsible for delivering this transmission to the intended recipient, you are hereby notified that you must not read this transmission and that any disclosure, copying, printing, distribution or use of this transmission is strictly prohibited. If you have received this transmission in error, please immediately notify the sender by telephone or return email and delete the original transmission and its attachments without reading or saving in any manner. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] ASM over OCFS2 vs. Standard locally managed tablespaces
Karim, I dont see why run ASM over OCFS2. It seems to be a useless overhead. Either you run ASM or OCFS2. Btw, neither ASM nor OCFS2 are smart enough to detect that some LUNs are faster than others. ASM expects each diskgroup to be comprised of LUNs of similar performance in order for it's load balancing algorithms to work. OCFS2, as far as I know doesnt have this type of management built in. See: http://www.oracle.com/technology/products/database/asm/pdf/take%20the%20guesswork%20out%20of%20db%20tuning%2001-06.pdf Section: ASM Best practices and principals. About the performance, ASM is said to have similar performance to raw devices in a SAME layout, being tightly integrated to Oracle. OCFS2 has some overheads that are inherent to a file system, like cache management, locking, context switching, so it is likely to use more CPU power than ASM. But I dont remember any specific benchmark comparing those. Also, keep in mind that when you use a filesystem you are using part of the memory for the filesystem cache. When using RAW or ASM you would need to allocate this memory to the block buffer in order to compare results. Regards, Luis Hello All, Are there any benchmarks with respect to performance with respect to ASM over OCFS2 vs. standard locally managed tablespaces? In our environment, data files hosting tables/lobs are stored on a RAID6 disk array with 10K rpm disks, whilst indices are stored on a different RAID6 disk array with 15K rpm disks. We’re using oracle managed files for the rollback/undo and temporary tablespaces. Would ASM over OCFS2 be smart enough to detect the fast LUNs? Appreciate your thoughts. Best regards, Karim ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] ASM over OCFS2 vs. Standard locally managed tablespaces
Karim, This is one big environment. I dont see how ASM over OCFS2 would give easier administration than only ASM or only OCFS2. The only situation I see that is reasonable to use ASM over a cooked filesystem would be when using a NAS device that doesnt support direct block access. Also I dont understand why you say that RAC needs a shared filesystem. When you use ASM you dont need to have a shared filesystem. If you go ASM, you will need to install the cluster services on each server that shares a ASM diskgroup, even if it has no RAC databases. The same goes to OCFS2, you will need to install the OCFS2 services on each server that shares a OCFS2 filesystem. If you do both, you will have to install both. ASM has some interesting storage like features, for example extended clusters and online disk reorganization. You can do some of these with OCFS2. For example, adding a disk. But try to remove a OCFS2 volume with the database online and not disrupt your users. ASM can do that. On the other hand ASM is less transparent. You have little control on how the data is layout, and the only tool to manage files is a ftp like client, that you need to use to delete dangling files or if you need to backup something manually. Database backups would usually need to go through RMAN. On OCFS2 you can use standard operating system commands to manage the datafiles. ASM also has no recovery tools, like fsck. Regards, Luis --- On Mon, 2/9/09, Karim Alkhayer kkha...@gmail.com wrote: From: Karim Alkhayer kkha...@gmail.com Subject: RE: [Ocfs2-users] ASM over OCFS2 vs. Standard locally managed tablespaces To: lfreita...@yahoo.com, ocfs2-users@oss.oracle.com Date: Monday, February 9, 2009, 10:47 AM We’re using OCFS2 for RAC on top of SLES9, which we’re going to upgrade to SLES10. Around 10 TB RAID6 multi disk arrays, 5 databases on RAC, and 5 single instances standby for the primary site As there is no AI component in ASM to detect the fast LUNs, and RAC on SLES requires a shared file system. Therefore, on a set of identical LUNs, in terms of capacity and speed, ASM should take care of distributing the balance over LUNs, and OCFS2 is expected to work even better if these LUNs are placed on several disk groups (arrays) How would this scenario (ASM over OCFS2) work? What are the cons and pros? Keep in mind that the goal of such a concept is provide performance and reliability with the least possible administration Appreciate your thoughts Best regards, Karim From: ocfs2-users-boun...@oss.oracle.com [mailto:ocfs2-users-boun...@oss.oracle.com] On Behalf Of Luis Freitas Sent: Monday, February 09, 2009 2:16 PM To: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] ASM over OCFS2 vs. Standard locally managed tablespaces Karim, I dont see why run ASM over OCFS2. It seems to be a useless overhead. Either you run ASM or OCFS2. Btw, neither ASM nor OCFS2 are smart enough to detect that some LUNs are faster than others. ASM expects each diskgroup to be comprised of LUNs of similar performance in order for it's load balancing algorithms to work. OCFS2, as far as I know doesnt have this type of management built in. See: http://www.oracle.com/technology/products/database/asm/pdf/take%20the%20guesswork%20out%20of%20db%20tuning%2001-06.pdf Section: ASM Best practices and principals. About the performance, ASM is said to have similar performance to raw devices in a SAME layout, being tightly integrated to Oracle. OCFS2 has some overheads that are inherent to a file system, like cache management, locking, context switching, so it is likely to use more CPU power than ASM. But I dont remember any specific benchmark comparing those. Also, keep in mind that when you use a filesystem you are using part of the memory for the filesystem cache. When using RAW or ASM you would need to allocate this memory to the block buffer in order to compare results. Regards, Luis Hello All, Are there any benchmarks with respect to performance with respect to ASM over OCFS2 vs. standard locally managed tablespaces? In our environment, data files hosting tables/lobs are stored on a RAID6 disk array with 10K rpm disks, whilst indices are stored on a different RAID6 disk array with 15K rpm disks. We’re using oracle managed files for the rollback/undo and temporary tablespaces. Would ASM over OCFS2 be smart enough to detect the fast LUNs? Appreciate your thoughts. Best regards, Karim ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] ocfs2 hangs during webserver usage
David, You probably could get away with writing the log files to OCFS2, but in different files. The access_log gets a small write on each access, and writes will be serialized, since no two processes can write to it at the same time. So it should be related more to the size of the disk queues and the speed of the interconnect than to the speed of the SAN itself. If you keep writing to the same file from different nodes, OCFS2 will need to keep flushing pages and bouncing write locks from one node to another on each page hit. This is very nonscalable, as you got a single serialized write on each page access. The extent size would play a role on this also, as Sunil pointed. You could check if the extent size is different on your test environment. The mkfs tool might have defaulted a larger extent size if the total size of the filesystem is larger. (Sunil, correct me on this if I am wrong). Regards, Luis --- On Tue, 1/27/09, jmose...@corp.xanadoo.com jmose...@corp.xanadoo.com wrote: From: jmose...@corp.xanadoo.com jmose...@corp.xanadoo.com Subject: Re: [Ocfs2-users] ocfs2 hangs during webserver usage To: David Johle djo...@industrialinfo.com Cc: lfreita...@yahoo.com, ocfs2-users@oss.oracle.com, ocfs2-users-boun...@oss.oracle.com Date: Tuesday, January 27, 2009, 10:32 PM As others have indicated, I don't think that's going to work very well. You've got two different nodes trying to write to the same file constantly. I would keep each server's log on a locally mounted file system, or simply keep the logs on the OCFS2 filesystem, but have each node write to different log files. Yeah, that makes parsing access_logs slightly more of a problem for producing hit reports, etc, but I think you'll notice performance improve. James Moseley David Johle djo...@industria linfo.com To Sent by: lfreita...@yahoo.com ocfs2-users-bounc cc e...@oss.oracle.com ocfs2-users@oss.oracle.com Subject Re: [Ocfs2-users] ocfs2 hangs 01/27/2009 04:38 during webserver usage PM Yes, that is the case, multiple nodes with the same log file open being written to at once. Worked well during all my testing environment stress tests, and even worked great in production for over a month. At 02:56 PM 1/27/2009, Luis Freitas wrote: David, You said you were keeping the apache log files on OCFS2. Are you using the same log file (access_log and error_log) for all the nodes? That is a single access_log that is writen by both nodes simultaneosly? Regards, Luis ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] ocfs2 hangs during webserver usage
Sunil, Hmm, I thought you meant the cluster size. Regards, Luis --- On Wed, 1/28/09, Sunil Mushran sunil.mush...@oracle.com wrote: From: Sunil Mushran sunil.mush...@oracle.com Subject: Re: [Ocfs2-users] ocfs2 hangs during webserver usage To: lfreita...@yahoo.com Cc: ocfs2-users@oss.oracle.com Date: Wednesday, January 28, 2009, 1:08 AM Luis Freitas wrote: The extent size would play a role on this also, as Sunil pointed. You could check if the extent size is different on your test environment. The mkfs tool might have defaulted a larger extent size if the total size of the filesystem is larger. (Sunil, correct me on this if I am wrong). The extend size I was referring to is not related to any ondisk size. It depends on the location the write() is issued to. Say you have a 100K sized file and you issue a write() at offset 1G. If the fs supports sparse files, it will allocate (and init) blocks only around the 1G location, leaving a hole in between. A non-sparsefile supporting fs, on the other hand, will need to allocate (and init) the entire space between 100K and 1G. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] OCFS2 replication
Brian, We have EMC Mirrorview here, but it doesnt allow an active-active mirror, we mount only the primary lun on both servers, and the mirror lun needs to be promoted to primary, and then mounted in case of a disaster recovery. We have a script to do that. Regards, Luis --- On Thu, 1/22/09, Brian Kroth bpkr...@gmail.com wrote: From: Brian Kroth bpkr...@gmail.com Subject: Re: [Ocfs2-users] OCFS2 replication To: CA Lists li...@creativeanvil.com Cc: ocfs2-users@oss.oracle.com ocfs2-users@oss.oracle.com Date: Thursday, January 22, 2009, 9:35 PM Our iscsi San (equallogic) does block level replication so we were thinking of trying to set something up soon so that we could have some nodes in another building connected via fiber to provide site level failover. I'll report back our experiences when we do that, but I imagine it would be similar to drdb with nice interconnect. Brian On Jan 22, 2009, at 2:08 PM, CA Lists li...@creativeanvil.com wrote: Can't say I've replicated it between two sites, but definitely between two physical servers. I used drbd in my particular case. Here's a small blog entry I put together a while back about what I did. Hopefully it's helpful: http://www.creativeanvil.com/blog/2008/how-to-create-an-iscsi-san-using-heartbeat-drbd-and-ocfs2/ Joe Koenig Creative Anvil, Inc. Phone: 314.692.0338 1346 Baur Blvd. Olivette, MO 63132 j...@creativeanvil.com http://www.creativeanvil.com David Schüler wrote: What about drbd? Von: ocfs2-users-boun...@oss.oracle.com [mailto:ocfs2-users-boun...@oss.oracle.com] Im Auftrag von Garcia, Raymundo Gesendet: Donnerstag, 22. Januar 2009 20:46 An: ocfs2-users@oss.oracle.com Betreff: Re: [Ocfs2-users] OCFS2 replication RSYNC is not real time…. any other suggestion…? I treid RSYNC already… From: ocfs2-users-boun...@oss.oracle.com [mailto:ocfs2-users-boun...@oss.oracle.com] On Behalf Of Sérgio Surkamp Sent: Thursday, January 22, 2009 12:32 PM Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] OCFS2 replication Try rsync. Garcia, Raymundo wrote: Hello… I am trying to replicate a OCFS2 filesystem in a site A to another OCFS@ based partition in another site B … I have tried several products, inmage, steeleye.. etc without any luck.. those programs help me to replicate the filesystem but nnot the OCFS2 mounted … I assume that this is because that most software based replication system work on the block level instead of the file leve… I wonder is anyone have tried to replicate OCFS2 between 2 sites…. Thanks Raymundo Garcia The information contained in this message may be confidential and legally protected under applicable law. The message is intended solely for the addressee(s). If you are not the intended recipient, you are hereby notified that any use, forwarding, dissemination, or reproduction of this message is strictly prohibited and may be unlawful. If you are not the intended recipient, please contact the sender by return e-mail and destroy all copies of the original message. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users Regards, -- .:':. .:' ` Sérgio Surkamp | Gerente de Rede :: ser...@gruposinternet.com.br `:. .:' `:, ,.:' *Grupos Internet S.A.* `: :' R. Lauro Linhares, 2123 Torre B - Sala 201 : : Trindade - Florianópolis - SC :.' :: +55 48 3234-4109 : ' http://www.gruposinternet.com.br Virus checked by G DATA AntiVirusKit Version: AVF 19.226 from 18.01.2009 Virus news: www.antiviruslab.com ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Filesystem Block Size w// DB_BLOCK_SIZE
Karin, Even if they wont upgrade the entire system, I see no reason to not upgrade the kernel. On Suse, upgrading the kernel would upgrade the OCFS2 version that is being used. And if there is trouble you could revert to the older kernel. Regards, Luis --- On Sat, 12/20/08, Karim Alkhayer kkha...@gmail.com wrote: From: Karim Alkhayer kkha...@gmail.com Subject: RE: [Ocfs2-users] Filesystem Block Size w// DB_BLOCK_SIZE To: 'Sunil Mushran' sunil.mush...@oracle.com Cc: lfreita...@yahoo.com, ocfs2-users@oss.oracle.com Date: Saturday, December 20, 2008, 8:57 AM Folks, We're using two StoreWay FDA storage with around 5 TB on each. The connection from the servers to the storage is through fiber channel RAID 6 is used with 7400RPM for non-index data files whereas 10K RPM is for the indices ASM is not used, the tablespaces are sized and organized based on the business requirements The platform provider is a hesitating to upgrade SLES from SP3 to SP4, so that we can benefit from the latest possible version of OCFS2. Therefore, any potential enhancement for the current scenario would be the best chance for now Appreciate your thoughts Cheers, Karim -Original Message- From: Sunil Mushran [mailto:sunil.mush...@oracle.com] Sent: Friday, December 19, 2008 11:57 PM To: Karim Alkhayer Cc: lfreita...@yahoo.com; ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] Filesystem Block Size w// DB_BLOCK_SIZE True. The only point I would like to add is that you are using a 2+ yr old version of the fs. You should upgrade to atleast SLES9 SP4. Luis Freitas wrote: Karim, This is not OCFS2 related, it is more related to the disk hardware capabilities and how it works. That will depend on your OS, HBAs and storage, and the workload. There is a maximum queue depth associated with each LUN, so if you use several LUNs on the same device, you could achieve more outstanding scsi commands open on the controller. But each port will also have a maximum queue depth that cannot be excedeed, so at some point using extra LUNs wont give you this advantage. If the storage/disk have more outstanding requests it could provide a better performance by reordering them to provide a larger overall throughput, given that the storage hardware supports this. Probably it supports it, since even low end SATA disks supports reordering nowadays. On the other hand the database (Or ASM for that matter) has no ideia that these luns are from the same device, so it will spread the data evenly accross them, and your data will end up scattered accross the disk, instead of concentrated at the start of the disks. This will bring a performance penalty, most noticeable on full table scan operations since they read the data sequentially from the start to the end. If you tune the tables extent size you can work around this problem. Regards, Luis --- On Thu, 12/18/08, Karim Alkhayer kkha...@gmail.com wrote: From: Karim Alkhayer kkha...@gmail.com Subject: [Ocfs2-users] Filesystem Block Size w// DB_BLOCK_SIZE To: ocfs2-users@oss.oracle.com Date: Thursday, December 18, 2008, 9:20 PM Hello All, We're hosting DB1 and DB2 with db_block_size set to 8K, 16K respectively File system creation is done with mkfs.ocfs2 -b 4K -C 32K -N 4 -L LABLE /dev/mapper/ Mount is done with: ocfs2 _netdev,datavolume,nointr 0 0 I'd like to know if we can separate most of the tablespaces on different LUNs, even if they're on the same disk group sometimes, is it possible to gain better performance? Is the impact limited to the time of creating the tablespaces only (assuming they're pre-sized properly)? Current OCFS2 version is 1.2.1 Current OCFS2 components: ocfs2-tools-1.1.4-0.5 ocfs2console-1.1.4-0.5 # uname -r Kernel 2.6.5-7.257-default # cat /etc/SuSE-release SUSE LINUX Enterprise Server 9 (ia64) VERSION = 9 PATCHLEVEL = 3 Oracle 10.1.0.5 Appreciate your input Best regards, Karim ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] hardware needed for OCFS
If your NAS can do iSCSI you could use it to provide a shared block device. The performance wont be as good as a SAN as the data has to go through the kernel TCP/IP stack, it can be comparable if you have a iSCSI IP accelerator board that simulates a hba. Regards, Luis --- On Fri, 12/19/08, David Coulson da...@davidcoulson.net wrote: From: David Coulson da...@davidcoulson.net Subject: Re: [Ocfs2-users] hardware needed for OCFS To: Pete Kay pete...@gmail.com Cc: ocfs2-users@oss.oracle.com Date: Friday, December 19, 2008, 9:57 AM OCFS2 requires a block device (local disk, direct attached, SAN), so it's not going to work with a NAS (Samba, NFS, etc) mount. You can, of course, share a OCFS2 filesystem using Samba or NFS. Pete Kay wrote: Hi, Does OCFS require NAS hardware to run or does normal PC hard disk work? Thanks, Pete ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] NFS Failover
I found some IBM papers on GPFS that indicates they have a working solution for this. They seem to use a standard feature of the NFS client, NFS lock recovery, but the server was modified to initiate this process when one of the nodes die, and also to maintain lock coeherence. I am pasting text from the one of the papers below: http://www.redbooks.ibm.com/redpapers/pdfs/redp4400.pdf The following system prerequisites must be met before you begin the installation and configuration: A Linux 2.6 kernel Distributions currently supported are Red Hat Enterprise Linux (RHEL) versions 4 and 5 and SUSE Linux Enterprise Server (SLES) versions 9 and 10. Operating system patches – If NLM locking is required, a kernel patch that updates the lockd daemon to propagate locks to the clustered file system must be applied. This patch is currently available at: http://sourceforge.net/tracker/?atid=719124group_id=130828func=browse/ Depending on the version of SLES you are using, this patch might exist partially. If this condition exists, you might need to resolve certain conflicts. Contact your support organization if necessary. – To permit NFS clients to reclaim their locks with a new server after failover, the reclaim message from statd must appear to come from the IP address of the failing node (and not the node that took over, which is the one that will actually send the message). On SUSE, statd runs in the kernel and does not implement the interface to support this requirement (notification only, -N option). Therefore, on SUSE, the common NFS utilities (sm-notify in the user space) are needed to implement this function. The patches required for the util-linux package are: • Support statd notification by name (patch-10113) http://support.novell.com/techcenter/psdb/2c7941abcdf7a155ecb86b309245e468.html • Specify a host name for the -v option (patch-10852) http://support.novell.com/techcenter/psdb/e6a5a6d9614d9475759cc0cd033571e8.html • Allow selection of IP source address on command line (patch-9617) http://support.novell.com/techcenter/psdb/c11e14914101b2debe30f242448e1f5d.html/ – For RHEL, use of nfs-utils 1.0.7 is required for rpc.statd fixes. See: http://www.redhat.com/ They also use load balancing using DNS round robin (?!), not sure how they can make this work. Regards, Luis --- On Tue, 12/9/08, Sunil Mushran [EMAIL PROTECTED] wrote: From: Sunil Mushran [EMAIL PROTECTED] Subject: Re: [Ocfs2-users] NFS Failover To: [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com Date: Tuesday, December 9, 2008, 5:09 PM I forgot about fsid. That's how it identifies the device. Yes, it needs to be the same. Yes, the inode numbers are consistent. It is the block number of the inode on disk. Afraid cannot help you with failover lockd. Sunil Luis Freitas wrote: Sunil, They are not waiting, the kernel reconnects after a few seconds, but just dont like the other nfs server, any attempt to access directories or files after the virtual IP failover to the other nfs server was resulting in errors. Unfortunatelly I dont have the exact error message here anymore. We found a parameter on the nfs server that seems to fix it, fsid. If you set this to the same number on both servers it forces both of them to use the same identifiers. Seems that if you dont, you need to guarantee that the mount is done on the same device on both servers, and we cannot do this since we are using powerpath. I would like to confirm if the inode numbers are consistent accross servers? That is: [EMAIL PROTECTED] concurrents]$ ls -il total 8 131545 drwxr-xr-x 2 100 users 4096 Dec 9 12:12 admin 131543 drwxrwxrwx 2 root dba 4096 Dec 4 08:53 lost+found [EMAIL PROTECTED] concurrents]$ Directory admin (Or other directories/files) is always be inode number 131545, no mater on what server we are? Seems to be so, but I would like to confirm. About the metadata changes, this share will be used for log files (Actually, for a Oracle eBusiness Suite concurrent log and output files), so we can tolerate if a few of the latest files are lost during the failover. The user can simply run his report again. Also if some processes hang or die during the failover it can be tolerated, as the internal manager can restart them. Preferably processes should die instead of hanging. But I am concerned about dangling locks on the server. Not sure on how to handle those. On the NFS-HA docs some files on /var/lib/nfs are copied using scripts every few seconds, but this does not seem to be a foolprof way. I overviewed the docs from NFS-HA sent to the list, they are usefull, but also very Linux HA centric, and require the heartbeat2 package. I wont install another cluster stack, since I already have CRS here. Do anyone has pointers on a similar setup with CRS? Best Regards, Luis
Re: [Ocfs2-users] NFS Failover
Sunil, They are not waiting, the kernel reconnects after a few seconds, but just dont like the other nfs server, any attempt to access directories or files after the virtual IP failover to the other nfs server was resulting in errors. Unfortunatelly I dont have the exact error message here anymore. We found a parameter on the nfs server that seems to fix it, fsid. If you set this to the same number on both servers it forces both of them to use the same identifiers. Seems that if you dont, you need to guarantee that the mount is done on the same device on both servers, and we cannot do this since we are using powerpath. I would like to confirm if the inode numbers are consistent accross servers? That is: [EMAIL PROTECTED] concurrents]$ ls -il total 8 131545 drwxr-xr-x 2 100 users 4096 Dec 9 12:12 admin 131543 drwxrwxrwx 2 root dba 4096 Dec 4 08:53 lost+found [EMAIL PROTECTED] concurrents]$ Directory admin (Or other directories/files) is always be inode number 131545, no mater on what server we are? Seems to be so, but I would like to confirm. About the metadata changes, this share will be used for log files (Actually, for a Oracle eBusiness Suite concurrent log and output files), so we can tolerate if a few of the latest files are lost during the failover. The user can simply run his report again. Also if some processes hang or die during the failover it can be tolerated, as the internal manager can restart them. Preferably processes should die instead of hanging. But I am concerned about dangling locks on the server. Not sure on how to handle those. On the NFS-HA docs some files on /var/lib/nfs are copied using scripts every few seconds, but this does not seem to be a foolprof way. I overviewed the docs from NFS-HA sent to the list, they are usefull, but also very Linux HA centric, and require the heartbeat2 package. I wont install another cluster stack, since I already have CRS here. Do anyone has pointers on a similar setup with CRS? Best Regards, Luis --- On Mon, 12/8/08, Sunil Mushran [EMAIL PROTECTED] wrote: From: Sunil Mushran [EMAIL PROTECTED] Subject: Re: [Ocfs2-users] NFS Failover To: [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com Date: Monday, December 8, 2008, 11:47 PM While the nfs protocol is stateless and thus should handle failing-over, the procedures themselves are synchronous. Meaning, I am not sure how a nfs client will handle getting a ok for some metadata change (mkdir, etc) just before a server dies and is recovered by another node. If the op did not make it to the journal, it would be a null op. But the nfs client would not know that as the server has failed-over. This is a qs for nfs. What is the stack of the nfs clients? As in, what are they waiting on? Luis Freitas wrote: Hi list, I need to implement a High available NFS server. Since we already have OCFS2 here for RAC, and already have a virtual IP on the RAC server that failovers automatically to the other node, it seems a natural choice to use it too for our NFS needs. We are using OCFS2 1.2. (Upgrade to 1.4 is not on our current plans) We did a preliminary failover test, and the client that mounts the filesystem (Actually a solaris box) doesnt like the failover. We expect some errors and minor data loss and can tolerate them as a transient condition, but the problem is that the mounted filesystem on the client becomes useless until we umount and remount it again. I suspect that NFS uses inode numbers on underlying filesystem to create handles that it passes on to clients, but I am not sure on how this is done. Anyone know if we can achieve a failover without needing to remount he nfs share on the clients? Any special options are needed mounting the OCFS2 filesystem and also for exporting it as NFS, or on the client? Best Regards, Luis ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
[Ocfs2-users] NFS Failover
Hi list, I need to implement a High available NFS server. Since we already have OCFS2 here for RAC, and already have a virtual IP on the RAC server that failovers automatically to the other node, it seems a natural choice to use it too for our NFS needs. We are using OCFS2 1.2. (Upgrade to 1.4 is not on our current plans) We did a preliminary failover test, and the client that mounts the filesystem (Actually a solaris box) doesnt like the failover. We expect some errors and minor data loss and can tolerate them as a transient condition, but the problem is that the mounted filesystem on the client becomes useless until we umount and remount it again. I suspect that NFS uses inode numbers on underlying filesystem to create handles that it passes on to clients, but I am not sure on how this is done. Anyone know if we can achieve a failover without needing to remount he nfs share on the clients? Any special options are needed mounting the OCFS2 filesystem and also for exporting it as NFS, or on the client? Best Regards, Luis ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Lost write in archive logs: has it ever happened?
Depending on your configuration, dataguard will transfer the modifications on the online log directly to the standby, so that the archived logs are recreated there. It doesnt transfer the archivedlogs from disk, instead it transfers the redo log entries directly to the other host, as they are generated, and the archivelogs are recreated there. So the remote copy is created independently from the local copy. The problem could be on the process writing the archivelog to disk or on the operating system/hardware. If it is a operating system or hardware issue, it probably is corrupting your datafiles as well. You should find fractured blocks or other corruptions indicating lost writes as well, like discrepancies between tables and indexes and rollback errors due to incorrect block scn. You can check this running a ANALYZE TABLE VALIDATE STRUCTURE CASCADE; job on all the tables. This command will also compare all the tables with the indexes, so it will effectively read all data blocks on the database and complain if it finds corrupted blocks or if the tables and indexes have discrepancies. Some types of tables cant be verified like this, so some errors indicating this are normal. (Btw this command locks the tables while it is running). Can you post the error that appears when applying the archivelog? Regards, Luis --- On Wed, 12/3/08, Silviu Marin-Caea [EMAIL PROTECTED] wrote: From: Silviu Marin-Caea [EMAIL PROTECTED] Subject: Re: [Ocfs2-users] Lost write in archive logs: has it ever happened? To: ocfs2-users@oss.oracle.com Date: Wednesday, December 3, 2008, 1:17 PM On Monday 22 September 2008 15:02:36 Silviu Marin-Caea wrote: We have 2 nodes with OCFS2 1.2.3 (SLES9). The archive logs are generated on an OCFS2 volume (mounted with nointr,datavolume). It has happened 3 times in one year that some archivelog had a lost write. We have detected this when applying the archivelogs on the standby database (with dataguard). We had to copy some datafiles from the production database to the standby and let it resume the recovery process. Has it ever occurred a data loss of this kind (lost write) on an OCFS2 volume, version 1.2.3 x86_64? We had 32 bit servers before with OCFS2 that was even older than 1.2.3 and those servers never had such a problem with archivelogs. The storage is Dell/EMC Clariion CX3-40. The storage on the old servers was CX300. We are worried that this lost writes could occur not only in archivelogs but in the datafiles as well... Not saying that OCFS2 is the cause, the problem might be with something else, but we must investigate everything. OCFS2 is not the cause. The error just occurred again and this time we had the archivelogs multiplexed on both OCFS2 and local storage (reiserfs). Both archives have identical MD5 sums. There was no lost write, just some bullshit that Oracle support tries to feed us. There is still an unknown, but it's not related to OCFS2. The unknown is that archives on the standby database have different MD5 sums than the ones on production. All the archives, not just the corrupt one. Does dataguard intervenes in some way in the archives during transmission? I thought it was just supposed to transfer them unchanged, then apply them on the standby. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Please urgent help required - OCFS2 and VPN again
Lorenzo, My 2 cents. This is purely speculation, since I never worked with a environment like what you have there. You have a very different configuration from what is usual with OCFS2. The filesystem is tested on systems that have a fast network connection between nodes, so it is probably not tuned to environments where the network bandwith is low. You might get some improvement if you change some of the VM tunables (/proc/sys/vm/*). On 2.6 there are not much of them for the filesystem cache, and some seem to have no effect. vfs_cache_pressure could give you some control on the quantity of inodes cached by the kernel. Try either increasing or decreasing. Some of the problems you relate, like the long time for umount might actually be caused by keeping a large amount of structures on memory, and since the network is slow a long time is needed to clear all of them. Swappiness controls how aggressivelly pages are swapped out. Since you dont have swap on the OCFS2 filesystem it should not have much impact. (You dont have swap there, right?) But you may be able to force the kernel to release memory so that it can be used by the OCFS2. Again this could actually cause the problem to become worse. There used to be a parameter that controls how much of the cache is used for inodes, and how much of it is used for data blocks, dcache_priorty. But it no longer available on 2.6 and I could not find an equivalent. Regards, Luis --- On Tue, 12/2/08, Lorenzo Milesi [EMAIL PROTECTED] wrote: From: Lorenzo Milesi [EMAIL PROTECTED] Subject: [Ocfs2-users] Please urgent help required - OCFS2 and VPN again To: ocfs2-users@oss.oracle.com Date: Tuesday, December 2, 2008, 8:59 AM Hi all... I already wrote before on the list about the solution I have at a customer running DRBD8+OCFS2 on two remote sites connected via VPN. The different suggestions helped improving the situation, but still we're having big troubles. We've also upgraded the old server with a new and much more powerful one but there was nearly no improvement at all! The situation is resumed as: SITE A: Dual Core 2GHz Pentium, 1Gb ram, 1 SATA hdd for /, 3 SATA hdd in software raid5, DRBD on /dev/md0. SITE B: Quad Core 2.4HGz Pentium, 2Gb ram, 3 SATA HDD in software raid5, DRBD on /dev/md1. The two sites are connected using two ADSL, with TWO bonded VPN. Both machines run Debian Etch fully updated, kernel 2.6.26-bpo.1-686 SMP with deadline scheduler, DRBD 8.0.13, OCFS2 1.4.1-1. The shared data partition is 187G, 30 of which used. The recent upgrade to OCFS2 1.4 and kernel 2.6.26 didn't improve the performances as much as I expected. The main problems we have are: 1. very high load average: this was previously caused by very high iowait percentages, but with the new server the load is high while top says the machine is 99-100% idle! 2. very slow dir browsing: Sunil pointed me to the user guide, where he talks about inode stat. How can I raise inode cache memory? I've done several searches without result... The server actually uses less than 300Mb of ram out of the 1Gb installed... 3. very long umount time: I often (not always) experience an extremely long umount time. During the period while the process is executing iftop says there's a high usage of network transfer. I suppose it's transfering file locks, but is it possible that stays stuck for more than one hour, and still going? This is the configuration file of OCFS2. The quad-core is file-server-2. #/etc/ocfs2/cluster.conf node: ip_port = ip_address = 192.168.0.1 number = 0 name = file-server-1 cluster = ocfs2 node: ip_port = ip_address = 192.168.2.31 number = 1 name = file-server-2 cluster = ocfs2 cluster: node_count = 2 name = ocfs2 What is stunning me is that on file-server-2 we run a rsync backup during the night on a local machine on the network, and it takes less than 20m! Doing the same on the other server throws the load average to the stars! We're in a critical situation because this solution is deployed since a long time and it's not yet working as expected. If nobody has suggestion we have no problem in paying qualified support for solving these problems. In this case please contact me directly. Sunil, can I get Oracle support for this? Thank you. -- Lorenzo Milesi - [EMAIL PROTECTED] YetOpen S.r.l. - http://www.yetopen.it/ C.so E. Filiberto, 74 23900 Lecco - ITALY - Tel 0341 220 205 - Fax 178 607 8199 GPG/PGP Key-Id: 0xE704E230 - http://keyserver.linux.it D.Lgs. 196/2003 Si avverte che tutte le informazioni contenute in questo messaggio sono riservate ed a uso esclusivo del destinatario. Nel caso in cui questo messaggio Le fosse pervenuto per errore, La invitiamo ad eliminarlo senza copiarlo, a non inoltrarlo a terzi e ad avvertirci non appena possibile. Grazie. ___
Re: [Ocfs2-users] New node..new problems
Dante, Your old debian is running OCFS 1.4 and your new Centos is running OCFS 1.2, right? If you are running Centos 5.0 you should be able to install OCFS 1.4. If not you will need to umount your debian before mounting the Centos. Beware that there are functionalities on OCFS 1.4 that are not available on 1.2, that might impact your applications. Also I am not sure if the disk layout is fully compatible if certain OCFS 1.4 filesystem options were enabled on your old cluster. The best option would be to upgrade to OCFS 1.4 on the Centos cluster. Regards, Luis --- On Fri, 10/10/08, Dante Garro [EMAIL PROTECTED] wrote: From: Dante Garro [EMAIL PROTECTED] Subject: Re: [Ocfs2-users] New node..new problems To: 'Tao Ma' [EMAIL PROTECTED] Cc: 'ocfs2-users@oss.oracle.com' ocfs2-users@oss.oracle.com Date: Friday, October 10, 2008, 9:05 AM Thanks Tao, I've setup the same on both nodes and the cluster becomes online. Now, when I try to mount the following errors appears on node 1 (new CentOS): (2512,1):o2net_connect_expired:1585 ERROR: no connection established with node 0 after 30.0 seconds, giving up and returning errors. (3022,1):dlm_request_join:901 ERROR: status = -107 (3022,1):dlm_try_to_join_domain:1049 ERROR: status = -107 (3022,1):dlm_join_domain:1321 ERROR: status = -107 (3022,1):dlm_register_domain:1514 ERROR: status = -107 (3022,1):ocfs2_dlm_init:2024 ERROR: status = -107 (3022,1):ocfs2_mount_volume:1133 ERROR: status = -107 ocfs2: Unmounting device (147,0) on (node 1) And the following on node 0 (old Debian) (2228,0):o2net_check_handshake:1093 node nodo2 (num 1) at 192.168.168.2: advertised net protocol version 103 but 2 is required, disconnecting I believe the Debian message is clear, protocol version incompatibility. Are there a way to resolve it? Thanks Dante -Mensaje original- De: Tao Ma [mailto:[EMAIL PROTECTED] Enviado el: viernes, 10 de octubre de 2008 1:25 Para: Dante Garro CC: 'Sunil Mushran'; 'ocfs2-users@oss.oracle.com' Asunto: Re: [Ocfs2-users] New node..new problems Hi, Dante Garro wrote: Sunil, now I fall in count of messages are related to node 0, but the new is node 1 and does not care about the value I've setup allways says 14000 ms. Do this change your diagnostic? Node1 start connection with node0, so you see the messages related to node0 on node1. It looks like your configuration in node1 is wrong. Please make sure that value of O2CB_HEARTBEAT_THRESHOLD in /etc/sysconfig/o2cb of node1 is the same as that in node0. Regards, Tao -Mensaje original- De: Sunil Mushran [mailto:[EMAIL PROTECTED] Enviado el: Jueves, 09 de Octubre de 2008 06:02 p.m. Para: Dante Garro CC: 'ocfs2-users@oss.oracle.com' Asunto: Re: [Ocfs2-users] New node..new problems Yeah the cluster timeouts are not consistent. Update and restart the cluster on the new node (or all nodes as the case might be). Hint: cat /sys/kernel/config/cluster/clustername/idle_timeout_ms to see the active heartbeat threshold. Dante Garro wrote: Hi all, because problems with ocfs2 release of Debian distribution decided to remake my cluster replacing it by CentOS based installation. Started replacing one of the nodes keeping the other working. On this recently created node the following errors appears: drbd0: Writing meta data super block now. (2558,1):o2hb_check_slot:881 ERROR: Node 0 on device drbd0 has a dead count of 14000 ms, but our count is 13000 ms. Please double check your configuration values for 'O2CB_HEARTBEAT_THRESHOLD' OCFS2 1.2.9 Wed Sep 24 19:26:41 PDT 2008 (build a693806cb619dd7f225004092b675ede) (2520,1):o2net_connect_expired:1585 ERROR: no connection established with node 0 after 30.0 seconds, giving up and returning errors. (2556,1):dlm_request_join:901 ERROR: status = -107 (2556,1):dlm_try_to_join_domain:1049 ERROR: status = -107 (2556,1):dlm_join_domain:1321 ERROR: status = -107 (2556,1):dlm_register_domain:1514 ERROR: status = -107 (2556,1):ocfs2_dlm_init:2024 ERROR: status = -107 (2556,1):ocfs2_mount_volume:1133 ERROR: status = -107 ocfs2: Unmounting device (147,0) on (node 1) (2591,1):o2hb_check_slot:881 ERROR: Node 0 on device drbd0 has a dead count of 14000 ms, but our count is 13000 ms. Please double check your configuration values for 'O2CB_HEARTBEAT_THRESHOLD' (2520,1):o2net_connect_expired:1585 ERROR: no connection established with node 0 after 30.0 seconds, giving up and returning errors. (2589,1):dlm_request_join:901 ERROR: status = -107 (2589,1):dlm_try_to_join_domain:1049 ERROR: status = -107 (2589,1):dlm_join_domain:1321 ERROR: status = -107 (2589,1):dlm_register_domain:1514 ERROR: status = -107 (2589,1):ocfs2_dlm_init:2024 ERROR: status = -107 (2589,1):ocfs2_mount_volume:1133 ERROR: status = -107 ocfs2: Unmounting device (147,0) on (node 1) I've changed the parameter O2CB_HEARTBEAT_THRESHOLD according O2CB adviced me, but It don't resolve the issue. I hope someone could
Re: [Ocfs2-users] Lost write in archive logs: has it ever happened?
Silviu, When I had this kind of issues it usually was caused by a bad hba, or a power failure. I am assuming it is not the latter as you would be aware of it. It is a difficult situation, since the controller only malfunctions sporadically it is difficult to prove that it is the cause or to get it changed on warranty. And your database slowly gets corrupted, until someday it crashes and wont startup. If this is the cause it surely is happening on the datafiles also. To be safe you should run a ANALYZE TABLE ... VALIDATE STRUCTURE CASCADE; on all your database tables, and look for fractured or bad blocks on the datafiles using dbv or rman. A fractured block is one that has a different timestamp on the begin and the end, so it was only partially writen to the disk. You also could try to change the hba with some other server to see if the problem disappears. Regards, Luis --- On Mon, 9/22/08, Silviu Marin-Caea [EMAIL PROTECTED] wrote: From: Silviu Marin-Caea [EMAIL PROTECTED] Subject: [Ocfs2-users] Lost write in archive logs: has it ever happened? To: ocfs2-users@oss.oracle.com Date: Monday, September 22, 2008, 9:02 AM We have 2 nodes with OCFS2 1.2.3 (SLES9). The archive logs are generated on an OCFS2 volume (mounted with nointr,datavolume). It has happened 3 times in one year that some archivelog had a lost write. We have detected this when applying the archivelogs on the standby database (with dataguard). We had to copy some datafiles from the production database to the standby and let it resume the recovery process. Has it ever occurred a data loss of this kind (lost write) on an OCFS2 volume, version 1.2.3 x86_64? We had 32 bit servers before with OCFS2 that was even older than 1.2.3 and those servers never had such a problem with archivelogs. The storage is Dell/EMC Clariion CX3-40. The storage on the old servers was CX300. We are worried that this lost writes could occur not only in archivelogs but in the datafiles as well... Not saying that OCFS2 is the cause, the problem might be with something else, but we must investigate everything. Thank you ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Lost write in archive logs: has it ever happened?
Silviu, Just so you be warned, the ANALYZE TABLE... command locks the tables during its execution. Regards, Luis --- On Mon, 9/22/08, Luis Freitas [EMAIL PROTECTED] wrote: From: Luis Freitas [EMAIL PROTECTED] Subject: Re: [Ocfs2-users] Lost write in archive logs: has it ever happened? To: ocfs2-users@oss.oracle.com, Silviu Marin-Caea [EMAIL PROTECTED] Date: Monday, September 22, 2008, 12:54 PM Silviu, When I had this kind of issues it usually was caused by a bad hba, or a power failure. I am assuming it is not the latter as you would be aware of it. It is a difficult situation, since the controller only malfunctions sporadically it is difficult to prove that it is the cause or to get it changed on warranty. And your database slowly gets corrupted, until someday it crashes and wont startup. If this is the cause it surely is happening on the datafiles also. To be safe you should run a ANALYZE TABLE ... VALIDATE STRUCTURE CASCADE; on all your database tables, and look for fractured or bad blocks on the datafiles using dbv or rman. A fractured block is one that has a different timestamp on the begin and the end, so it was only partially writen to the disk. You also could try to change the hba with some other server to see if the problem disappears. Regards, Luis --- On Mon, 9/22/08, Silviu Marin-Caea [EMAIL PROTECTED] wrote: From: Silviu Marin-Caea [EMAIL PROTECTED] Subject: [Ocfs2-users] Lost write in archive logs: has it ever happened? To: ocfs2-users@oss.oracle.com Date: Monday, September 22, 2008, 9:02 AM We have 2 nodes with OCFS2 1.2.3 (SLES9). The archive logs are generated on an OCFS2 volume (mounted with nointr,datavolume). It has happened 3 times in one year that some archivelog had a lost write. We have detected this when applying the archivelogs on the standby database (with dataguard). We had to copy some datafiles from the production database to the standby and let it resume the recovery process. Has it ever occurred a data loss of this kind (lost write) on an OCFS2 volume, version 1.2.3 x86_64? We had 32 bit servers before with OCFS2 that was even older than 1.2.3 and those servers never had such a problem with archivelogs. The storage is Dell/EMC Clariion CX3-40. The storage on the old servers was CX300. We are worried that this lost writes could occur not only in archivelogs but in the datafiles as well... Not saying that OCFS2 is the cause, the problem might be with something else, but we must investigate everything. Thank you ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Problems building ocfs2 rpm on Fedora 9
Tina, The raw devices are being deprecated on Linux. Since you are using fedora instead of a enterprise distro these changes are already done. You can use disk devices directly with the 11g clusterware. Also it probably would work with the main tree OCFS2, that doesnt has the datavolume option, since this implies that the clusterware writes on these devices using O_DIRECT, but AFAIK this not tested on the main OCFS2 and since the clusterware uses some heuristics to turn on these options it could not work with OCFS2. If you have metalink access check note 401132.1. Beware that this is broken on 10gR1 and probably the initial 10gR2 patchsets. I didnt found references of patches to include the O_DIRECT on the 10g version of the CRS so I am not sure when this was fixed. Regards, Luis --- On Tue, 7/29/08, Tina Soles [EMAIL PROTECTED] wrote: From: Tina Soles [EMAIL PROTECTED] Subject: Re: [Ocfs2-users] Problems building ocfs2 rpm on Fedora 9 To: [EMAIL PROTECTED], Tao Ma [EMAIL PROTECTED] Cc: Randy Gustin [EMAIL PROTECTED], ocfs2-users@oss.oracle.com Date: Tuesday, July 29, 2008, 4:19 PM OK, another snag. Fedora 9 does not support RAW devices, so I can't configure the voting disk or OCR disk to be as such. Any suggestions? I think I'm up a creek here... -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Sunday, June 29, 2008 11:20 PM To: Tao Ma; Tina Soles Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] Problems building ocfs2 rpm on Fedora 9 datavolume mount option is only on ocfs2 for enterprise kernels. For most part, you shouldn't require it for using it as a datastore. Instead set init.ora param filesystemio_options to directio (or is it odirect). The only bit that won't work here is using ocfs2 for the voting disk and ocr... as datavolume is necessary to force enable odirect. But you can always use raw for that. Sunil ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Recommended block size for a mail environment
Funny, I could not find a option to specify the number of inodes or the inodes/bytes ratio on mkfs.ocfs2? Regards, Luis --- On Tue, 7/22/08, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: From: [EMAIL PROTECTED] [EMAIL PROTECTED] Subject: [Ocfs2-users] Recommended block size for a mail environment To: ocfs2-users@oss.oracle.com Date: Tuesday, July 22, 2008, 9:22 AM Hello all, I have a mail environment, with a situation seemed with a previous mail Different size with du and ls. My environment is: Debian Etch 4.0 Postfix 2.3.8-2 ocfs2-tools 1.2.1-1.3 kernel 2.6.18-4-amd64 My df -h shows me the follow: df -Th /dev/sdb1ocfs21,0T 894G 131G 88% /mails df -Thi File Syst. Type Inodes IUsed IFree IUse% Mount on /dev/sdb1ocfs2 16M 14M2,1M 88% /mails but when I do a du -sh . /mails it shows me 480GB, so much different from 849G. The OCFS2 block size is 4K. I made a script to average the medium mail size, and it's the follow: 0k to 1k - 8% 0k to 2k - 13% 0k to 3k - 22% 0k to 4k - 35% 4.1k to 6k - 15% 6.1 to 8k - 10% 8.1k to 10k - 9% 10.1k to 12k - 7% rest - 24% My debug.ocfs2 doesn't has the -R option: # debugfs.ocfs2 -R stats /dev/sdb1 debugfs.ocfs2: invalid option -- R debugfs.ocfs2 1.2.1 My doubts are: 4k is a good block size for this situation? How to avoid waste space as now (~50%)? If I choose to use 2k, will the available Inodes double (16M to 32M)? Thanks a lot! Jeronimo Universidade Federal da Bahia - http://www.portal.ufba.br ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] ocfs2 datavolume option and oracle
Jamin, If you are using 10g you can create two 170Mb partitions for the OCR and three 20Mb partitions for the voting disk, and you can leave the rest on a single partition for the OCFS2 filesystem. I never installed 11g RAC, but the manual says that each partition must be 280Mb or larger, so you will need at least two 280Mb partitions for one OCR and one voting disk, preferrably five, for two OCR copies and three voting disks. Regards, Luis --- On Tue, 7/8/08, Sunil Mushran [EMAIL PROTECTED] wrote: From: Sunil Mushran [EMAIL PROTECTED] Subject: Re: [Ocfs2-users] ocfs2 datavolume option and oracle To: [EMAIL PROTECTED] [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com Date: Tuesday, July 8, 2008, 12:18 PM [EMAIL PROTECTED] wrote: By raw are you meaning raw device access without a filesystem like ocfs2 on the volume for the voting disk? Or am I not following? Raw means specifying the block device directly. So make two partitions, say, sdd1 and sdd2, and feed that (/dev/sdd1, etc) to the tool that creates the voting disk and ocr. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
[Ocfs2-users] ocfs2 limits
Sunil, This applies only to subdirectories or also to the number of files inside a directory? Oracle Applications can easily have hundreds of thousands files inside the concurrent output and log directories. Regards, Luis --- On Thu, 7/3/08, Sunil Mushran [EMAIL PROTECTED] wrote: From: Sunil Mushran [EMAIL PROTECTED] Subject: Re: [Ocfs2-users] ocfs2 fencing problem To: Gabriele Di Giambelardini [EMAIL PROTECTED] Cc: Kuang, Howard [WHQKT] [EMAIL PROTECTED], ocfs2-users@oss.oracle.com Date: Thursday, July 3, 2008, 1:54 PM Gabriele Di Giambelardini wrote: Hi to all, some time ago, I read that the ocfs have a limit for the subfolder. Is it possible this whren this limit gone exceeded the ocfs2 have those problem??? Or some boby know the limit number? http://oss.oracle.com/projects/ocfs2/dist/documentation/ocfs2_faq.html#LIMITS ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] RAC Shared Disk..?
Abhishek, Depends on what you intend to use OCFS2 for. Since you want to test Oracle RAC, the best options would be to go with a iSCSI setup or use a firewire shared disk, to get a configuration close to what is supported by Oracle . For both options you will need a third computer to "simulate" the storage, and for the firewire also the firewire adapters for the three computers. For the iSCSI you can use "OpenFiler", which has a browser based interface to create the iSCSI volume, and it can run on the same ethernet where you run the cluster interconnect, or on a third separate adapter on each server. Nowadays iSCSI seems to be prefered over firewire, the Oracle people here may have a better indication on which to use. If you only intended to run other services, like Apache you could try with DRDB. But for installing RAC I would not recommend it as would be very different from what a RAC production environment should look like. Also you can simulate the two nodes and the "shared storage" on a single machine using VMWare server. There are some howtos on this subject on www.oracle-base.com (But none for the particular setup you want to achieve).Regards,Luis--- On Mon, 6/16/08, Abhishek Sahu [EMAIL PROTECTED] wrote:From: Abhishek Sahu [EMAIL PROTECTED]Subject: [Ocfs2-users] RAC Shared Disk..?To: Ocfs2-users@oss.oracle.comDate: Monday, June 16, 2008, 9:06 AM Dear All, I am preparing a two node cluster on simple LINUX desktop systems. Both systems have 80 GB of hard disk; I don’t have any additional storage. Now I want to know the procedure for making a shared disk space out or these two available hard disks. Can any body help me out in this experiment? Also if any body can send me a detailed installation document for RAC on Linux (other than given on oracle website), then it will be a great help for me. Thanks in Advance, Abhishek ___Ocfs2-users mailing listOcfs2-users@oss.oracle.comhttp://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] CRS/CSS and OCFS2
Martin, Sunil and Mark are Oracle employees and involved on the development of OCFS2, I am just a user, :-).Regards,Luis--- On Fri, 6/6/08, Schmitter, Martin [EMAIL PROTECTED] wrote:From: Schmitter, Martin [EMAIL PROTECTED]Subject: AW: [Ocfs2-users] CRS/CSS and OCFS2To: "[EMAIL PROTECTED]" [EMAIL PROTECTED], "[EMAIL PROTECTED]" [EMAIL PROTECTED]Cc: "ocfs2-users@oss.oracle.com" ocfs2-users@oss.oracle.comDate: Friday, June 6, 2008, 4:31 AM Hi Luis, now I am a bit confused, because I asked a few months ago this question. Howmust be the timing setting from OCFS2 and CRS? Sunil and Mark stated that OCFS2 must be theleading system! If I get it right, the SAN fail over comes first, then OCFS2 und least but not last CRS. BR Martin Schmitter -- OPITZ CONSULTING Gummersbach GmbH Martin Schmitter - Fachinformatiker Kirchstr. 6 - 51647 Gummersbach Telefon: +49 2261 6001-0 Mobil: +49 173 2808193 http://www.opitz-consulting.de Geschäftsführer: Bernhard Opitz, Dr. Jürgen Abel, Ulrich Kramer HRB-Nr.39163 Amtsgericht Köln Von: [EMAIL PROTECTED] [EMAIL PROTECTED] im Auftrag von Luis Freitas [EMAIL PROTECTED] Gesendet: Donnerstag, 5. Juni 2008 18:32 An: [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com Betreff: Re: [Ocfs2-users] CRS/CSS and OCFS2 Alexandra, I usually make sure that one of the timeouts is large enought so that the other node death is detected before the other node "self-fence". To solve the problem you could configure the OCFS timeouts to be larger than the CRS timeouts, so that the CRS fences the node and OCFS detects the other node as dead before it takes any action. Maybe Sunil has a better solution that I am not aware of. This is particular of OCFS2 and CRS, which is kind of funny since both are developed by Oracle. With vendor clusterware (Sun cluster, Veritas, etc) CRS is integrated with the vendor clusterware stack so that this kind of situation does not occur. Btw, CRS is kind of picky about its interfaces, if it detects a link down on the interface, it will shutdown the services on the node. This is why I asked about the crossover cable, when using a crossover cable and one node goes down, the inteface goes down on the other node and things does not work as expected. Regards, Luis --- On Thu, 6/5/08, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: From: [EMAIL PROTECTED] [EMAIL PROTECTED] Subject: [Ocfs2-users] CRS/CSS and OCFS2 To: ocfs2-users@oss.oracle.com Date: Thursday, June 5, 2008, 12:38 PM Hi Sunil, sorry for the delay but I was ill the last 10 days. a. We do not use a crossover cable between the two nodes. The two systems are seated in two SANs in different building with redundant switches and HBA's inbetween. b.ocfs2-node numbers: [EMAIL PROTECTED] etc]$ cat /etc/ocfs2/cluster.conf node: ip_port = ip_address = 10.190.59.5 number = 0 name = byaz05.bayer-ag.com cluster = ocfs2 node: ip_port = ip_address = 10.190.59.6 number = 1 name = byaz10.bayer-ag.com cluster = ocfs2 cluster: node_count = 2 name = ocfs2 Clusterconfiguration css/crs: /u01/app/oracle/product/crs/log/byaz05/crsd 2008-04-25 09:29:01.855: [ OCRMAS][1210108256]th_master:12: I AM THE NEW OCR MASTER at incar 1. Node Number 2 2008-04-25 09:29:01.862: [ OCRRAW][1210108256]proprioo: for disk 0 (/dev/raw/raw101), id match (1), my id set (1723799148,1710759834) total id sets (1), 1st set (1723799148,1710759834), 2nd set (0,0) my votes (1), total votes (2) 2008-04-25 09:29:01.862: [ OCRRAW][1210108256]proprioo: for disk 1 (/dev/raw/raw201), id match (1), my id set (1723799148,1710759834) total id sets (1), 1st set (1723799148,1710759834), 2nd set (0,0) my votes (1), total votes (2) /u01/app/oracle/product/crs/log/byaz10/crsd 2008-04-25 10:21:28.781: [ OCRMAS][1210108256]th_master:13: I AM THE NEW OCR MASTER at incar 4. Node Number 1 2008-04-25 10:21:28.781: [ OCRMSG][1505941856]prom_rpc:1: NULL con. Probably got disconnected due to a remote server failure. 2008-04-25 10:21:29.324: [ OCRRAW][1210108256]proprioo: for disk 0 (/dev/raw/raw101), id match (1), my id set (1723799148,1710759834) total id sets (1), 1st set (1723799148,1710759834), 2nd set (0,0) my votes (1), total votes (2) 2008-04-25 10:21:29.324: [ OCRRAW][1210108256]proprioo: for disk 1 (/dev/raw/raw201), id match (1), my id set (1723799148,1710759834) total id sets (1), 1st set (1723799148,1710759834), 2nd set (0,0) my votes (1), total votes (2) 2008-04-25 10:21:29.351: [ OCRMAS][1210108256]th_master: Deleted ver keys from cache (master) So the two nodes have the following nodenumbers: Fencing the node with the higher node number ocfs2 would have fenced byaz10 and crs/css would have fenced byaz05. This is exactly the behaviour we watched. But how to solve this? Oracle says it's certified to use ocfs2 with RAC. Then the used software combination is nearly the same as we us
Re: [Ocfs2-users] CRS/CSS and OCFS2
Alexandra, I usually make sure that one of the timeouts is large enought so that the other node death is detected before the other node "self-fence". To solve the problem you could configure the OCFS timeouts to be larger than the CRS timeouts, so that the CRS fences the node and OCFS detects the other node as dead before it takes any action. Maybe Sunil has a better solution that I am not aware of. This is particular of OCFS2 and CRS, which is kind of funny since both are developed by Oracle. With vendor clusterware (Sun cluster, Veritas, etc) CRS is integrated with the vendor clusterware stack so that this kind of situation does not occur. Btw, CRS is kind of picky about its interfaces, if it detects a link down on the interface, it will shutdown the services on the node. This is why I asked about the crossover cable, when using a crossover cable and one node goes down, the inteface goes down on the other node and things does not work as expected. Regards,Luis--- On Thu, 6/5/08, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:From: [EMAIL PROTECTED] [EMAIL PROTECTED]Subject: [Ocfs2-users] CRS/CSS and OCFS2To: ocfs2-users@oss.oracle.comDate: Thursday, June 5, 2008, 12:38 PM Hi Sunil, sorry for the delay but I was ill the last 10 days. a. We do not use a crossover cable between the two nodes. The two systems are seated in two SANs in different building with redundant switches and HBA's inbetween. b.ocfs2-node numbers: [EMAIL PROTECTED] etc]$ cat /etc/ocfs2/cluster.conf node: ip_port = ip_address = 10.190.59.5 number = 0 name = byaz05.bayer-ag.com cluster = ocfs2 node: ip_port = ip_address = 10.190.59.6 number = 1 name = byaz10.bayer-ag.com cluster = ocfs2 cluster: node_count = 2 name = ocfs2 Clusterconfiguration css/crs: /u01/app/oracle/product/crs/log/byaz05/crsd 2008-04-25 09:29:01.855: [ OCRMAS][1210108256]th_master:12: I AM THE NEW OCR MASTER at incar 1. Node Number 2 2008-04-25 09:29:01.862: [ OCRRAW][1210108256]proprioo: for disk 0 (/dev/raw/raw101), id match (1), my id set (1723799148,1710759834) total id sets (1), 1st set (1723799148,1710759834), 2nd set (0,0) my votes (1), total votes (2) 2008-04-25 09:29:01.862: [ OCRRAW][1210108256]proprioo: for disk 1 (/dev/raw/raw201), id match (1), my id set (1723799148,1710759834) total id sets (1), 1st set (1723799148,1710759834), 2nd set (0,0) my votes (1), total votes (2) /u01/app/oracle/product/crs/log/byaz10/crsd 2008-04-25 10:21:28.781: [ OCRMAS][1210108256]th_master:13: I AM THE NEW OCR MASTER at incar 4. Node Number 1 2008-04-25 10:21:28.781: [ OCRMSG][1505941856]prom_rpc:1: NULL con. Probably got disconnected due to a remote server failure. 2008-04-25 10:21:29.324: [ OCRRAW][1210108256]proprioo: for disk 0 (/dev/raw/raw101), id match (1), my id set (1723799148,1710759834) total id sets (1), 1st set (1723799148,1710759834), 2nd set (0,0) my votes (1), total votes (2) 2008-04-25 10:21:29.324: [ OCRRAW][1210108256]proprioo: for disk 1 (/dev/raw/raw201), id match (1), my id set (1723799148,1710759834) total id sets (1), 1st set (1723799148,1710759834), 2nd set (0,0) my votes (1), total votes (2) 2008-04-25 10:21:29.351: [ OCRMAS][1210108256]th_master: Deleted ver keys from cache (master) So the two nodes have the following nodenumbers: Fencing the node with the higher node number ocfs2 would have fenced byaz10 and crs/css would have fenced byaz05. This is exactly the behaviour we watched. But how to solve this? Oracle says it's certified to use ocfs2 with RAC. Then the used software combination is nearly the same as we use it. How can the combination of the two systems (ocfs2/css) fencing different nodes avoided then? Greets, Alex In such a situation, ocfs2 fences the higher node number. afaik, css does the same. What are the css node numbers for the two nodes? alexandra.strauss at bayerbbs.com wrote: Hello, I refer to you hoping you may help me with my problem... We have got an issur here and opened a SR at Metalink but until now, we got no useful information in solving our problem. SR-Number is 6855815.994... We wanted to protect 9i Single-Instance Databases with 10g Clusterware following the third-party-tool approach. There are no RAC-databases involved. But we want to achieve high availability as the databases are business critical systems. We want to make the systems able to relocate to another machine in case of failure to keep downtimes low... To achieve this we want to use OCFS2 for the filesystem. Relocate is done by script with help of CRS. So we took two systems (byaz05 and byaz10) and installed the following software: 10g CRS (10.2.0.3) and Oracle Software 9.2.0.8 and OCFS2 1.2.8 We found the following Metalinknotes and adjusted the heartbeat and timeouts for OCFS2: Metalink Note 395878.1: Heartbeat/Voting/Quorum Related Timeout Configuration for Linux, OCFS2, RAC Stack to avoid unnessary node fencing, panic and reboot Metalink Note
Re: [Ocfs2-users] huge something problem
Alexandre, nbsp; What are you running on the servers? (NFS Server? apache? Oracle?) Regards, Luis --- On Mon, 6/2/08, Alexandre Racine lt;[EMAIL PROTECTED]gt; wrote: From: Alexandre Racine lt;[EMAIL PROTECTED]gt; Subject: Re: [Ocfs2-users] huge something problem To: Sunil Mushran lt;[EMAIL PROTECTED]gt; Cc: ocfs2-users@oss.oracle.com Date: Monday, June 2, 2008, 11:17 AM Ok. I have the same problem now (load of server at 8.00, no processors higher then 5% of utilization, and a user can't access his folder). I did your commands, but there no real data here... [EMAIL PROTECTED] /mnt/data/testOCFS2 $ sudo ./scanlocks2.sh racinea@ srv2 /mnt/data/testOCFS2 $ w 10:16:44 up 9 days, 19:10, 2 users, load average: 8.43, 8.36, 8.14 USER TTYLOGIN@ IDLE JCPU PCPU WHAT racinea pts/1 10:130.00s 0.02s 0.00s w racinea@ srv2 /mnt/data/testOCFS2 $ sudo ./listdomains.sh 41535574BDEB4720B2CE7819A631DF10 /dev/sdd /home What else could I try? Thanks. Alexandre Racine [EMAIL PROTECTED] 514-461-1300 poste 3303 gt; -Original Message- gt; From: Sunil Mushran [mailto:[EMAIL PROTECTED] gt; Sent: 27 mai 2008 14:26 gt; To: Alexandre Racine gt; Cc: ocfs2-users@oss.oracle.com gt; Subject: Re: [Ocfs2-users] huge something problem gt; gt; Alexandre Racine wrote: gt; gt; Excellent, that works great! Now that I have the locks and the domain gt; gt; name what should I do to unlock them? (Or fix the problem). gt; gt; gt; The locking and unlocking is handled by the dlm. I'm working gt; on updating the wikis with more information on debugging such gt; issues. For the time being, ping me with the info. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] CRS/CSS and OCFS2
Alexandra, nbsp;nbsp; You could use only CRS and ext3 instead of ocfs2 for this kind of use. You would need to register a script to force umount the filesystem on the primary node and mount it on the node you are failing over to, it would be nice to be able to check if the filesystem is mounted before atempting to mount it, but I am not sure on how to do this) nbsp;nbsp; Are you using a cross-over cable for the private interconnect? Regards, Luis --- On Fri, 6/27/08, [EMAIL PROTECTED] lt;[EMAIL PROTECTED]gt; wrote: From: [EMAIL PROTECTED] lt;[EMAIL PROTECTED]gt; Subject: [Ocfs2-users] CRS/CSS and OCFS2 To: ocfs2-users@oss.oracle.com Date: Friday, June 27, 2008, 10:41 AM Hello, I refer to you hoping you may help me with my problem... We have got an issur here and opened a SR at Metalink but until now, we got no useful information in solving our problem. SR-Number is 6855815.994... We wanted to protect 9i Single-Instance Databases with 10g Clusterware following the third-party-tool approach. There are no RAC-databases involved. But we want to achieve high availability as the databases are business critical systems. We want to make the systems able to relocate to another machine in case of failure to keep downtimes low... To achieve this we want to use OCFS2 for the filesystem. Relocate is done by script with help of CRS. So we took two systems (byaz05 and byaz10) and installed the following software: 10g CRS (10.2.0.3) and Oracle Software 9.2.0.8 and OCFS2 1.2.8 We found the following Metalinknotes and adjusted the heartbeat and timeouts for OCFS2: Metalink Note 395878.1: Heartbeat/Voting/Quorum Related Timeout Configuration for Linux, OCFS2, RAC Stack to avoid unnessary node fencing, panic and reboot Metalink Note 391771.1: OCFS2 - FREQUENTLY ASKED QUESTIONS (hier insbesondere der Abschnitt zu Fencing und Quorum) Metalink Note 434255.1: Common reasons for OCFS2 Kernel Panic or Reboot Issues Metalink Note 457423.1: OCFS2 Fencing, Network, and Disk Heartbeat Timeout Configuration We did no changes to the CRS/CSS default settings until now. During HA-testing we watched unexpected behaviour of the system. We deactivated the bond for private interconnect and expected only one node to go down. But we faced both nodes going down. As it seems to me one node was rebooted from OCFS2 and the other one from CRS/CSS. Timestamp nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; -- 10:21:06 nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp;bond1 disabled (eth1) nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; /var/log/messages byaz05 Apr 25 10:21:06 byaz05 kernel: bonding: bond1: link status definitely down for interface eth1, disabling it Apr 25 10:21:06 byaz05 kernel: bonding: bond1: making interface eth5 the new active one. 10:21:09 nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp;bond1 disabled (eth5) nbsp; nbsp; nbsp; nbsp; /var/log/messages byaz05 Apr 25 10:21:09 byaz05 kernel: bonding: bond1: link status definitely down for interface eth5, disabling it Apr 25 10:21:09 byaz05 kernel: bonding: bond1: now running without any active interface ! 10:21:23 nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp;o2net – no longer connected nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; /var/log/messages byaz05 Apr 25 10:21:23 byaz05 kernel: o2net: no longer connected to node byaz10.bayer-ag.com (num 1) at 10.190.59.6: /var/log/messages byaz10 Apr 25 10:21:23 byaz10 kernel: o2net: no longer connected to node byaz05.bayer-ag.com (num 0) at 10.190.59.5: 10:21:27 nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp;CSSD failure 134 10:21:29 nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp;Reboot initiated by CRS /var/log/messages byaz05 Apr 25 10:21:27 byaz05 logger: Oracle clsomon failed with fatal status 12. Apr 25 10:21:27 byaz05 logger: Oracle CSSD failure 134. Apr 25 10:21:27 byaz05 su(pam_unix)[25839]: session closed for user oracle Apr 25 10:21:27 byaz05 logger: Oracle CRS failure. nbsp;Rebooting for cluster integrity. Apr 25 10:21:27 byaz05 kernel: md: stopping all md devices. Apr 25 10:21:27 byaz05 kernel: md: md0 switched to read-only mode. Apr 25 10:21:29 byaz05 logger: Oracle CRS failure. nbsp;Rebooting for cluster integrity. Apr 25 10:21:29 byaz05 kernel: e1000: eth2: e1000_watchdog_task: NIC Link is Up 1000 Mbps Full Duplex Apr 25 10:21:29 byaz05 logger: Oracle init script ceding reboot to sibling 27383. 10:21:58 nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp; nbsp;Reboot initiated by OCFS2(?) /var/log/messages byaz10 Apr 25 10:21:58 byaz10 su(pam_unix)[4595]: session opened for user oracle by (uid=0) Apr 25 10:21:58 byaz10 su(pam_unix)[4595]: session closed for user oracle Apr 25 10:25:58 byaz10 syslogd 1.4.1: restart. Apr 25 10:25:58 byaz10 syslog: syslogd startup succeeded Apr 25 10:25:58 byaz10 kernel: klogd 1.4.1, log
Re: [Ocfs2-users] CRS/CSS and OCFS2
Hmm, nbsp; There is a lazy umount: nbsp;nbsp;nbsp;nbsp;nbsp;nbsp; -lnbsp;nbsp;nbsp;nbsp; Lazy unmount. Detach the filesystem from the filesystem hierar- nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp; chynbsp; now,nbsp; and cleanup all references to the filesystem as soon nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp; as it is not busy anymore. This option allows a âbusyâ filesys- nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp;nbsp; tem to be unmounted.nbsp; (Requires kernel 2.4.11 or later.) nbsp; Not sure if this prevents writes on the filesystem after the umount completes, or if there is some way to fence the device, so these problems would need to be verified. nbsp; This is the way any other cold failover HA solution works, they dont use cluster filesystems. If there is no way to force the filesystem to be umounted or fence the device, it should be possible to configure CRS to evict the primary node before the filesystem is mounted on the secondary node. (Or use some external fence device, like a SAN switch or power appliance). nbsp; Btw, is OCFS2 officially supported with Oracle 9i single instance databases? Regards, Luis --- On Tue, 5/27/08, Sunil Mushran lt;[EMAIL PROTECTED]gt; wrote: From: Sunil Mushran lt;[EMAIL PROTECTED]gt; Subject: Re: [Ocfs2-users] CRS/CSS and OCFS2 To: [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com, [EMAIL PROTECTED] Date: Tuesday, May 27, 2008, 5:07 PM AFAIK: a. There is no force umount in Linux. b. There is no way to know whether a local fs is mounted on another node. Luis Freitas wrote: gt; Alexandra, gt; gt;You could use only CRS and ext3 instead of ocfs2 for this kind of gt; use. You would need to register a script to force umount the gt; filesystem on the primary node and mount it on the node you are gt; failing over to, it would be nice to be able to check if the gt; filesystem is mounted before atempting to mount it, but I am not sure gt; on how to do this) gt; gt;Are you using a cross-over cable for the private interconnect? gt; gt; Regards, gt; Luis gt; gt; --- On *Fri, 6/27/08, [EMAIL PROTECTED] gt; /lt;[EMAIL PROTECTED]gt;/* wrote: gt; gt; From: [EMAIL PROTECTED] lt;[EMAIL PROTECTED]gt; gt; Subject: [Ocfs2-users] CRS/CSS and OCFS2 gt; To: ocfs2-users@oss.oracle.com gt; Date: Friday, June 27, 2008, 10:41 AM gt; gt; gt; Hello, gt; gt; I refer to you hoping you may help me with my problem... We have gt; got an issur here and opened a SR at Metalink but until now, we gt; got no useful information in solving our problem. SR-Number is gt; 6855815.994... gt; gt; We wanted to protect 9i Single-Instance Databases with 10g gt; Clusterware following the third-party-tool approach. There are no gt; RAC-databases involved. But we want to achieve high availability gt; as the databases are business critical systems. We want to make gt; the systems able to gt; relocate to another machine in case of failure to keep downtimes gt; low... To achieve this we want to use OCFS2 for the filesystem. gt; Relocate is done by script with help of CRS. gt; gt; So we took two systems (byaz05 and byaz10) and installed the gt; following software: 10g CRS (10.2.0.3) and Oracle Software 9.2.0.8 gt; and OCFS2 1.2.8 gt; gt; We found the following Metalinknotes and adjusted the heartbeat gt; and timeouts for OCFS2: Metalink Note 395878.1: gt; Heartbeat/Voting/Quorum Related Timeout Configuration for Linux, gt; OCFS2, RAC Stack to avoid unnessary node fencing, panic and reboot gt; Metalink Note 391771.1: OCFS2 - FREQUENTLY ASKED QUESTIONS (hier gt; insbesondere der Abschnitt zu Fencing und Quorum) gt; Metalink Note 434255.1: Common reasons for OCFS2 Kernel Panic or gt; Reboot Issues gt; Metalink Note 457423.1: OCFS2 Fencing, Network, and Disk Heartbeat gt; Timeout Configuration gt; gt; We did no changes to the CRS/CSS default settings until now. gt; gt; During HA-testing we watched unexpected behaviour of the system. gt; We deactivated the bond for private interconnect and expected only gt; one node to go down. But we faced both nodes going down. As it gt; seems to me one node was rebooted from OCFS2 and the other one gt; from CRS/CSS. gt; gt; Timestamp gt; -- gt; gt; 10:21:06bond1 disabled (eth1) gt; */var/log/messages byaz05* gt; Apr 25 10:21:06 byaz05 kernel: bonding: bond1: link status gt; definitely down for interface eth1, disabling it gt; Apr 25 10:21:06 byaz05 kernel: bonding: bond1: making interface gt; eth5 the new active one. gt; gt; 10:21:09bond1 disabled (eth5) gt; */var/log/messages byaz05* gt; Apr 25 10:21:09 byaz05 kernel: bonding
Re: [Ocfs2-users] Node reboot during network outage
If you expect one of the switches to remain alive you should configure your bonding driver timeout low to force it to failover soon to the other switch. What kind of bonding are you using? There are several different modes on Linux, for different types of switch configurations. I have used balance-tlb with good results, also active-backup could work too. The main problem I see is the time the switch takes to identify the different path. I had some problems with virtual IP failover due to the kernel not broadcasting something to the switch if forwarding is not enabled. Although virtual IPs are very different from network bonding you might want to try enabling ip_forward to see if the switches reconfigure faster. (echo 1 /proc/sys/net/ipv4/ip_forward). Some people here have asked the O2CB to allow for multiple heartbeat interfaces, it would provide an alternative to this problem, as the switches would not need to reconfigure.Regards,Luis--- On Tue, 4/22/08, Brendan Beveridge [EMAIL PROTECTED] wrote:From: Brendan Beveridge [EMAIL PROTECTED]Subject: Re: [Ocfs2-users] Node reboot during network outageTo: Cc: "ocfs2-users@oss.oracle.com" ocfs2-users@oss.oracle.comDate: Tuesday, April 22, 2008, 8:06 PMSTP Shouldnt have anything to do with the nodes still seeing each other when the switch failsHave you checked that your bonding config is correct?iecat /proc/net/bonding/bond0and check that it fails over to the other eth when the switch goes downCheersBrendanSunil Mushran wrote: The issue is not the time the switch takes to reboot. The issue is the amount of time the secondary switch takes to find a unique path. http://en.wikipedia.org/wiki/Spanning_tree_protocol Mick Waters wrote:Thanks Sunil, The network switch is brand new but has a fairly complex configurationdue to us running a number of VLANs - however, we have found that it has alwaystaken quite a while to reboot. I'll try increasing the idle timeout as suggested and let you knowwhat happens. However, surely this is only treating the symptoms of what is,after all, a contrived scenario. Rebooting the switch is supposed to test whatwould happen if we had a real network outage. What if the switch were to staydown? My issue is that we have an alternative route via the other NIC in thebond and the other switch. The affected nodes in cluster shouldn't fencebecause they should still be able to see all of the other nodes in the clustervia this other route. Does this make sense? Regards, Mick. -Original Message- From: Sunil Mushran [mailto:[EMAIL PROTECTED] Sent: 22 April 2008 17:40 To: Mick Waters Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] Node reboot during network outage The interface died at 14:25:44 and recovered at 14:27:43. That's two minutes. One solution is to increase o2cb_idle_timeout to 2mins. Better solution would be to look into your router setup to determinewhy it is taking 2 minutes for the router to reconfigure. Mick Waters wrote: Hi, my company is in the process of moving our web and database servers to new hardware. We have a HP EVA 4100 SAN which is being used by two database servers running in an Oracle 10g cluster andthat works fine. We have gone to extreme lengths to ensure high availability. The SAN has twin disk arrays, twin controllers, andall servers have dual fibre interfaces. Networking is (should be) similarly redundant with bonded NICs connected in two-switch configuration, two firewalls and so on. We also want to share regular Linux filesystems between ourservers - HP DL580 G5s running RedHat AS 5 (kernel 2.6.18-53.1.14.el5) andwe chose OCFS2 (1.2.8) to manage the cluster. As stated, each server in the 4 node cluster has a bondedinterface set up as bond0 in a two-switch configuration (each NIC in thebond is connected to a different switch). Because this is a two-switch configuration, we are running the bond in active-standby mode andthis works just fine. Our problem occurred when we were doing failover testing where we simulated the loss of one of the network switches by powering itoff. The result was that the servers rebooted and this make a mockeryof our attempts at a HA solution. Here is a short section from /var/log/messages following a rebootof one of the switches to simulate an outage:-- Apr 22 14:25:44 mtkws01p1 kernel: bonding: bond0: backup interface eth0 is now down Apr 22 14:25:44 mtkws01p1 kernel: bnx2: eth0 NIC Link is Down Apr 22 14:26:13 mtkws01p1 kernel: o2net: connection to node mtkdb01p2 (num 1) at 10.1.3.50: has beenidle for 30.0 seconds, shutting it down. Apr 22 14:26:13 mtkws01p1 kernel: (0,12):o2net_idle_timer:1426here are some times that might help debug the situation: (tmr 1208870743.673433 now 1208870773.673192 dr 1208870743.673427 adv 1208870743.673433:1208870743.673434 func (97690d75:2) 1208870697.670758:1208870697.670760) Apr 22 14:26:13 mtkws01p1 kernel: o2net: no longer
Re: [Ocfs2-users] AoE+ocfs2 = Heartbeat write timeout to device
Sunil, Can I configure this heartbeat to use a high priority (realtime) schedulling? If I simply increase the timeout it still could timeout on heavy I/O situations, like several different threads queuing large amounts of writes. The kernel should know this is a high priority write so that it is put ahead of the queue. Regards, Luis Sunil Mushran [EMAIL PROTECTED] wrote: The older 12 sec default timeout was too low. It has been bumped up to 60 secs. The FAQ has details on this. [EMAIL PROTECTED] wrote: Hi, I got a problem regarding 100Mbit Ethernet, AoE and ocfs2. I setup 2 boxes connected per 100Mbit ethernet to their Ata-over-Ethernet storage. The ocfs filesystem resides on such an AoE-Partition. If I produce high troughput to that ocfs-partition on one node, it reboots after some seconds. I use dd for testing, like dd if=/dev/zero of=test bs=1M count=1000 If I write 100Mb of data to the disk everything is fine. If I write 1Gb of data to the disk, the node reboots after some seconds and prints the following error: (9,0):o2hb_write_timeout:167 ERROR: Heartbeat write timeout to device etherd/e402.0 after 12000 milliseconds (9,0):o2hb_stop_all_regions:1865 ERROR: stopping heartbeat on all active regions. This couldn't be caused by lost heartbeat packets. I setup a seperate network for heartbeat to track this problem. Actually I know that 100Mbit Ethernet is a bottleneck, but this should not cause the system to reboot, right? Even if I could switch to Gigbit Ethernet it may be the bottleneck in future.. Someone experienced this already? Do you know how to solve this issue? Please help, I need to do some tests.. Your help is really appreciated. Cheers, Holger ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Never miss a thing. Make Yahoo your homepage.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Possibility for networkless installation (for long distance shared storage)
Michael, You might be able to route IP over your FC network. I know that some unixes do this, but I am not sure on how to do it on Linux. Regards, Luis Mark Fasheh [EMAIL PROTECTED] wrote: On Thu, Jan 03, 2008 at 08:47:37AM -0800, Michael M. wrote: As we have fiber channel, and we have fiber connections, sometimes direct p-to-p, where we could run it into our fiber channel switch, and share a disk array over hundreds of feet, to hundreds of miles, but where servers wouldnât be able to talk via fast enough network connection, does (or will) ocfs2 support a mode where there is no network connection needed, and all voting, and cluster âtalkingâ can be done to the block device itself, including the heartbeat? Ocfs (v1) used to do that, but we moved away from it for Ocfs2 because it quickly became unusably slow once all the additional locking traffic for general purpose usage was added. --Mark -- Mark Fasheh Principal Software Developer, Oracle [EMAIL PROTECTED] ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Never miss a thing. Make Yahoo your homepage.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Large difference between space used on reiserfs vs ocfs2
If you have lots of small files, it could happen. The basic space allocation unit is one block. If I am not mistaken on OCFS2 the default is 4k. A quick search on google shows that Reiserfs can coalesce small files together and save space. You could try to format your OCFS partition with a smaller block size, near the average file size. Other possible explanation for this (That is, without filesystem corruption) is a large file that was deleted but is still in use by some process. The space remains allocated until you kill the process. Regards, Luis Michael M. [EMAIL PROTECTED] wrote:Dear list, I have all of my web files for my apache servers on multiple machines placed on ocfs2 volumes. I recently did an rsync to a reiserfs volume on an external usb harddrive, and df reports the following: (ocfs2 is on top, reiserfs on bottom) /dev/sde5218757056 176814592 41942464 81% /mnt/www /dev/sdd1732549604 117493932 615055672 17% /mnt/pba-web Is this normal? Thanks, Michael ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Looking for last minute shopping deals? Find them fast with Yahoo! Search.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Fwd: Returned mail: see transcript for details
Anjan, Yes, it should be like that. Also see if you can follow Mark suggestion as he knows this stuff better than me. You could get a iSCSI server running with a third machine and openfiler and test to see if you get the same results. (Btw I never installed openfiler, last time I installed a test system like this I used NFS :-P, neither are supported of course) Regards, Luis Anjan Chakraborty [EMAIL PROTECTED] wrote: Luis, Let me understanfd your suggestion. Are you suggesting: $ORA_CRS_HOME -- without datavolume,nointr $ORACLE_HOME -- without datavolume,nointr But place OCR, Voting Disks future datafiles with datavolume,nointr Did I understand correctly? I have also gone back to OCFS2 1.2.5-6. Okay, if this is the suggestion -- I will surely try. Thanks. Anjan Luis Freitas [EMAIL PROTECTED] wrote: Anjan, The two installs I did with x86_64 were with ASM, with kernel versions that today are very old, some years ago. Here I am using x86 (even as the hardware is x86_64 capable) due to a managment decision not to go to x86_64. Anyone else on the list? About the mount points, the CRS home is a ORACLE_HOME, so you shouldnt mount it with datavolume. If you want to keep the CRS ORACLE_HOME and the RDBMS ORACLE_HOME each on one filesystem you should have at least a third partition for the cluster registry, voting disks and datafiles, mounted with the datavolume option, and mount those two without it. If you need to get this working fast, I think you should try a older OCFS2 module version. If not keep working with Mark and the other Oracle people here on the list as they can diagnose these errors and get it fixed. What is the firewire disk you are using? Regards, Luis Anjan Chakraborty [EMAIL PROTECTED] wrote: Luis Freitas [EMAIL PROTECTED] wrote: Luis, As far as I know, it's not recommended to mount ORACLE_HOME with datavolume,nointr but that's not the case with CRS. So, I have different mount options for ORA_CRS_HOME (home/oracle/ocrvotcrs) ORACLE_HOME (home/oracle/orasys). If I am wrong, please correct me. Also, I don't have any datafile yet because I don't have the basic thing -- Oracle binary under ORACLE_HOME. Once I have that, I will have datafiles mount will contain datavolume,nointr. Do you know anybody has successfully implemented OCFS2 1.2.7-1 on Red Hat4.5 (kernel 2.6.9-55.EL) under 86_64 architecture? What exactly you have? Thanks for your continuous effort to help me. I hope somebody else will also start joining this discussion help me. I am really stuck -- can't advance at all -- need help desperately. Anjan Anjan, Well it is only a sugestion. About those mount options, perhaps the Oracle people can give you a feedback on that also. But I noticed that you have two mount points, and seems to be installing CRS binaries on one, and RDBMS binaries on the other. So, where are datafiles? You shouldnt be installing the CRS home on /home/oracle/ocrvotcrs as it has the datavolume option. It is not clear if you are trying this. The filesystems where you have binaries should not have the datavolume option, and the filesystems with datafiles, voting disks and the cluster registry need to have the datavolume option. Regards, Luis Anjan Chakraborty [EMAIL PROTECTED] wrote:Mark, Thanks to both of you for trying to help me. I have alreaddy communicated to Luis that I have to have OCFS2 because my software needs Clustering technology that is absent in EXT3. Moreover, both CRS RDBMS homes should also be on shared/clustered system -- so, OCFS2 is the only choice. If you can help me to understand/resolve this issue, I will really appreciate that. Please note that my FireWire shared drive works perfectly when I use RAW but as soon as I am trying to use OCFS2, all the problems started. Thanks. Anjan Mark Fasheh [EMAIL PROTECTED] wrote: Luis, Thanks for helping here. I have one comment regarding moving the shared home to ext3. If Anjan's setup is having issues like this, moving his oracle home to a local disk would only hide problems which take longer to reproduce for the crs and data files partitions. What needs to happen is that the root cause of his problems is discovered and fixed. Keep in mind, I'm not saying someone couldn't use ext3 for their oracle install - it's an excellent choice for running a non-shared oracle home. It's just not a good shared disk diagnostic tool ;) Once again, thanks for stepping up and lending a hand. --Mark On Tue, Nov 27, 2007 at 08:50:28AM -0800, Luis Freitas wrote: Anjan, You dont need to share the database binaries, only the CRS and the datafiles. You can do it to save disk space, but it is not mandatory. The CRS and datafiles are much less stressfull to the filesystem structures as there is a reduced number of large files, although they usually have a heavy i/o load, and stress the disk subsystem and the locking
Re: [Ocfs2-users] OCFS2 on CentOS 4.5 for CRS/RAC
Anjan, You dont need to share the database binaries, only the CRS and the datafiles. You can do it to save disk space, but it is not mandatory. The CRS and datafiles are much less stressfull to the filesystem structures as there is a reduced number of large files, although they usually have a heavy i/o load, and stress the disk subsystem and the locking algorithms. So you can have two separated ext3 filesystems located at the same place on each server, and one or more ocfs2 shared filesystems for the CRS and the database datafiles. The Oracle installer takes care of copying the binaries between the servers during the installation. It might be usefull to try a lower version like 1.2.6, as you are using the latest version available. I am using 1.2.4-2 here with RH 4.0 and kernel 2.6.9-42 and it seems rather stable, only needed to increase the timeouts. (But I dont have the oracle_home shared.) Also you might have a hardware problem somewhere on the SAN. And I still have to check those mount options you sent... One detail. I dont know if the Centos distro includes the OCFS2 module. Are you using the modules downloaded from the oss.oracle.com site for the equivalent RH 4.0 kernel, or modules built by Centos? If using CENTOS modules you might get better results by changing to the Oracle built modules for the equivalent RH 4.0 kernel. Regards, Luis Anjan Chakraborty [EMAIL PROTECTED] wrote: Luis, I am intending to use CRS/RAC that needs a Cluster File System. How does EXT3 falls into that area? Thanks. Anjan Thanks a lot for the response. Here is what I am doing: 1. mkfs.ocfs2 -b 4K -C 32K -N 4 -L ocrvotcrs /dev/sdb3 -- for CRS mkfs.ocfs2 -b 4K -C 32K -N 4 -L orasys /dev/sdb4 -- for RDBMS 2. Then mounting using /etc/fstab: /dev/sdb3 /home/oracle/ocrvotcrsocfs2_netdev,datavolume,nointr 0 0 /dev/sdb4 /home/oracle/orasysocfs2_netdev 0 0 If you find anything wrong here, can you please tell what to do? It's a non-production system so I can experiment with whatever you suggest and won't held you responsible for that. Thanks. Anjan Luis Freitas [EMAIL PROTECTED] wrote: Anjan, Are you installing the binaries on OCSF2 too? How are you mounting the filesystem? You might want to try using ext3 for the binaries and OCF2 only for datafiles and archives, until you get this fixed. Regards, Luis Anjan Chakraborty [EMAIL PROTECTED] wrote:Hi, I sent an email to Mark Fisheh of Oracle Corp. posted this issue at OTN under Linux thread this morning. I hope that someone among you might have experienced this and can help. On that basis, I am sending this to you too. I am stuck will really appreciate if you can shed some light on this. Thanks. Anjan *** I have a 2 node CentOS 4.5 86_64 system (kernel 2.6.9-55.EL). On this I installed Oracle OCFS2 1.2.7-1 (with exact kernel matching). After this I installed Oracle CRS 10.2.0.1 and that installation went fine. Then I tried to install Oracle RDBMS 10.2.0.1 and all the problems started from there. The /var/log/messages file got filled up with messages (giving some to avoid confusion): ocfs2_read_locked_inode: .. : ERROR: Invalid dinode #0 signature = ocfs2_lookup: .. : ERROR: Unable to create inode Then OUI gave several error messages, e.g. Invalid stored block length on file ./em/em.war followed by I/O error in file Errors in invoking to files ins_rdbms.mk and ins_ldap.mk Then /var/log/messages gave: OCFS2: ERROR (device ): ocfs2_extend_file: Dinode # .. has bad signature O' # I And the installation failed CRS died. And the machines reboot. I ran fsck.ocfs2 -n /dev/, it came clean. I have tested this several timnes always same thing happening. If I use RAW partitions, everything works fine. So, the problem may be in the OCFS2 OS/Oracle -- but, not sure how to bypass this. I have to have OCFS2 -- can't use RAW for various reasons. Can somebody please help me to resolve this? Thanks. *** - Be a better pen pal. Text or chat with friends inside Yahoo! Mail. See how.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Be a better pen pal. Text or chat with friends inside Yahoo! Mail. See how. - Never miss a thing. Make Yahoo your homepage. - Get easy, one-click access to your favorites. Make Yahoo! your homepage.___ Ocfs2-users mailing list Ocfs2-users
Re: [Ocfs2-users] OCFS2 on CentOS 4.5 for CRS/RAC
Anjan, Well it is only a sugestion. About those mount options, perhaps the Oracle people can give you a feedback on that also. But I noticed that you have two mount points, and seems to be installing CRS binaries on one, and RDBMS binaries on the other. So, where are datafiles? You shouldnt be installing the CRS home on /home/oracle/ocrvotcrs as it has the datavolume option. It is not clear if you are trying this. The filesystems where you have binaries should not have the datavolume option, and the filesystems with datafiles, voting disks and the cluster registry need to have the datavolume option. Regards, Luis Anjan Chakraborty [EMAIL PROTECTED] wrote: Mark, Thanks to both of you for trying to help me. I have alreaddy communicated to Luis that I have to have OCFS2 because my software needs Clustering technology that is absent in EXT3. Moreover, both CRS RDBMS homes should also be on shared/clustered system -- so, OCFS2 is the only choice. If you can help me to understand/resolve this issue, I will really appreciate that. Please note that my FireWire shared drive works perfectly when I use RAW but as soon as I am trying to use OCFS2, all the problems started. Thanks. Anjan Mark Fasheh [EMAIL PROTECTED] wrote: Luis, Thanks for helping here. I have one comment regarding moving the shared home to ext3. If Anjan's setup is having issues like this, moving his oracle home to a local disk would only hide problems which take longer to reproduce for the crs and data files partitions. What needs to happen is that the root cause of his problems is discovered and fixed. Keep in mind, I'm not saying someone couldn't use ext3 for their oracle install - it's an excellent choice for running a non-shared oracle home. It's just not a good shared disk diagnostic tool ;) Once again, thanks for stepping up and lending a hand. --Mark On Tue, Nov 27, 2007 at 08:50:28AM -0800, Luis Freitas wrote: Anjan, You dont need to share the database binaries, only the CRS and the datafiles. You can do it to save disk space, but it is not mandatory. The CRS and datafiles are much less stressfull to the filesystem structures as there is a reduced number of large files, although they usually have a heavy i/o load, and stress the disk subsystem and the locking algorithms. So you can have two separated ext3 filesystems located at the same place on each server, and one or more ocfs2 shared filesystems for the CRS and the database datafiles. The Oracle installer takes care of copying the binaries between the servers during the installation. It might be usefull to try a lower version like 1.2.6, as you are using the latest version available. I am using 1.2.4-2 here with RH 4.0 and kernel 2.6.9-42 and it seems rather stable, only needed to increase the timeouts. (But I dont have the oracle_home shared.) Also you might have a hardware problem somewhere on the SAN. And I still have to check those mount options you sent... One detail. I dont know if the Centos distro includes the OCFS2 module. Are you using the modules downloaded from the oss.oracle.com site for the equivalent RH 4.0 kernel, or modules built by Centos? If using CENTOS modules you might get better results by changing to the Oracle built modules for the equivalent RH 4.0 kernel. Regards, Luis Anjan Chakraborty wrote: Luis, I am intending to use CRS/RAC that needs a Cluster File System. How does EXT3 falls into that area? Thanks. Anjan Thanks a lot for the response. Here is what I am doing: 1. mkfs.ocfs2 -b 4K -C 32K -N 4 -L ocrvotcrs /dev/sdb3 -- for CRS mkfs.ocfs2 -b 4K -C 32K -N 4 -L orasys /dev/sdb4 -- for RDBMS 2. Then mounting using /etc/fstab: /dev/sdb3 /home/oracle/ocrvotcrs ocfs2 _netdev,datavolume,nointr 0 0 /dev/sdb4 /home/oracle/orasys ocfs2 _netdev 0 0 If you find anything wrong here, can you please tell what to do? It's a non-production system so I can experiment with whatever you suggest and won't held you responsible for that. Thanks. Anjan Luis Freitas wrote: Anjan, Are you installing the binaries on OCSF2 too? How are you mounting the filesystem? You might want to try using ext3 for the binaries and OCF2 only for datafiles and archives, until you get this fixed. Regards, Luis Anjan Chakraborty wrote: Hi, I sent an email to Mark Fisheh of Oracle Corp. posted this issue at OTN under Linux thread this morning. I hope that someone among you might have experienced this and can help. On that basis, I am sending this to you too. I am stuck will really appreciate if you can shed some light on this. Thanks. Anjan *** I have a 2 node CentOS 4.5 86_64 system (kernel 2.6.9-55.EL). On this I installed Oracle OCFS2 1.2.7-1 (with exact kernel matching). After this I installed Oracle CRS 10.2.0.1 and that installation
[Ocfs2-users] Fwd: Returned mail: see transcript for details
Anjan, The two installs I did with x86_64 were with ASM, with kernel versions that today are very old, some years ago. Here I am using x86 (even as the hardware is x86_64 capable) due to a managment decision not to go to x86_64. Anyone else on the list? About the mount points, the CRS home is a ORACLE_HOME, so you shouldnt mount it with datavolume. If you want to keep the CRS ORACLE_HOME and the RDBMS ORACLE_HOME each on one filesystem you should have at least a third partition for the cluster registry, voting disks and datafiles, mounted with the datavolume option, and mount those two without it. If you need to get this working fast, I think you should try a older OCFS2 module version. If not keep working with Mark and the other Oracle people here on the list as they can diagnose these errors and get it fixed. What is the firewire disk you are using? Regards, Luis Anjan Chakraborty [EMAIL PROTECTED] wrote: Luis Freitas [EMAIL PROTECTED] wrote:Luis, As far as I know, it's not recommended to mount ORACLE_HOME with datavolume,nointr but that's not the case with CRS. So, I have different mount options for ORA_CRS_HOME (home/oracle/ocrvotcrs) ORACLE_HOME (home/oracle/orasys). If I am wrong, please correct me. Also, I don't have any datafile yet because I don't have the basic thing -- Oracle binary under ORACLE_HOME. Once I have that, I will have datafiles mount will contain datavolume,nointr. Do you know anybody has successfully implemented OCFS2 1.2.7-1 on Red Hat4.5 (kernel 2.6.9-55.EL) under 86_64 architecture? What exactly you have? Thanks for your continuous effort to help me. I hope somebody else will also start joining this discussion help me. I am really stuck -- can't advance at all -- need help desperately. Anjan Anjan, Well it is only a sugestion. About those mount options, perhaps the Oracle people can give you a feedback on that also. But I noticed that you have two mount points, and seems to be installing CRS binaries on one, and RDBMS binaries on the other. So, where are datafiles? You shouldnt be installing the CRS home on /home/oracle/ocrvotcrs as it has the datavolume option. It is not clear if you are trying this. The filesystems where you have binaries should not have the datavolume option, and the filesystems with datafiles, voting disks and the cluster registry need to have the datavolume option. Regards, Luis Anjan Chakraborty [EMAIL PROTECTED] wrote: Mark, Thanks to both of you for trying to help me. I have alreaddy communicated to Luis that I have to have OCFS2 because my software needs Clustering technology that is absent in EXT3. Moreover, both CRS RDBMS homes should also be on shared/clustered system -- so, OCFS2 is the only choice. If you can help me to understand/resolve this issue, I will really appreciate that. Please note that my FireWire shared drive works perfectly when I use RAW but as soon as I am trying to use OCFS2, all the problems started. Thanks. Anjan Mark Fasheh [EMAIL PROTECTED] wrote: Luis, Thanks for helping here. I have one comment regarding moving the shared home to ext3. If Anjan's setup is having issues like this, moving his oracle home to a local disk would only hide problems which take longer to reproduce for the crs and data files partitions. What needs to happen is that the root cause of his problems is discovered and fixed. Keep in mind, I'm not saying someone couldn't use ext3 for their oracle install - it's an excellent choice for running a non-shared oracle home. It's just not a good shared disk diagnostic tool ;) Once again, thanks for stepping up and lending a hand. --Mark On Tue, Nov 27, 2007 at 08:50:28AM -0800, Luis Freitas wrote: Anjan, You dont need to share the database binaries, only the CRS and the datafiles. You can do it to save disk space, but it is not mandatory. The CRS and datafiles are much less stressfull to the filesystem structures as there is a reduced number of large files, although they usually have a heavy i/o load, and stress the disk subsystem and the locking algorithms. So you can have two separated ext3 filesystems located at the same place on each server, and one or more ocfs2 shared filesystems for the CRS and the database datafiles. The Oracle installer takes care of copying the binaries between the servers during the installation. It might be usefull to try a lower version like 1.2.6, as you are using the latest version available. I am using 1.2.4-2 here with RH 4.0 and kernel 2.6.9-42 and it seems rather stable, only needed to increase the timeouts. (But I dont have the oracle_home shared.) Also you might have a hardware problem somewhere on the SAN. And I still have to check those mount options you sent... One detail. I dont know if the Centos distro includes the OCFS2 module. Are you using the modules downloaded from the oss.oracle.com site for the equivalent RH
Re: [Ocfs2-users] Cluster setup
Well, I stand corrected, and do think he has a point. OCFS2 is heavilly tested on Oracle RAC related scenarios, large files, high I/O rates, etc. On other scenarios, dealing with ACLs, and multiple user ids, it might not be so throrougly tested. But I do see a large effort to resolve any issues that are reported. Also there is the point that it is designed for Oracle database use. So the fencing and reboots are intentional, and they happen fast. All nodes in Oracle RAC freeze when one of the nodes is unreachable, either by SAN or network failures, this is inevitable due to the architecture of the database itself, so it is highly desirable that the misbehaving node go down fast, so that the other nodes can start crash recovery of the failed node, during which the database is still hung, since the crashed node memory structures are lost and need to be recovered from the redo and rollback segments. One thing that I think could be improved is that there is no communication between the O2CB and CRS, so if the timeouts are not setup correctly, one can have a situation where O2CB resolves to fence one node, and CRS fences the other. This seems to not happen with vendor clusterware, that always have to be integrated with CRS to be supported with Oracle RAC. Other thing that would be interesting, if technically feasible, would be to have a option for OCFS2 behave more like NFS, and instead of rebooting, fence all I/O to the filesystems on the evicted node, waiting to see if it can rejoin the cluster. When this situation occurs the fenced node would need to consider all the blocks on the buffer cache as invalid, so that other nodes could modify them safely. Not sure how filesystem locks could be handled. Regards, Luis Sunil Mushran [EMAIL PROTECTED] wrote: Randy Ramsdell wrote: I am not taking sides but I think Alexei's postings are a positive contribution to this project and more of a contribution than the lurkers who write nothing to the list. His feedback does have merit and should be considered valuable although it is critical of ocfs2. We, although I strongly disagreed, have stopped using ocfs2 and we WERE using it in a production environment. The so-called lurkers file bugs, validate patches. That helps everybody. His feedback is useless. Ranting about the core design does not help anyone. Dreaming up possible bugs, without actually looking for them, is useless. When you encounter a bug, file a bugzilla with as much relevant information as possible. If you do that, we can fix the bug and everyone benefits. Sending emails like a broken record is not a positive contribution. BTW, I hear great things about Vista. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Need a vacation? Get great deals to amazing places on Yahoo! Travel. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Cluster setup
Alexei, I do not agree on the heavily loaded part. Oracle run certification tests for their database, so the OCFS2 must have passed through this certification process, which must include high load scenarios. Last time I checked LVM2 was not supported for RAC use, so it probably was not tested. I do agree on the part of OCFS2+LVM and OCFS2+LVM+NFS not being well tested. Also I suspect one cannot run a Active-Active NFS cluster without special NFS software. It would need to be Active-Passive. Regards, Luis Alexei_Roudnev [EMAIL PROTECTED] wrote: Yes, it can be done. Question is in reliability: - OCFSv2 is not very stable when it is about millions of files; - OCFSv2 cluster tend to self-fence after a small SAN storage glitches (it is by design so you can't heal it even if you fix all timeouts - just to improve); - OCFSv2 + LVM +_ NFS is not well tested territory. It should work - in theory, IT works practically, under average load and FS size. No one knows how it behaves on a very big storage and a very big file systems in 1 - 2 years of active usage. I manage ti get stable OCFSv2 system here, after applying few pathces and discovering few issues, BUT I use it on lightly-loaded file system (which is not critical at all)to get more statistics on behavior before I wil use it for anything else. If comparing with heartbeat + LVM + reiserfs + NFS: - all technologies in stack are well tested and heavily used; - heartbeat have external fencing (stonith) so it is extremely reliable in the long term - it can recover from almost any failure (sometimes it dont feel failure, it's true); - ReiserFS (or ext3) proved to be very stable on a huge file systems (it is widely used, so we dont expect any problems here). One problem comes from Novell - since they stoped using it as a default, I can';t trust to ReiserFS on SLES10 (because it is not default) but we stil can trust into it on SLES9 etc... (where it is default). Common rule - if you want reliable system, use defaults where possible. OCFSv2 + NFS is not default yet (through OCFSv2 improved dramatically during last 2 years). - Original Message - From: Pavel Georgiev To: Sent: Wednesday, October 10, 2007 1:25 AM Subject: Re: [Ocfs2-users] Cluster setup How about using just OCFSv2 as I described in my first mail - two servers export their storage, the rest of the servers mount it and a failure of any of the two storage servers remains transparent to the clients. Can this be done with OCFSv2? On Tuesday 09 October 2007 21:46:15 Alexei_Roudnev wrote: You better use LVM + heartbeat + NFS + cold failover cluster. It works 100% stable and is 100% safe from the bugs (and it allows online resizing, if your HBA or iSCSI can add lun's on the fly). Combining NFS + LVM + OCFSv2 can cause many unpredictable problems, esp. on the unusual (for OCFSv2) system (such as Ubuntu). - Original Message - From: Brian Anderson To: Sent: Tuesday, October 09, 2007 11:35 AM Subject: RE: [Ocfs2-users] Cluster setup Not exactly. I'm in a similar boat right now. I have 3 NFS servers all mounting an OCFS2 volume. Each NFS server has its own IP, and the clients load balance manually... some mount fs1, others fs2, and the rest fs3. In an ideal world, I'd have the NFS cluster presenting a single IP, and failing over / load balancing some other way. I'm looking at NFS v4 as one potential avenue (no single IP, but it does let you fail over from 1 server to the next in line), and commercial products such as IBRIX. Brian -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Sunil Mushran Sent: Tuesday, October 09, 2007 2:27 PM To: Luis Freitas Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] Cluster setup Unsure what you mean. If the two servers mount the same ocfs2 volume and export them via nfs, isn't that clustered nfs? Luis Freitas wrote: Is there any cluster NFS solution out there? (Two NFS servers sharing the same filesystem with distributed locking and failover capability) Regards, Luis */Sunil Mushran /* wrote: Appears what you are looking for is a mix of ocfs2 and nfs. The storage servers mount the shared disks and the reexport them via nfs to the remaining servers. ubuntu 6.06 is too old. If you are stuck on Ubuntu LTS, the next version 7.10 should have all you want. Pavel Georgiev wrote: Hi List, I`m trying to build a cluster storage with commodity hardware in a way that the all the data would be on 1 server. It should have the meet the following requirements: 1) If one of the servers goes down, the cluster should continue to work with rw access from all clients. 2) Clients that mount the storage should not be part of cluster (not export any disk storage) - I have few servers
Re: [Ocfs2-users] Re: OCF2 and LVM
In OCFSv1 this could only be done with the oracle patched binutils, but it should be safe to backup OCFSv2 files directly. Any input from the developers? Regards, Luis Shivaprasad Kambalimath [EMAIL PROTECTED] wrote: We use EMC's BCV for OCFS2 split mirror and it works just fine (10.2.0.3 and RH Linux 4). The only problem is we can't backup to tape directly from ocfs2 (we are using legato and I haven't heard of any other third party s/w doing this without rman). So, we copy files to ext3, zip them and then take tape backup. Thanks, Shiva Kambalimath -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Monday, October 08, 2007 12:00 PM To: ocfs2-users@oss.oracle.com Subject: Ocfs2-users Digest, Vol 46, Issue 7 Send Ocfs2-users mailing list submissions to ocfs2-users@oss.oracle.com To subscribe or unsubscribe via the World Wide Web, visit http://oss.oracle.com/mailman/listinfo/ocfs2-users or, via email, send a message with subject or body 'help' to [EMAIL PROTECTED] You can reach the person managing the list at [EMAIL PROTECTED] When replying, please edit your Subject line so it is more specific than Re: Contents of Ocfs2-users digest... Today's Topics: 1. OCF2 and LVM (Riccardo Paganini) 2. Re: OCF2 and LVM (Jordi Prats) 3. Re: OCF2 and LVM (Alexei_Roudnev) -- Message: 1 Date: Mon, 08 Oct 2007 12:35:54 +0200 From: Riccardo Paganini Subject: [Ocfs2-users] OCF2 and LVM To: ocfs2-users@oss.oracle.com Message-ID: Content-Type: text/plain;charset=iso-8859-1;format=flowed Does anybody knows if is there a certified procedure in to backup a RAC DB 10.2.0.3 based on OCFS2 , via split mirror or snaphots technology ? Using Linux LVM and OCFS2, does anybody knows if is possible to dinamically extend an OCFS filesystem, once the underlying LVM Volume has been extended ? Thanks in advance Riccardo Paganini -- Message: 2 Date: Mon, 08 Oct 2007 20:20:08 +0200 From: Jordi Prats Subject: Re: [Ocfs2-users] OCF2 and LVM To: Riccardo Paganini Cc: ocfs2-users@oss.oracle.com Message-ID: [EMAIL PROTECTED] Content-Type: text/plain; charset=ISO-8859-1; format=flowed Hi, I'm not an expert on oracle, but I don't think using a snapshot is a recommended backup solution. You should use data pump or exp utilities to backup your database. To extend your fs, tune2fs.ocfs2 requires you to umount it, according to it's man page: == tunefs.ocfs2 is used to adjust OCFS2 file system parameters on disk. In order to prevent data loss, tunefs.ocfs2 will not perform any action on the specified device if it is mounted on any node in the cluster. This tool requires the O2CB cluster to be online. == The concrete command should be: tune2fs.ocfs2 -S /dev/LVM_volume I don't know if there's any way to extend it without umounting it. If you want to extend your fs without umounting it you should use ASM instead of OCFS2. regards, Jordi Riccardo Paganini wrote: Does anybody knows if is there a certified procedure in to backup a RAC DB 10.2.0.3 based on OCFS2 , via split mirror or snaphots technology ? Using Linux LVM and OCFS2, does anybody knows if is possible to dinamically extend an OCFS filesystem, once the underlying LVM Volume has been extended ? Thanks in advance Riccardo Paganini ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users -- Message: 3 Date: Mon, 8 Oct 2007 11:49:29 -0700 From: Alexei_Roudnev Subject: Re: [Ocfs2-users] OCF2 and LVM To: Jordi Prats , Riccardo Paganini Cc: ocfs2-users@oss.oracle.com Message-ID: [EMAIL PROTECTED] Content-Type: text/plain; format=flowed; charset=iso-8859-1; reply-type=response I dont knoiw any lvm technology which can make a snapshort in clustered environment, so better forget this idea. Recommended method to do oracle backups is rman. Any violation of this brings you into the swamp of possible oracle bugs. - Original Message - From: Jordi Prats To: Riccardo Paganini Cc: Sent: Monday, October 08, 2007 11:20 AM Subject: Re: [Ocfs2-users] OCF2 and LVM Hi, I'm not an expert on oracle, but I don't think using a snapshot is a recommended backup solution. You should use data pump or exp utilities to backup your database. To extend your fs, tune2fs.ocfs2 requires you to umount it, according to it's man page: == tunefs.ocfs2 is used to adjust OCFS2 file system parameters on disk. In order to prevent data loss, tunefs.ocfs2 will not perform any action on the specified device if it is mounted on any node in the cluster. This tool requires the O2CB cluster to be online. == The concrete command should be: tune2fs.ocfs2 -S /dev/LVM_volume I don't know if there's any way to extend it without
RE: [Ocfs2-users] OCFS2 mount problem at Linux reboot when device names are non persistent.
Anyone got this working on emc powerpath? I remeber seeing somewhere some special configuration to get mounting by label working with multipath devices, but could not find it again. Regards, Luis Ricardo Fernandez [EMAIL PROTECTED] wrote: Hi people, Thanks a lot for your answer! It worked perfectly. Best regards Ricardo -Original Message- From: Sunil Mushran [mailto:[EMAIL PROTECTED] Sent: viernes, 24 de agosto de 2007 14:26 To: Randy Ramsdell Cc: ocfs2-users@oss.oracle.com; Ricardo Fernandez Subject: Re: [Ocfs2-users] OCFS2 mount problem at Linux reboot when device names are non persistent. While mount-by-uuid will work, mount-by-label should also work. The one gotcha in the latter is that it expects the device to be partitioned. As in, it will not mount-by-label if the device is /dev/sda but will if the device is /dev/sda1 or sda2, etc. Randy Ramsdell wrote: Ricardo Fernandez wrote: Hi, I have the following problem when the servers accessing OCFS2 reboot: as the Linux device names are non persistent, at reboot they usually change, and then OCFS2 can't mount the device because it is expecting a different device name as stated in the fstab file. (it is specified in the format /dev/sdx as the instructions of the OCFS2 installation manual say) If I change the device name to the new name, it works fine. But this is not an acceptable solution, as each node should be able to start in a fully automatic way. (without human intervention) I thought that the purpose of the disk LABEL that I added when formatting the partition with OCFS2 was exactly this. (Am I right?) I changed the fstab to use the LABEL option, and also try to mount it from the command line using the LABEL option but it didn't work. Is there any bug or known issue on this topic. I guess that if I glue the device name with udev it will work, but I really would expect OCF2 to solve this problem (because it is not a new one, and most of the file systems I know can handle it) I would appreciate any help on this topic. Thanks a lot Ricardo I work with: RHEL 4 Local SCSI devices External devices locates in an EVA8000 SAN, accessed through a fibre channel bus. The OCFS2 file system is on one of these. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users Do not use label use UUID name and _netdev_ fstab option. This is the UUID of a volume we have. /dev/disk/by-uuid/be12775a-ec1c-4ed7-a06b-f30a081a0603 UUID's are unique and never change so they are ideal for what you are describing. Randy Ramsdell ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Need a vacation? Get great deals to amazing places on Yahoo! Travel. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Urgent :: 11i on OCFS2.. I mean APPL_TOP, COMMON_TOP etc..
Anyone have coments on the shared mmap? What could happen if a file is mmaped by say, 3000 different processes in a 4 node cluster, and someone went there and changed its contents? This is a possible situation on a large production APPS environment, or any forms based application. Regards, Luis Alexei_Roudnev [EMAIL PROTECTED] wrote: It's really a good question. We should know _which exactly combinations are used internally_. (OS, File System, NAS/SAN). In reality therse particular combinations are most stable and preffered ones - no matter what is _officially certified_. - Original Message - From: Luis Freitas To: [EMAIL PROTECTED] ; ocfs2-users@oss.oracle.com Sent: Wednesday, August 22, 2007 6:46 PM Subject: Re: [Ocfs2-users] Urgent :: 11i on OCFS2.. I mean APPL_TOP,COMMON_TOP etc.. Prabhakar, This post is only my personal opinion. I do not work for Oracle nor have any close contact with the support and development groups. But I never heard of anyone using this combination. This is kind of a gray area, since Oracle usually provides support for OCFS2 in conjunction with RAC database. The APPS people is a different group internally at Oracle, so even if this is officially supported you could get into situations where the support analysts start to redirect the TAR between these groups, or could argue that since the issues are not related to Oracle Database or Oracle database binaries they would not fix it. I would not do it unless with a very large customer that could escalate any issues to get a fast resolution with support. It probably is better to get a NAS storage and map it using NFS. That said since it is supported for the database binaries, there should not be immediate problems with other files. The FAQ about shared APPL_TOP on metalink says that any disk sharing technology can be used. One thing that comes to my attention is that Oracle Forms uses memory mapped files extensively. I dont know how completelly this is implemented on OCFS2 and if it works well. Anyone can comment on this? There are some posts on this list about problems with shared mmap. Regards, Luis [EMAIL PROTECTED] wrote: All, Is Shared APPL_TOP certified with 11i on OCFS2? Please send me the information as soon as possible.. We have two RAC nodes on ASM file system and one middle tier right now. Now want to setup Shared APPL_TOP so need your help whether this is a certied with OCFS2 or not. OS is SUSE 9 SP3. Regards, Prabhakar Tammina. (614)598-3487(cell). ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Got a little couch potato? Check out fun summer activities for kids. - ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Fussy? Opinionated? Impossible to please? Perfect. Join Yahoo!'s user panel and lay it on us.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Using OCFS2 for Shared APPL_TOP
I cant read Russian... Does it work??? Regards, Luis Alexei_Roudnev [EMAIL PROTECTED] wrote: There is discussion about it on www.sql.ru, but it is in Russian. - Original Message - From: Santosh Udupa To: Sent: Thursday, August 23, 2007 12:04 PM Subject: [Ocfs2-users] Using OCFS2 for Shared APPL_TOP Hello all, Has anybody seen any issue with using OCFS2 for shared APPL_TOP in E-Business Suite 11i and Release 12? Thanks, Santosh CAUTION - Disclaimer * This e-mail contains PRIVILEGED AND CONFIDENTIAL INFORMATION intended solely for the use of the addressee(s). If you are not the intended recipient, please notify the sender by e-mail and delete the original message. Further, you are not to copy, disclose, or distribute this e-mail or its contents to any other person and any such actions are unlawful. This e-mail may contain viruses. Infosys has taken every reasonable precaution to minimize this risk, but is not liable for any damage you may sustain as a result of any virus in this e-mail. You should carry out your own virus checks before opening the e-mail or attachment. Infosys reserves the right to monitor and review the content of all messages sent to or from this e-mail address. Messages sent to or from this e-mail address may be stored on the Infosys e-mail system. ***INFOSYS End of Disclaimer INFOSYS*** ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Fussy? Opinionated? Impossible to please? Perfect. Join Yahoo!'s user panel and lay it on us.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] cluvfy fails in pre crs install
Harry, Do you have DISPLAY set on the root shell? Try to run xclock and see if it appears correctly. I am not sure about this but I think that you need to load the environment variables that you have defined on you oracle account when running these tools as root. Usually I logon with the oracle account and do a su without the - parameter so the environment is not overwriten. Also when you are with the root user, check if you have any java executable on the path that is not from the JDK bundled in ORACLE_HOME/jdk. This could be either GNU gcj or IBM Jdk, and can cause problems with the graphical tools if they are used by mistake. which java java -version Regards, Luis Harry Ronis [EMAIL PROTECTED] wrote: body{font-family: Geneva,Arial,Helvetica,sans-serif;font-size:9pt;background-color: #ff;color: black;}body{font-family: Geneva,Arial,Helvetica,sans-serif;font-size:9pt;background-color: #ff;color: black;}body{font-family: Geneva,Arial,Helvetica,sans-serif;font-size:9pt;background-color: #ff;color: black;}body{font-family: Geneva,Arial,Helvetica,sans-serif;font-size:9pt;background-color: #ff;color: black;} Did this,blew away clusterware as per metalink note, reinstalled clustrware this time properly filling in two node cluster and got clusterware installled on both. Thks. $ ssh hoffman /apps/crs/oracle/product/10/app/bin/olsnodes -n hackman 1 hoffman 2 [EMAIL PROTECTED] /apps/crs/oracle/product/10/app/bin $ ./olsnodes -n hackman 1 hoffman 2 Now this: Need to manually run vipca as ROOT to config virtual ips. in runing vip this is what happens: AS ROOT [EMAIL PROTECTED] bin]# ./vipca Exception in thread main [EMAIL PROTECTED] bin]# [EMAIL PROTECTED] bin]# ./vipca Exception in thread main [EMAIL PROTECTED] bin]# I believe this is the LAST HUMP in clusterware -Original Message- From: Luis Freitas Sent: Aug 21, 2007 3:35 PM To: Harry Ronis , [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] cluvfy fails in pre crs install Harry, From this it looks like you forgot to fill the second node details on the installer. There is a screen on the begining of the installation where the installer asks for all the nodes hostnames, public node interfaces and private node interfaces, and it brings only the first node details as default. This root.sh and the rest of the installation get the node names from that screen. But the only way to confirm this would be to look at the installer logs. You probably can add the second node but it may be easier to cleanup and reinstall. There is a note with a procedure to manually remove CRS on metalink, you need to stop the services, restore inittab and remove some rc.d scripts. Regards, Luis Harry Ronis [EMAIL PROTECTED] wrote: Why is it only seeing 1 node when it run root.sh second node empty -Original Message- From: Harry Ronis Sent: Aug 21, 2007 1:03 PM To: Luis Freitas , [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] cluvfy fails in pre crs install running install --asking to run root.sh from hackman only Didn't populated anything on Hoffman yet -- is this normal ?? Checking to see if Oracle CRS stack is already configured /etc/oracle does not exist. Creating it now. Setting the permissions on OCR backup directory Setting up NS directories Oracle Cluster Registry configuration upgraded successfully Successfully accumulated necessary OCR keys. Using ports: CSS=49895 CRS=49896 EVMC=49898 and EVMR=49897. node nodenumber: nodename private interconnect name hostname node 1: hackman hackman-priv hackman Creating OCR keys for user 'root', privgrp 'root'.. Operation successful. Now formatting voting device: /ocfs2/crs/vdisk1 Now formatting voting device: /ocfs2/crs/vdisk2 Now formatting voting device: /ocfs2/crs/vdisk3 Format of 3 voting devices complete. Startup will be queued to init within 90 seconds. Adding daemons to inittab Expecting the CRS daemons to be up within 600 seconds. CSS is active on these nodes. hackman CSS is active on all nodes. Waiting for the Oracle CRSD and EVMD to start Oracle CRS stack installed and running under init(1M) Running vipca(silent) for configuring nodeapps The given interface(s), eth0 is not public. Public interfaces should be used to configure virtual IPs. -Original Message- From: Luis Freitas Sent: Aug 21, 2007 11:25 AM To: Harry Ronis , [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] cluvfy fails in pre crs install Harry, This tool fails here too if I put these -c and -q parameters for the storage checks. Even so RAC is installed and running. I would not pay much attention to this at this point. Also as far as I know these parameters are originaly intended to work with raw devices. It is a generic tool that run in all
Re: [Ocfs2-users] cluvfy fails in pre crs install
Harry, This tool fails here too if I put these -c and -q parameters for the storage checks. Even so RAC is installed and running. I would not pay much attention to this at this point. Also as far as I know these parameters are originaly intended to work with raw devices. It is a generic tool that run in all platforms and probably the developer did not pay much attention to OCFS2, which is specific on the Linux platform. That error on the VIP checking also is documented and you can ignore it at this point. But you probably will have a error in the end of the CRS install and in this case there is manual procedure that need to be done with vipca before you start to install the database. touch /ocfs2/ocr2 touch /ocfs2/vdisk1 Regards, Luis Harry Ronis [EMAIL PROTECTED] wrote: body{font-family: Geneva,Arial,Helvetica,sans-serif;font-size:9pt;background-color: #ff;color: black;} Am at a LOSS Shared storage seems ok -- yet cluvfy for ANY shared storage commands FAIL ala [EMAIL PROTECTED] /apps/oracle/stage/clusterware/cluvfy $ ./runcluvfy.sh comp ssa -n hackman,hoffman -verbose Verifying shared storage accessibility Checking shared storage accessibility... Shared storage check failed on nodes hoffman,hackman. Verification of shared storage accessibility was unsuccessful on all the nodes. [EMAIL PROTECTED] /apps/oracle/stage/clusterware/cluvfy -Original Message- From: Marcos E. Matsunaga Sent: Aug 21, 2007 8:26 AM To: hr Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] cluvfy fails in pre crs install Try to create the ocr/voting under a directory on /ocfs2. If I'm not wrong, there was a bug on CRS install identifying ocr/voting disks directly under the mountpoint. Also, make sure both nodes can mount the same partition. Create a file on one node and see if you can remove it from the other node. Regards,Marcos Eduardo MatsunagaOracle USA Linux Engineering hr wrote: /runcluvfy.sh stage -pre crsinst -n hackman,hoffman -c /ocfs2/ocr2 -q /ocfs2/vdisk1 -verbose output Performing pre-checks for cluster services setup Checking node reachability... Check: Node reachability from node hackman Destination Node Reachable? hackman yes hoffman yes Result: Node reachability check passed from node hackman. Checking user equivalence... Check: User equivalence for user oracle Node Name Comment hoffman passed hackman passed Result: User equivalence check passed for user oracle. Checking administrative privileges... Check: Existence of user oracle Node Name User Exists Comment hoffman yes passed hackman yes passed Result: User existence check passed for oracle. Ch Administrative privileges check failed. Checking node connectivity... Suitable interfaces for the private interconnect on subnet 10.32.250.0: hoffman eth0:10.32.250.175 hackman eth0:10.32.250.155 Suitable interfaces for the private interconnect on subnet 192.168.1.0: hoffman eth1:192.168.1.2 hackman eth1:192.168.1.1 ERROR: Could not find a suitable set of interfaces for VIPs. Result: Node connectivity check failed. Checking shared storage accessibility... ERROR: /ocfs2/ocr2 Could not find the storage Shared storage check failed on nodes hoffman,hackman. Checking shared storage accessibility... ERROR: /ocfs2/vdisk1 Could not find the storage Shared storage check failed on nodes hoffman,hackman. Checking system requirements for 'crs'... No checks registered for this product. - ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users Harry Roniscell 646 529 8853Home 718 224 9793work 212 465v 2826 ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Moody friends. Drama queens. Your life? Nope! - their life, your story. Play Sims Stories at Yahoo! Games. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] cluvfy fails in pre crs install
Harry, From this it looks like you forgot to fill the second node details on the installer. There is a screen on the begining of the installation where the installer asks for all the nodes hostnames, public node interfaces and private node interfaces, and it brings only the first node details as default. This root.sh and the rest of the installation get the node names from that screen. But the only way to confirm this would be to look at the installer logs. You probably can add the second node but it may be easier to cleanup and reinstall. There is a note with a procedure to manually remove CRS on metalink, you need to stop the services, restore inittab and remove some rc.d scripts. Regards, Luis Harry Ronis [EMAIL PROTECTED] wrote: body{font-family: Geneva,Arial,Helvetica,sans-serif;font-size:9pt;background-color: #ff;color: black;}body{font-family: Geneva,Arial,Helvetica,sans-serif;font-size:9pt;background-color: #ff;color: black;}body{font-family: Geneva,Arial,Helvetica,sans-serif;font-size:9pt;background-color: #ff;color: black;} Why is it only seeing 1 node when it run root.sh second node empty -Original Message- From: Harry Ronis Sent: Aug 21, 2007 1:03 PM To: Luis Freitas , [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] cluvfy fails in pre crs install running install --asking to run root.sh from hackman only Didn't populated anything on Hoffman yet -- is this normal ?? Checking to see if Oracle CRS stack is already configured /etc/oracle does not exist. Creating it now. Setting the permissions on OCR backup directory Setting up NS directories Oracle Cluster Registry configuration upgraded successfully Successfully accumulated necessary OCR keys. Using ports: CSS=49895 CRS=49896 EVMC=49898 and EVMR=49897. node nodenumber: nodename private interconnect name hostname node 1: hackman hackman-priv hackman Creating OCR keys for user 'root', privgrp 'root'.. Operation successful. Now formatting voting device: /ocfs2/crs/vdisk1 Now formatting voting device: /ocfs2/crs/vdisk2 Now formatting voting device: /ocfs2/crs/vdisk3 Format of 3 voting devices complete. Startup will be queued to init within 90 seconds. Adding daemons to inittab Expecting the CRS daemons to be up within 600 seconds. CSS is active on these nodes. hackman CSS is active on all nodes. Waiting for the Oracle CRSD and EVMD to start Oracle CRS stack installed and running under init(1M) Running vipca(silent) for configuring nodeapps The given interface(s), eth0 is not public. Public interfaces should be used to configure virtual IPs. -Original Message- From: Luis Freitas Sent: Aug 21, 2007 11:25 AM To: Harry Ronis , [EMAIL PROTECTED] Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] cluvfy fails in pre crs install Harry, This tool fails here too if I put these -c and -q parameters for the storage checks. Even so RAC is installed and running. I would not pay much attention to this at this point. Also as far as I know these parameters are originaly intended to work with raw devices. It is a generic tool that run in all platforms and probably the developer did not pay much attention to OCFS2, which is specific on the Linux platform. That error on the VIP checking also is documented and you can ignore it at this point. But you probably will have a error in the end of the CRS install and in this case there is manual procedure that need to be done with vipca before you start to install the database. touch /ocfs2/ocr2 touch /ocfs2/vdisk1 Regards, Luis Harry Ronis [EMAIL PROTECTED] wrote: Am at a LOSS Shared storage seems ok -- yet cluvfy for ANY shared storage commands FAIL ala [EMAIL PROTECTED] /apps/oracle/stage/clusterware/cluvfy $ ./runcluvfy.sh comp ssa -n hackman,hoffman -verbose Verifying shared storage accessibility Checking shared storage accessibility... Shared storage check failed on nodes hoffman,hackman. Verification of shared storage accessibility was unsuccessful on all the nodes. [EMAIL PROTECTED] /apps/oracle/stage/clusterware/cluvfy -Original Message- From: Marcos E. Matsunaga Sent: Aug 21, 2007 8:26 AM To: hr Cc: ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] cluvfy fails in pre crs install Try to create the ocr/voting under a directory on /ocfs2. If I'm not wrong, there was a bug on CRS install identifying ocr/voting disks directly under the mountpoint. Also, make sure both nodes can mount the same partition. Create a file on one node and see if you can remove it from the other node. Regards,Marcos Eduardo MatsunagaOracle USA Linux Engineering hr wrote: /runcluvfy.sh stage -pre crsinst -n hackman,hoffman -c /ocfs2/ocr2 -q /ocfs2/vdisk1 -verbose output Performing pre-checks for cluster services setup Checking node reachability
Re: [Ocfs2-users] Replication not works
Yohan The device you are using for OCFS2 need to be shared between nodes, either a shared device on a external storage or a mirrored device using DRDB or iSCSI. For Oracle database use on a certified configuration you need to use a external storage, as drdb is not a officially supported option for Oracle Database. The OCFS does not replicate the data, it only manages the filesystem so that it can be mounted simultaneosly on both nodes. Regards, Luis Yohan [EMAIL PROTECTED] wrote: Dr J Pelan a écrit : On Fri, 10 Aug 2007, Yohan wrote: [EMAIL PROTECTED]:~# mounted.ocfs2 -f Device FS Nodes /dev/sda3 ocfs2 t1 [EMAIL PROTECTED]:~# mounted.ocfs2 -f Device FS Nodes /dev/sda3 ocfs2 t2 Tell us about /dev/sda. /dev/sda is the same on the two nodes Disk /dev/sda: 73.4 GB Device Boot Start End Blocks Id System /dev/sda1 * 1 2432 19535008+ 83 Linux /dev/sda2 2433 2675 1951897+ 82 Linux swap / Solaris /dev/sda3 2676 2800 1004062+ 83 Linux did i make a mistake here ? The use of the word 'replication' is interesting. My 1st goal is to have clusters of 3 nodes. Each node need to be the mirror of the other one. All of them are read/write accessed. It's not what ocfs2 does ? ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Looking for a deal? Find great prices on flights and hotels with Yahoo! FareChase.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Kernel panic with OCFS2 1.2.6 for EL5
I always get these annoying messages with EMC storages. They seem to have a unusable lun that appears for the operating system and the kernel keeps trying to access it. On RH 4.0 these messages appears on dmesg, I remember that RH 3.0 showed them on /var/log/messages also: SCSI error : 0 0 0 0 return code = 0x2 SCSI error : 0 0 0 0 return code = 0x2 SCSI error : 1 0 0 1 return code = 0x2 SCSI error : 0 0 0 0 return code = 0x2 SCSI error : 0 0 0 0 return code = 0x2 SCSI error : 1 0 0 0 return code = 0x2 Anyone knows how to prevent the kernel from trying to access this lun? Regards, Luis Freitas Daniel [EMAIL PROTECTED] wrote: Hello System: Two brand new Dell 1950 servers with dual Intel Quadcore Xeon connected to an EMC CX3-20 SAN. Running CentOS 5 x86_64 - both with kernel 2.6.18-8.1.6-el5 x86_64. I just noticed a panic on one of the servers: Jul 2 04:08:52 megasrv2 kernel: (3568,2):dlm_drop_lockres_ref:2289 ERROR: while dropping ref on 87B24E40651A4C7C858EF03ED6F3595F:M00021af916b7dfbde4 (master=0) got -22. Jul 2 04:08:52 megasrv2 kernel: (3568,2):dlm_print_one_lock_resource:294 lockres: M00021af916b7dfbde4, owner=0, state=64 Jul 2 04:08:52 megasrv2 kernel: (3568,2):__dlm_print_one_lock_resource:309 lockres: M00021af916b7dfbde4, owner=0, state=64 Jul 2 04:08:52 megasrv2 kernel: (3568,2):__dlm_print_one_lock_resource:311 last used: 4747810336, on purge list: yes Jul 2 04:08:52 megasrv2 kernel: (3568,2):dlm_print_lockres_refmap:277 refmap nodes: [ ], inflight=0 Jul 2 04:08:52 megasrv2 kernel: (3568,2):__dlm_print_one_lock_resource:313 granted queue: Jul 2 04:08:52 megasrv2 kernel: (3568,2):__dlm_print_one_lock_resource:328 converting queue: Jul 2 04:08:52 megasrv2 kernel: (3568,2):__dlm_print_one_lock_resource:343 blocked queue: Jul 2 04:08:52 megasrv2 kernel: --- [cut here ] - [please bite here ] - After booting the server I'm getting a lot of the following messages: Jul 5 11:09:54 megasrv2 kernel: Additional sense: Logical unit not ready, manual intervention required Jul 5 11:09:54 megasrv2 kernel: end_request: I/O error, dev sdd, sector 0 Jul 5 11:09:54 megasrv2 kernel: Buffer I/O error on device sdd, logical block 0 Jul 5 11:09:54 megasrv2 kernel: sd 1:0:0:2: Device not ready: 6: Current: sense key: Not Ready But I guess this one has something to do with EMC PowerPath as sdd is not a valid device. And there is no PowerPath for use with RHEL5 yet... I'm sorry I haven't had the time to investigate this much. But right now I have no clue what caused this panic, or if it will happen again... ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Luggage? GPS? Comic books? Check out fitting gifts for grads at Yahoo! Search.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Best way to use multiple ocvfs2 volumes
Randy, If you have one big volume with several partitions you will have to stop all servers if you want to create new partitions or delete one, or be extremelly careful. With LVM2 and CLVM you could work with one big disk without this kind of trouble. Personally I prefer having multiple small volumes and letting the storage do the hard work. You will still need to reboot the server or reload the hba module to force a bus scan in order to see new volumes, but you can do this one server at a time. Before doing this kind of reorganization on production with everything running I would recommend testing it well. RedHat 2.1 used to rescan the bus dynamically and I had some bad experiences, like sda suddenly becoming sdb with the filesystems mounted. I believe current kernels are better with this kind of stuff, and you should have no trouble if you are using a software like powerpath. I do not know if there is any way to force a bus rescan without reloading the hba module. Sugestions, anyone? Regards, Luis Marcos E. Matsunaga [EMAIL PROTECTED] wrote:Randy, You can either create multiple partitions in a volume and mount them or just use the whole volume, it doesn't matter. Make sure the disks are accessible on all nodes, define your cluster, format and mount. Works just like other FS, with the exception of the cluster part. A single cluster can handle a large number of partitions without problem. Randy Ramsdell wrote: Hi, What would be the best way to implement multi-use of a ocfs2 cluster? We need to have several (2-4) different mount points that provide clustered data for different services. I don't think we can have 2 different clusters on 1 machine so would mounting multiple ocfs2 partitions/volumes for each of these services on the same cluster work? Or what would be the best way to handle this? We could use 1 large volume with multiple partitions or several volumes with a partition/raw device. Thanks, Randy Ramsdell ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users -- Regards, Marcos Eduardo Matsunaga Oracle USA Linux Engineering ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Building a website is a piece of cake. Yahoo! Small Business gives you all the tools to get online.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] PBL with RMAN and ocfs2
Gaetano, If o2cb or CRS is killing the machine, it usually shows on /var/log/messages with lines explaining what happened. Take a look on the /var/log/messages just before the last syslogd x.x.x: restart. Regards, Luis Gaetano Giunta wrote: Hello. On a 2 node RAC 10.2.0.3 setup, on RH ES 4.4 x86_64, with ocfs 1.2.5-1, we are experiencing some troubles with RMAN: when the archive log destination is on an ASM partition, and the backup detsination is on ocfs2, running backup archivelog all format '/home/SANstorage/oracle/backup/rman/dump_log/FULL_20070509_154916/arc_%d_%u' delete input; consistently causes a reboot. The rman catalog is clean, and has been crosschecked in every way. We tried on both nodes, and the node executing the backup always reboots. I am thus inclined to think that it is not the ocfs2 dlm that triggers the reboot, because in that case the victim would always be the second node. I also tested the same command using as backup destination /tmp, and all was fine. The backup file of the archived logs is 1249843712 in size. Our local oracle guy went through metalink and said there is no open bug/patch for that at this time. Any suggestions ??? Thanks Gaetano Giunta ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Ahhh...imagining that irresistible new car smell? Check outnew cars at Yahoo! Autos.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
RE: [Ocfs2-users] Hi
Ulf, I have implemented a few RAC databases and worked with OCFS, OCFS2, NetApp and ASM on different machines. In my opinion, currently, OCFS2 seems to be rather stable for Oracle use, except for some race conditions with CRS when the CRS files are on OCFS2 and the small default timeouts. Both can be tuned or worked around so as to not cause major headaches. One thing that caused me trouble was the fact that 10g has asyncronous I/O linked in by default, this was changed from what was on 9i, and the async I/O handlers deplete very fast if they are not tuned up when using OCFS2. Somehow this seems to be completely undocumented except for some ancient RedHat papers, and seems to cause problems in CRS. This can be seen as /proc/sys/fs/aio-nr reaches the value on /proc/sys/fs/aio-max-nr. Aside from that usually the problems I have are on changes of the optimizer or some 10g specific bugs causing ORA-600. One thing that is very important is to gather statistics, including system statistics, as there is a big change on the optimizer that now considers also the CPU cost for choosing the execution plan. This change on the optimizer also appears when applying 9.2.0.7 and 9.2.0.8. Hope you have better luck on the next try. Regards, Luis Ulf Zimmermann [EMAIL PROTECTED] wrote: -Original Message- From: [EMAIL PROTECTED] [mailto:ocfs2-users- [EMAIL PROTECTED] On Behalf Of Sunil Mushran Sent: 05/07/2007 10:47 To: Alexei_Roudnev Cc: Ocfs2-users@oss.oracle.com Subject: Re: [Ocfs2-users] Hi None of what you have written allows you to use our resources to spread your opinions as official recommendation. Alexei_Roudnev wrote: Oracle itself have not a SINGLE opinion (to be curious, I hear a strong recommendation against OCFSv2 from oracle support, which I can not agree with), so we can't treat your recommendations as official as well - you are interested in OCFSv2 while users are not (users are interested in making our data centers run smoothly). The only _official_ thing is _certification matrix_. - Original Message - Alexei, While you are free to use this forum to share your opinions, do not couch these opinions as official recommendations. When push comes to shove, we are helping users not you. We develop, build, distribute the software, not you. So it may serve to community better if you let us offer the official recommendations and not you. Sunil Just to add some comments from a user of Oracle 9i with OCFSv1 on RedHat AS2.1 who tried to upgrade to EL4 and OCFSv2 and failed miserable: Oracle support pretty much told us the problems we were running into are problems of OCFSv2 and they weren't really willing to help us. The feeling we were getting was that two Oracle departments (the one writing the Database RAC engine and the one writing OCFSv2) are fighting with each other. In general I have a very low opinion of Oracle and their quality of code and tools. Like patch revision numbering? Does not exist. Patch tools suppose to patch all machines in clusters? You wish. Decent error messages? They never heard about that. We ended up with staying on AS2.1 and OCFSv1 for now and just migrating our data to a new SAN. Regards, Ulf. - ATC-Onlane Inc., T: 650-532-6382, F: 650-532-6441 4600 Bohannon Drive, Suite 100, Menlo Park, CA 94025 - ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Ahhh...imagining that irresistible new car smell? Check outnew cars at Yahoo! Autos.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] High on buffers and deep on swap
Alexei, How can I relate the information on slabtop to the actual memory used by buffers? I see this on slabtop: Active / Total Objects (% used): 603822 / 649643 (92.9%) Active / Total Slabs (% used) : 47216 / 47216 (100.0%) Active / Total Caches (% used) : 97 / 133 (72.9%) Active / Total Size (% used) : 176601.08K / 181508.49K (97.3%) Minimum / Average / Maximum Object : 0.01K / 0.28K / 128.00K OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME 202461 202451 99%0.54K 289237115692K ext3_inode_cache 232206 232153 99%0.15K 8931 26 35724K dentry_cache 29974 26092 87%0.27K 2141 14 8564K radix_tree_node 68250 62400 91%0.05K910 75 3640K buffer_head 855855 100%4.00K8551 3420K pmd 647647 100%4.00K6471 2588K size-4096 8595 7694 89%0.25K573 15 2292K filp 20835 17331 83%0.09K463 45 1852K vm_area_struct 780767 98%2.00K3902 1560K size-2048 18849 7962 42%0.06K309 61 1236K size-64 2440632 25%0.50K3058 1220K size-512 2926 2889 98%0.34K266 11 1064K inode_cache 256256 100%3.00K1282 1024K biovec-(256) 600592 98%1.38K1205 960K task_struct 515512 99%1.38K1035 824K pirpIo 5084 2365 46%0.12K164 31 656K size-128 The largest area is about 100Mb, but on free there are over 3Gb on the cached column: [EMAIL PROTECTED] ~]$ free total used free sharedbuffers cached Mem: 51907364461420 729316 0 1418363265464 -/+ buffers/cache:10541204136616 Swap: 2048248 02048248 [EMAIL PROTECTED] ~]$ Regards, Luis Alexei_Roudnev [EMAIL PROTECTED] wrote: Did you run slabtop ? It can show unreleased buffers in the system. - Original Message - From: Luis Freitas To: Alexei_Roudnev ; Brian Sieler ; ocfs2-users@oss.oracle.com Sent: Monday, April 09, 2007 2:29 PM Subject: Re: [Ocfs2-users] High on buffers and deep on swap Alexei, Yes, it seems to have no effect, which too is very strange. On 2.4 vm.freepages had a very easy to notice effect. There are other people having problems with buffers not being released on the list and some of them are forcing the kernel cache to be flushed with: echo 3 /proc/sys/vm/drop_caches But I dont see this parameter on RHAS 4.0. Also, to be fair this seems to be a generic VM issue, I see this on servers that are not running ocfs2 too. And I only see this behavior on machines with more than 2Gb of memory. Regards, Luis Regards, Luis Alexei_Roudnev [EMAIL PROTECTED] wrote: v\:* { BEHAVIOR: url(#default#VML) } o\:* { BEHAVIOR: url(#default#VML) } w\:* { BEHAVIOR: url(#default#VML) } .shape { BEHAVIOR: url(#default#VML) } Did you tried vm.swappiness parameter? (/proc/sys/vm/swappiness) - Original Message - From: Brian Sieler To: 'Luis Freitas' ; ocfs2-users@oss.oracle.com Sent: Sunday, April 08, 2007 10:52 PM Subject: RE: [Ocfs2-users] High on buffers and deep on swap Luis, yes I am experiencing what appears to be a similar problem you are describing. See my post from just a few minutes ago on another thread. I run a 2-node cluster with OCFS2/RAC on 2.6.9-34.0.2.ELsmp (RHEL 4.0) as well. total used free sharedbuffers cached Mem: 40444964005516 38980 0 341082236636 -/+ buffers/cache:17347722309724 Swap: 2097144 6482441448900 If youve uncovered anything since posting this message, please pass it along? - From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Luis Freitas Sent: Friday, February 23, 2007 5:32 PM To: ocfs2-users@oss.oracle.com Subject: [Ocfs2-users] High on buffers and deep on swap Hi, This is a bit off topic, hope there is not a problem. Anyone out there experiencing high swapping with the kernel retaining a large amount of buffers? This used to be a problem on 2.4, and I usually changed /proc/sys/vm/freepages to fix it. But on 2.6 this parameter no longer exists. One of the servers here is holding over 3.5Gb of cache even when using over 700Mb of swap, and free memory is always low. [EMAIL PROTECTED] ~]$ free total used free sharedbuffers cached Mem: 51907364810880 379856 0 1430323583868 -/+ buffers/cache:10839804106756 Swap: 2048248
Re: [Ocfs2-users] ocfs2 cluster becomes unresponsive
Andy, I found helpfull to diagnose this kind of hang to keep a priority 0 shell opened on the server. This shell usually keeps working even during heavy swapping or other situations where the system becomes unresponsive. You can start one with this command: nice -n -20 bash From this you could run a top or a vmstat to see what is happening when the server is unresponsive. Just be careful to not run any command that might generate a large output or have high CPU usage, as you might hang the server yourself. Regards, Luis Andy Kipp [EMAIL PROTECTED] wrote: I checked bugzilla and what is happening is almost identical to bug #819. However, the dead node continues to heartbeat, yet is unresponsive. No log output at all is generated on the dead node. This has been happening for a few months however frequency is increasing. Is there any information I can provide to hopefully figure this out? - Andy -- Andrew Kipp Network Administrator Velcro USA Inc. Email: [EMAIL PROTECTED] Work: (603) 222-4844 CONFIDENTIALITY NOTICE: This email is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply e mail and destroy all copies of the original message. If you are the intended recipient but do not wish to receive communications through this medium, please so advise immediately. On 3/9/2007 at 9:39 PM, in message [EMAIL PROTECTED], Sunil Mushran wrote: File a bugzilla with the messages from all three nodes. Appears node 2 went down but kept heartbeating. Strange. The messages from node 2 may shed more light. Andy Kipp wrote: We are running OCFS2 on SLES9 machines using a FC SAN. Without warning both nodes will become unresponsive. Can not access either machine via ssh or terminal (hangs after typing in username). However the machine still responds to pings. This continues until one node is rebooted, at which time the second node resumes normal operations. I am not entirely sure that this is an OCFS2 problem at all however the syslog shows it had issues Here is the log from the node that was not rebooted. The node that was rebooted contained no log information. The system appeared to have gone down at about 3AM, until the node was rebooted at around 7:15. Mar 8 03:06:32 groupwise-1-mht kernel: o2net: connection to node groupwise-2-mht (num 2) at 192.168.1.3: has been idle for 10 seconds, shutting it down. Mar 8 03:06:32 groupwise-1-mht kernel: (0,2):o2net_idle_timer:1310 here are some times that might help debug the situation: (tmr 1173341182.367220 now 1173341192.367244 dr 1173341182.367213 adv 1173341182.367228:1173341182.367229 func (05ce6220:2) 1173341182.367221:1173341182.367224) Mar 8 03:06:32 groupwise-1-mht kernel: o2net: no longer connected to node groupwise-2-mht (num 2) at 192.168.1.3: Mar 8 03:06:32 groupwise-1-mht kernel: (499,0):dlm_do_master_request:1330 ERROR: link to 2 went down! Mar 8 03:06:32 groupwise-1-mht kernel: (499,0):dlm_get_lock_resource:914 ERROR: status = -112 Mar 8 03:13:02 groupwise-1-mht kernel: (8476,0):dlm_send_proxy_ast_msg:458 ERROR: status = -107 Mar 8 03:13:02 groupwise-1-mht kernel: (8476,0):dlm_flush_asts:607 ERROR: status = -107 Mar 8 03:19:54 groupwise-1-mht kernel: (147,1):dlm_send_remote_unlock_request:356 ERROR: status = -107 Mar 8 03:19:54 groupwise-1-mht last message repeated 127 times Mar 8 03:19:55 groupwise-1-mht kernel: (873,0):dlm_do_master_request:1330 ERROR: link to 2 went down! Mar 8 03:19:55 groupwise-1-mht kernel: (873,0):dlm_get_lock_resource:914 ERROR: status = -107 Mar 8 03:19:55 groupwise-1-mht kernel: (901,0):dlm_do_master_request:1330 ERROR: link to 2 went down! Mar 8 03:19:55 groupwise-1-mht kernel: (901,0):dlm_get_lock_resource:914 ERROR: status = -107 Mar 8 03:19:56 groupwise-1-mht kernel: (929,0):dlm_do_master_request:1330 ERROR: link to 2 went down! Mar 8 03:19:56 groupwise-1-mht kernel: (929,0):dlm_get_lock_resource:914 ERROR: status = -107 Mar 8 03:45:29 groupwise-1-mht -- MARK -- Mar 8 04:15:02 groupwise-1-mht kernel: (147,1):dlm_send_remote_unlock_request:356 ERROR: status = -107 Mar 8 04:15:03 groupwise-1-mht last message repeated 383 times Mar 8 06:27:54 groupwise-1-mht kernel: (147,1):dlm_send_remote_unlock_request:356 ERROR: status = -107 Mar 8 06:27:54 groupwise-1-mht last message repeated 127 times Mar 8 06:27:54 groupwise-1-mht kernel: (147,1):dlm_send_remote_unlock_request:356 ERROR: status = -107 Mar 8 06:27:54 groupwise-1-mht last message repeated 127 times Mar 8 06:35:48 groupwise-1-mht kernel: (8872,0):dlm_do_master_request:1330 ERROR: link to 2 went down! Mar 8 06:35:48 groupwise-1-mht kernel: (8872,0):dlm_get_lock_resource:914 ERROR: status = -107 Mar 8 06:52:45
Re: [Ocfs2-users] High on buffers and deep on swap
Hi, This particular server is 2.6.9-34.ELsmp (RedHat 4.0) as this is our production cluster and we did not update it recently. Regards, Luis Sunil Mushran [EMAIL PROTECTED] wrote: Hmmm... the last time I saw your numbers, ocfs2's foot print was 15M. You'll have to do better than that. Anycase, Luis problem is the relationship between swap and cached buffers. Which kernel is this? John Lange wrote: It seems that ocfs has an unfixed memory leak even in the most recent version. I hope to make a more detailed bug report on Monday. John On Fri, 2007-02-23 at 15:31 -0800, Luis Freitas wrote: Hi, This is a bit off topic, hope there is not a problem. Anyone out there experiencing high swapping with the kernel retaining a large amount of buffers? This used to be a problem on 2.4, and I usually changed /proc/sys/vm/freepages to fix it. But on 2.6 this parameter no longer exists. One of the servers here is holding over 3.5Gb of cache even when using over 700Mb of swap, and free memory is always low. [EMAIL PROTECTED] ~]$ free total used free shared buffers cached Mem: 5190736 4810880 379856 0 143032 3583868 -/+ buffers/cache: 1083980 4106756 Swap: 2048248 723064 1325184 [EMAIL PROTECTED] ~]$ I am tuning /proc/sys/vm/swappiness, but this seems to have no effect at all. Changed from 60 to 10 and seems to have no effect. The server runs Oracle RAC with OCFS2. Regards, Luis __ Food fight? Enjoy some healthy debate in the Yahoo! Answers Food Drink QA. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Be a PS3 game guru. Get your game face on with the latest PS3 news and previews at Yahoo! Games.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
[Ocfs2-users] High on buffers and deep on swap
Hi, This is a bit off topic, hope there is not a problem. Anyone out there experiencing high swapping with the kernel retaining a large amount of buffers? This used to be a problem on 2.4, and I usually changed /proc/sys/vm/freepages to fix it. But on 2.6 this parameter no longer exists. One of the servers here is holding over 3.5Gb of cache even when using over 700Mb of swap, and free memory is always low. [EMAIL PROTECTED] ~]$ free total used free sharedbuffers cached Mem: 51907364810880 379856 0 1430323583868 -/+ buffers/cache:10839804106756 Swap: 2048248 7230641325184 [EMAIL PROTECTED] ~]$ I am tuning /proc/sys/vm/swappiness, but this seems to have no effect at all. Changed from 60 to 10 and seems to have no effect. The server runs Oracle RAC with OCFS2. Regards, Luis - Food fight? Enjoy some healthy debate in the Yahoo! Answers Food Drink QA.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: Hmm, here is an example. Re: [Ocfs2-users] Also just a comment to theOracle guys
Alexei, I think you got a point too, maybe OCFS2 could behave like Netapp, and simply hang when there is a problem and leave the fencing for CRS or whathever other clusterware is in use. Anyone from Oracle got a opinion on this? Regards, Luis Alexei_Roudnev [EMAIL PROTECTED] wrote: Absolutely. I know how redo and RAC interacts, you are absolutely correct. Sometimes CSSD reboots one node and that's all - good luck. Sometimes OCFS reboots one node and CSSD reboots another node - bad luck. That's why it is important do not mix different cluster managers on the same servers, or at least allow them to interact and make similar decision _who is master today_ (so who will survive split-brain situation). RAC is little simple case because Oracle is usually primary service - so if it decide to reboot, it's reasonable decision. OCFSv2 is another story - sometimes it is _secondary_ service (for example, it is used for the backups only), and if it is secondary then it should better stop working then reboot. It reveals 2 big problems (both, Oracle and OCFSv2, are affected): - single interface heartbeat is not reliable. You CAN NOT build reliable cluster using single heartbeat channel. Classical clusters (Veritas VCS) uses 2 - 4 different heartbeat media (we use 2 independent Ethernet HUBS and 2 generic Ethernet LAN-s in Veritas, use 2 Ethernets + 1 Serial in Linux clusters, use 2 Ethernet + 1 Serial in Cisco PIX cluster, and so on). Both OCFS and Oracle RAC can not use more then one (to be correct, you can configure few interfaces for RAC interconnection in SPFile, but it wil not affect CSSD). In addition, OCFS defaults are very strange and unrealistic - Ethernet and FC can not guarantee heartbeat times better than 1 minute, in average (I mean - it case of any network reconfiguration heartbeat wil experience 30 - 50 seconds delay, so if you configure 12 seconds timeout /default in OCFSv2/ you are at least naive. - too easy _self fencing_. Just again, if OCFS node lost connection to the disk, it should not self fence - it can send data to another nodes (or request them from another nodes), it can unmount file system and try to remount it, it can release control and resume operations. Immediate fencing is necessary in SOME cases but not in all. If FS have not pending operations, then Fencing by reboot don't make much difference with just _remount_. It's not so simple as I explain here, but the truth is that fencing decisions are not flexible enough and decrease reliability dramatically (I posted a list of scenarios when fencing should not happen). IN addition, I noticed other problems with OCFSv2 too (such as excessive CPU usage in some cases). I use OCFSv2, even in production. But I do it with a grain of salt, have a backup plan _how to run without it_, and don't use it for heavily loaded file systems with million files (I use heartbeat, reiserfs and APC switch fencing - and 3 independent heartbeats, with 40 seconds timeout). For now, I had one glitch on OCFSv2 (when it remounted read only on one node) and that's all - no other problems in production (OCFSv2 is used during start/stops only, so it is safe). But I run stress tests in the lab, I am running it in the lab clusters now (including RAC), and conclusion is simple - as a cluster, it is not reliable; as a file system, it may have hidden bugs so be extra careful with it. PS. Good point - it improves every month. Some problems are in the past already. PPS. All this lab reboots have been caused by extremely heavy load or by hardware failures (simulated or real). It works better in real life. But my experience says me, that if I can break something in the lab in 3 days, it's a matter of few month, when it broke in production. - Original Message - From:LuisFreitas To: Ocfs2-users@oss.oracle.com Sent: Saturday, February 10, 2007 4:52PM Subject: Re: Hmm,here is an example. Re:[Ocfs2-users] Also just a comment to theOracle guys Alexei, Actually your log seems to show that CSSD (OracleCRS) rebooted the node before OCFS2 got a chance to do it. On a RAC cluster, if the interconnect isinterrupted, all the nodes hang until a split brain resolution iscomplete and the recovery of all the crashed nodes is completed. This isneeded because every read on a Oracle datablock needs a ping to the othernodes. The view of the data must be consistent, when one noderead a particular data block, the Oracle Database first ping the othernodes to ensure that they did not modify the block and still have not flushedit to disk. Another node may even forward a reply with the block,preventing the disk access (Cache Fusion). When a split brain occurs, there is the loss of theseblocks not flushed
Re: Hmm, here is an example. Re: [Ocfs2-users] Also just a comment to the Oracle guys
Alexei, Actually your log seems to show that CSSD (Oracle CRS) rebooted the node before OCFS2 got a chance to do it. On a RAC cluster, if the interconnect is interrupted, all the nodes hang until a split brain resolution is complete and the recovery of all the crashed nodes is completed. This is needed because every read on a Oracle datablock needs a ping to the other nodes. The view of the data must be consistent, when one node read a particular data block, the Oracle Database first ping the other nodes to ensure that they did not modify the block and still have not flushed it to disk. Another node may even forward a reply with the block, preventing the disk access (Cache Fusion). When a split brain occurs, there is the loss of these blocks not flushed to disk, and they are rebuilt using the redo threads of the particular nodes that crashed. During this interval all the database instances freeze, since before the node recovery is complete there is no way to guarantee that a block read from disk has not been altered on the crashed node. So the fencing is needed even if there is no disk activity, as the entire cluster becomes hang the moment the interconnect is down. And the timeout for the fencing must be as small as possible to prevent a long cluster reconfiguration delay. Of course the timeout must be tuned so as to be larger than ethernet switch failovers, or storage controller or disk multipath failovers. Or if possible the failover times should be reduced. Now, on the other hand, I am too having problems with OCFS2. It seems much less robust than ASM and the previous version, OCFS, specially under heavy disk activity. But I do expect these problems to get solved in the near future, as did the 2.4 kernel VM problems. Regards, Luis Alexei_Roudnev [EMAIL PROTECTED] wrote: Additional info - node had not ANY active OCFSv2 operations (OCFSv2 used for backups only and from another node only). So, if system just SUSPEND all FS operations and try to rejoin to the cluster, it all could work (moreover, connection to the disk system was intact, so it could close file sytem gracefully). It reveals 3 problems at once: - single heartbeat link (instead of multiple links) - timeout too short (ethernet can't guarantee 10 seconds, it can guarantee 1 minute minimum); - fencing even if system is passive and can remount / reconnect instead of rebooting. All we did in the lab was _disconnect 1 of trunks between switches for a few seconds, then insert it back into the socket_. No one other application failed (including heartbeat clusters). Database cluster was not doing anything on OCFS in time of failure (even backups). I will try heartbeat between loopback interfaces (and OCFS protocol) next time (I am just curios if it can provide 10 seconds for network reconfiguration). ... Feb 1 12:19:13 testrac12 kernel: o2net: connection to node testrac11 (num 0) at 10.254.32.111: has been idle for 10 seconds, shutting it down. Feb 1 12:19:13 testrac12 kernel: (13,3):o2net_idle_timer:1310 here are some times that might help debug the situation: (tmr 1170361135.521061 now 1170361145.520476 dr 1170361141.852795 adv 1170361135.521063:1170361135.521064 func (c4378452:505) 1170361067.762941:1170361067.762967) Feb 1 12:19:13 testrac12 kernel: o2net: no longer connected to node testrac11 (num 0) at 10.254.32.111: Feb 1 12:19:13 testrac12 kernel: (1855,3):dlm_send_remote_convert_request:398 ERROR: status = -107 Feb 1 12:19:13 testrac12 kernel: (1855,3):dlm_wait_for_node_death:371 5AECFF0BBCF74F069A3B8FF79F09FB5A: waiting 5000ms for notification of death of node 0 Feb 1 12:19:13 testrac12 kernel: (1855,1):dlm_send_remote_convert_request:398 ERROR: status = -107 Feb 1 12:19:13 testrac12 kernel: (1855,1):dlm_wait_for_node_death:371 5AECFF0BBCF74F069A3B8FF79F09FB5A: waiting 5000ms for notification of death of node 0 Feb 1 12:22:22 testrac12 kernel: (1855,2):dlm_send_remote_convert_request:398 ERROR: status = -107 Feb 1 12:22:22 testrac12 kernel: (1855,2):dlm_wait_for_node_death:371 5AECFF0BBCF74F069A3B8FF79F09FB5A: waiting 5000ms for notification of death of node 0 Feb 1 12:22:27 testrac12 kernel: (13,3):o2quo_make_decision:144 ERROR: fencing this node because it is connected to a half-quorum of 1 out of 2 nodes which doesn't include the lowest active node 0 Feb 1 12:22:27 testrac12 kernel: (13,3):o2hb_stop_all_regions:1889 ERROR: stopping heartbeat on all active regions. Feb 1 12:22:27 testrac12 kernel: Kernel panic: ocfs2 is very sorry to be fencing this system by panicing Feb 1 12:22:27 testrac12 kernel: Feb 1 12:22:28 testrac12 su: pam_unix2: session finished for user oracle, service su Feb 1 12:22:29 testrac12 logger: Oracle CSSD failure. Rebooting for cluster integrity. Feb 1 12:22:32 testrac12 su: pam_unix2: session finished for user
Re: [Ocfs2-users] Success!
Brandon, Can you post details about the Disk layout you are using and disk models? (RAID5/RAID10, 15K RPM or 10K RPM, 36Gb or 76Gb, how many disks on each raid group) Best Regards, Luis Brandon Lamb [EMAIL PROTECTED] wrote: Well today I did clean installs on 3 test machines. For the server I used the openfiler (latest from sourceforge iso) as the server os, using a 500 gig sata drive and exporting as iscsi. I then installed CentOS 4.4 on two servers, ran the yum updates, and then installed the ocfs2 kernel and tools, devel and debug rpms. After my previous experience i was able to get up and going very quickly since i didnt have to relearn how to do the iscsi stuff and edit the ocfs2 config. I havent had a chance to really test it yet as the machines just finished updating awhile ago but they can mount the drive, so I am excited at seeing some progress. One one node I was able to do a copy from /tmp (hda3) of a 771 meg maildir to the ocfs2 mount over iscsi. Took about 1 minute, using gige (no jumbo frames, just a cheap switch). I used the -T mail option for mkfs.ocfs2 and mounted with _netdev,nointr in fstab. I briefly went over our PRTG logs and it looks our current NFS server that stores our mail data moves about 114 gigabytes a day of traffic on our switch (total in and out). Doing simple math and not taking into account spurts its about 1.6 megabytes per second I think, so I am hoping this will work out quite well for us. One thing I did notice is that the server (running openfiler) went up to 7+ load. Im wondering/thinking of just reinstalling it with centos and isntalling the iscsi-target from source (newer version i think). This is a pentium d 3ghz, 8gig ram machine. I was supprised to see it go to 7 with just 1 node copying data to it. But this is an iscsi issue i assume so i'll mess with that plus network tuning. All in all, I am happy with the results today, mostly because I have had no crashes or errors or kernels panics/reboots or anything. ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - Have a burning question? Go to Yahoo! Answers and get answers from real people who know.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] expand ocfs2
Antonio, Last time I looked, about 6 months ago, LVM2 still wasnt cluster aware. So changing volume sizes on a cluster disk could lead to data corruption. At least a reboot of all involved nodes would be required after any LVM change, or some other procedure to force LVM to reread the volumes from disk on the remote nodes. With fdisk, you can force the other nodes to re-read the partition table from disk by running fdisk and forcing a write of the partition table on each node. Regards, Luis Antonio Trujillo Carmona [EMAIL PROTECTED] wrote: El mié, 24-01-2007 a las 16:49 +0100, Antonio Trujillo Carmona escribió: Hello I'm try to install a 2 node acces with a proliant and hba conect to a EVA. my doubt is. is posibol expand a ocfs2 filesystem? and, what is better, in order to use the multipath device, format /dev/sda or use lvm to create a logical volumen and format it? OK. I found the answer to my first cuestion in the FAQ http://oss.oracle.com/projects/ocfs2/dist/documentation/ocfs2_faq.html#RESIZE. What abauot the second one? any one can guide me to leading with multipath? -- Antonio Trujillo Carmona ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users
Re: [Ocfs2-users] Correction
Well, I dont know. But I have this issue also, OCFS2 is much slower than OCFS when doing file copies. Regards, Luis GOKHAN [EMAIL PROTECTED] wrote: Why db2 is too slow than db1 ? Anybody have idea ? Thx Luis 4 the correction.. Message: 4 Date: Tue, 16 Jan 2007 03:17:17 -0800 (PST) From: Luis Freitas [EMAIL PROTECTED] Subject: Re: [Ocfs2-users] ocfs Vs ocfs2 To: ocfs2-users@oss.oracle.com Message-ID: [EMAIL PROTECTED] Content-Type: text/plain; charset=iso-8859-1 Hi, It seems that you inverted the times on the chart. Regards, Luis GOKHAN [EMAIL PROTECTED] wrote: Hi everbody this is my first post, I have two test server .(Both of them is idle) db1 : RHEL4 OCFS2 db2 : RHEL3 OCFS I test the IO both of them The result is below. db1(Time Spend) db2(Time Spend) OS Test Commanddd (1GB) (Yazma) 0m0.796s 0m18.420s time dd if=/dev/zero of=./sill.t bs=1M count=1000dd (1GB) (Okuma) 0m0.241s 8m16.406s time dd of=/dev/zero if=./sill.t bs=1M count=1000cp (1GB) 0m0.986s 7m32.452s time cp sill.t sill2.t Why db1 is too slow than db2 ? Anybody have idea ? My production database is oracle 9205 on RHEL3 and OCFS I test the 9208 on RHEL4 and OCFS2 . I will try to decide to upgrade or not... Thx.. - Need Mail bonding? Go to the Yahoo! Mail QA for great tips from Yahoo! Answers users.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com -- next part -- An HTML attachment was scrubbed... URL: http://oss.oracle.com/pipermail/ocfs2-users/attachments/20070116/beac5eb0/attachment-0001.html -- Message: 5 Date: Tue, 16 Jan 2007 09:20:28 -0600 From: Brian Sieler [EMAIL PROTECTED] Subject: [Ocfs2-users] OCFS2 crash To: ocfs2-users@oss.oracle.com Message-ID: [EMAIL PROTECTED] Content-Type: text/plain; charset=WINDOWS-1252; format=flowed Using 2-node clustered file system on DELL/EMC SAN/RHEL 2.6.9-34.0.2.ELsmp x86_64. Config: O2CB_HEARTBEAT_THRESHOLD=30 Kernel param: elavator=deadline (per FAQ) These log items appear and the server crashes. Has happened twice now at three week intervals, each time during a heavy IO operation: Jan 15 16:08:29 db100 kernel: (3898,6):o2hb_setup_one_bio:371 ERROR: Could not alloc slots BIO! Jan 15 16:08:29 db100 kernel: (3898,6):o2hb_read_slots:507 ERROR: status = -12 Jan 15 16:08:29 db100 kernel: (3898,6):o2hb_do_disk_heartbeat:973 ERROR: status = -12 Jan 15 16:08:29 db100 kernel: (3898,6):o2hb_setup_one_bio:371 ERROR: Could not alloc slots BIO! Jan 15 16:08:29 db100 kernel: (3898,6):o2hb_read_slots:507 ERROR: status = -12 Jan 15 16:08:29 db100 kernel: (3898,6):o2hb_do_disk_heartbeat:973 ERROR: status Can't find much on any of these errors what is 507 ERROR status = -12? Any help appreciated -- Brian -- ___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users End of Ocfs2-users Digest, Vol 37, Issue 12 *** - Don't pick lemons. See all the new 2007 cars at Yahoo! Autos.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users - The fish are biting. Get more visitors on your site using Yahoo! Search Marketing.___ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-users