Re: [ovirt-users] Fwd: RE: FC QLogic question
Hi, Sorry, I got waylaid at the end of yesterday - catching up... (sorry for length of thread now as well, but would like to get to the bottom of this) Have seen posts about ovirt-node... I personally like the idea and hope that it continues to develop. Having to load a Linux distro up and then put ovirt hypervisor on it is more work and then you need to worry about upgrading the linux distro and ovirt and suddenly ovirt doesn't run on the distro you have and you start over. IMHO anyway. All of the LUN names are listed in /dev/mapper /etc/multipath.conf is thus # RHEV REVISION 1.0 defaults { polling_interval5 getuid_callout /sbin/scsi_id --whitelisted --replace-whitespace --device=/dev/%n no_path_retry fail user_friendly_names no flush_on_last_del yes fast_io_fail_tmo5 dev_loss_tmo30 max_fds 4096 } devices { device { vendor HITACHI product DF.* getuid_callout /sbin/scsi_id --whitelisted --replace-whitespace --device=/dev/%n } device { vendor COMPELNT product Compellent Vol no_path_retry fail } } But on the multipath -ll round robin is enabled... I can't see how the firmware missing can be a problem if all my disks are appearing on node. Some cards don't need to be flashed, they can be if needed but if my LUNs are appearing, then I assume it is working. Sorry, I am very new to node and not much experience with FC under Linux, so trying to get up to speed on commands... Which ovirt am Iusing? Not sure what you want back, uname -a yields 2.6.32-431.3.1.el6.x86_6 The iso I used is ovirt-node-iso-3.4-20140423.0.el6.iso from the releases/stable/iso folder - I am fairly sure, there are two ISOs there. Unfortunately I am not in the office today and can't get to the desktop where I burned the ISOs from... Thanks Bill -Original Message- From: Fabian Deutsch [mailto:fdeut...@redhat.com] Sent: 11 September 2014 13:40 To: Daniel Helgenberger; Bill Dossett Cc: users@ovirt.org Subject: Re: [ovirt-users] Fwd: RE: FC QLogic question - Original Message - On 11.09.2014 14:23, Fabian Deutsch wrote: Hey Daniel, - Original Message - On 11.09.2014 14:09, Fabian Deutsch wrote: Hey Bill and Daniel, it would indeed be interesting - mandatory - to know if it's qla* or qxl (qxl is IMHO related to graphics). Once we know that I'll see if the bug below fixes it. Hello Fabian, indeed this was my mistake; while typing the mail I was reading on the spice website - so qla is the correct module of course! I already replied this to Bill but forgot to cc the list also ... Also, for the node iso it would make little sense to have the qlx module loaded. For completeness, I attach my answer to Bill. Thanks :) So to summarize: The adapter is recognized on Node? But it does not show up in Engine? Bill has to tell, but I think so. I suggested to him this might be Ah, I mixed you an Bill up in the thread. because of dm-multipath, multipath.conf might blacklist here. My nodes are plain EL6 and I have multipath enabled - but I do not know if ovirt-node filters for dm-mp devnodes only by default (this would make sense because of iscsi-multipathing in ovirt 3.4). Well, we did some multipath changes lately. Let#s see what Bill says: Bill, what node are you using? What does multipath -l say? How does /dev/mapper/* and /dev/multipath.conf look like? And is the disk appearing on Node and/or Engine? Thanks! fabian Maybe you can help out on this topic? Thanks! Thanks - fabian I'd need your help to verify that we are also including all relevant dependencies. Thanks - fabian - Original Message - On 11.09.2014 13:05, Bill Dossett wrote: Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one Indeed, sorry for that. lsscsi does list all of the LUNs... so that's working! Great! But nothing in ovirt engine when I try to add a storage domain. Please check back in ovirt engine, with Hosts / Hardware info. You should see the adapter there, along with WWN's ect. I am not quite sure, but AFAIK the node uses dm-multipath by default (for iscsi multipathing). So what does: # multipath -ll list? Thanks again, I guess I will try and get a bit further with the installs I have... Bill Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one lsscsi does list all of the LUNs... so that's working! But nothing in ovirt engine when I try to add a storage domain
Re: [ovirt-users] Fwd: RE: FC QLogic question
This could also be an selinux issue with FC on the node, I've hit that one before. Especially since the luns do show up with multipath -ll Could be as simple as flipping a sanlock related selinux boolean. A quick test would be to boot in permissive mode. See http://www.ovirt.org/Node_Troubleshooting#SELinux for some more information. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] Fwd: RE: FC QLogic question
Hmmm… I’ve just gone to add a storage domain and in the admin portal and bingo bango, all my LUNs appeared. Just wiping out everything from my last lab now. Thanks everyone, I am not quite sure why it started working, but knowing that the ovirt-node was seeing the LUNs prompted me to continue. Thanks all for the help. Bill From: sander.grendel...@gmail.com [mailto:sander.grendel...@gmail.com] On Behalf Of Sander Grendelman Sent: 12 September 2014 08:28 To: Bill Dossett Cc: users@ovirt.org Subject: Re: [ovirt-users] Fwd: RE: FC QLogic question This could also be an selinux issue with FC on the node, I've hit that one before. Especially since the luns do show up with multipath -ll Could be as simple as flipping a sanlock related selinux boolean. A quick test would be to boot in permissive mode. See http://www.ovirt.org/Node_Troubleshooting#SELinux for some more information. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] Fwd: RE: FC QLogic question
Bugger, spoke to soon…. One host sees the LUNs the other doesn’t. Same builds of ovirt-node. Both dell 2950s, and both qlogic cards… but one says Sep 9 08:12:23 localhost kernel: qla2xxx [:0c:00.0]-00fb:1: QLogic QLE2460 - QLogic 4GB FC Single-Port PCI-E HBA for IBM System x. And further down the log this (this one does not find the LUNS) Sep 9 08:12:24 localhost kernel: qla2xxx [:0c:00.0]-d001:1: Firmware dump saved to temp buffer (1/c90013d9e000). Sep 9 08:12:24 localhost kernel: qla2xxx [:0c:00.0]-101e:1: Mailbox cmd timeout occurred, cmd=0x69, mb[0]=0x69. Scheduling ISP abort Sep 9 08:12:24 localhost kernel: qla2xxx [:0c:00.0]-00af:1: Performing ISP error recovery - ha=88086ce53000. Sep 9 08:12:24 localhost kernel: qla2xxx :0c:00.0: firmware: requesting ql2400_fw.bin Sep 9 08:12:24 localhost kernel: qla2xxx [:0c:00.0]-0063:1: Failed to load firmware image (ql2400_fw.bin). Sep 9 08:12:24 localhost kernel: qla2xxx [:0c:00.0]-0090:1: Fimware image unavailable. Sep 9 08:12:24 localhost kernel: qla2xxx [:0c:00.0]-0091:1: Firmware images can be retrieved from: http://ldriver.qlogic.com/firmware/. Sep 9 08:12:24 localhost firmware.sh[4794]: Cannot find firmware file 'ql2400_fw.bin' Sep 9 08:12:26 localhost kernel: qla2xxx [:0c:00.0]-505f:1: Link is operational (4 Gbps). Sep 9 08:12:26 localhost kernel: qla2xxx [:0c:00.0]-1020:1: Failed mbx[0]=69, mb[1]=0, mb[2]=0, mb[3]=0, cmd=69 . Sep 9 08:12:32 localhost kdump: mkdumprd: failed to make kdump initrd Sep 9 08:12:32 localhost lldpad: config file failed to load, And the one that does find the LUNS is Sep 11 08:33:39 localhost kernel: qla2xxx [:0a:00.0]-00fb:0: QLogic QLE2460 - SG-(X)PCIE1FC-QF4, Sun StorageTek 4 Gb FC Enterprise PCI-Express Single Channel. And a little further down it discovers the LUNs Sep 11 08:33:39 localhost kernel: qla2xxx [:0e:00.0]-505f:1: Link is operational (4 Gbps). Sep 11 08:33:39 localhost kernel: scsi 1:0:0:0: Direct-Access DGC LUNZ 0216 PQ: 0 ANSI: 4 Sep 11 08:33:39 localhost kernel: scsi 1:0:0:1: Direct-Access DGC RAID 5 0216 PQ: 0 ANSI: 4 Sep 11 08:33:39 localhost kernel: scsi 1:0:0:2: Direct-Access DGC RAID 5 0216 PQ: 0 ANSI: 4 Sep 11 08:33:39 localhost kernel: scsi 1:0:0:3: Direct-Access DGC RAID 5 0216 PQ: 0 ANSI: 4 Sep 11 08:33:39 localhost kernel: scsi 1:0:0:5: Direct-Access DGC RAID 5 0216 PQ: 0 ANSI: 4 Sep 11 08:33:39 localhost kernel: scsi 1:0:0:11: Direct-Access DGC RAID 1 0216 PQ: 0 ANSI: 4 Sep 11 08:33:39 localhost kernel: scsi 1:0:0:51: Direct-Access DGC RAID 1 0216 PQ: 0 ANSI: 4 So, I am thinking maybe different firmware levels on these cards and the first one for IBM systems has different firmware? As it seems to be throwing more errors. Should I try to add the firmware? And if so I know it goes in /lib/firmware but then the boot image needs to be updated I think and I don’t know how to do that… or am I barking up the wrong tree. Thanks again for the help. From: users-boun...@ovirt.org [mailto:users-boun...@ovirt.org] On Behalf Of Bill Dossett Sent: 12 September 2014 08:35 To: Sander Grendelman Cc: users@ovirt.org Subject: Re: [ovirt-users] Fwd: RE: FC QLogic question Hmmm… I’ve just gone to add a storage domain and in the admin portal and bingo bango, all my LUNs appeared. Just wiping out everything from my last lab now. Thanks everyone, I am not quite sure why it started working, but knowing that the ovirt-node was seeing the LUNs prompted me to continue. Thanks all for the help. Bill From: sander.grendel...@gmail.commailto:sander.grendel...@gmail.com [mailto:sander.grendel...@gmail.com] On Behalf Of Sander Grendelman Sent: 12 September 2014 08:28 To: Bill Dossett Cc: users@ovirt.orgmailto:users@ovirt.org Subject: Re: [ovirt-users] Fwd: RE: FC QLogic question This could also be an selinux issue with FC on the node, I've hit that one before. Especially since the luns do show up with multipath -ll Could be as simple as flipping a sanlock related selinux boolean. A quick test would be to boot in permissive mode. See http://www.ovirt.org/Node_Troubleshooting#SELinux for some more information. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] Fwd: RE: FC QLogic question
Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one lsscsi does list all of the LUNs... so that's working! But nothing in ovirt engine when I try to add a storage domain. Thanks again, I guess I will try and get a bit further with the installs I have... Bill -Original Message- From: Daniel Helgenberger [mailto:daniel.helgenber...@m-box.de] Sent: 11 September 2014 11:53 To: Bill Dossett Subject: Re: [ovirt-users] FC QLogic question On 11.09.2014 12:17, Bill Dossett wrote: Thanks for responding Daniel. Never mind! I was wondering and thought the message may be irrelevant. But yes, it doesn't appear to be working. When I try to add a FC storage domain, no hosts are showing in the drop down box of hosts. I have loaded the latests ovirtnode (3.4.something...) on two hosts now and neither are working. Lets first try to isolate your FC issure. So you are using the latest ovirt-node? EL6? Keep in mind, the kernel is standard Linux and therefore should support Qlogic adapters out of the box (IMHO they are the best supported HBAs in Linux). Please try a: # lsmod | grep qxl to see whatever the qxl module is loaded. Keep in mind you can run into problems with other conrollers in your host, though I only had this with LSI SAS / FC controllers so far. Further check on the FC switch if the N port is logged in and the zoning is correct. Finally, check with # lsscsi if you can see the luns. Also having a lot of issues with configuring the network. I configure an interface I can ping it, but it shows as unconfigured... on one host I have added it to the manager and I see the interfaces anyway, but on the second host I configure the eth0, I can ping it, it added to the ovirt engine but it shows no NICs. Should I be using an earlier version maybe? thanks again and best regards. *Bill Dossett* *Virtualisation Architect* *Pitney Bowes - Tech Central*** 6 Hercules Way, 600 Leavesden Park Watford, WD25 7GS Phone(+44) (0)1923 279353 | VOIP: 54739353 Mobile(+44) (0)7775 908612 bill.doss...@pb.com mailto:bill.doss...@pb.com www.pb.com http://www.pb.com/ *In Engineering?**Raise a ticket via Remedy Anywhere [HERE http://teamsite.inside.pb.com/sites/TCPPM/GES/Lists/Create%20Incident /newIncident.aspx]takes less than a minute* *ManageIQ User Guide available* [*HERE https://dl.dropbox.com/u/6219441/ManageIQ%20-%20User%20Manual%20PB%20 v4.pdf*] *Check out the ManageIQ Blog for recent enhancements and information [HERE] http://teamsite.inside.pb.com/sites/TCPPM/GES/manageiq/default.aspx* On Sep 11, 2014, at 10:52 AM, Daniel Helgenberger daniel.helgenber...@m-box.de mailto:daniel.helgenber...@m-box.de wrote: Hello Bill, what are your problems acually? Is fibre channel not working (as expected)? The log snippet you provided is completly normal and the messages can safely be ignored. This is also unrelated to oVirt but rather CentOS (or the distro you are running) I suppose. Cheers! On 11.09.2014 10:51, Bill Dossett wrote: Hi, I seem to have run into a problem setting up ovirt with Fibre Channel storage. I have a 2MBs SAN and qlogic cards in the nodes. I used this same setup with RHEV about 6 months ago and it just worked, storage was there no problems. Now booting with ovirtnode distro, I am seeing Sep 9 08:01:27 localhost kernel: qla2xxx [:0c:00.0]-0091:1: Firmware images can be retrieved from: http://ldriver.qlogic.com/firmware/. Sep 9 08:01:27 localhost kernel: scsi1 : qla2xxx Sep 9 08:01:27 localhost kernel: qla2xxx [:0c:00.0]-00fb:1: QLogic QLE2460 - QLogic 4GB FC Single-Port PCI-E HBA for IBM System x. Sep 9 08:01:27 localhost kernel: qla2xxx [:0c:00.0]-00fc:1: ISP2432: PCIe (2.5GT/s x4) @ :0c:00.0 hdma+ host#=1 fw=4.00.26 (2). Sep 9 08:01:27 localhost kernel: ata_piix :00:1f.1: PCI INT A - GSI 16 (level, low) - IRQ 16 I have googled the firmware error and it says some cards need a firmware flash and some don't. These are pretty standard HBAs so I would think they would be supported out of the box, but maybe not. Looking at trying to add the firmware looked kind of complex - if I need to do this, does anyone know of a nice and easy howto guide on this. I'm just getting started with ovirt and appreciate any help Thanks *Bill Dossett* *Virtualisation Architect* *Pitney Bowes - Tech Central*** 6 Hercules Way, 600 Leavesden Park Watford, WD25 7GS Phone(+44) (0)1923 279353 | VOIP: 54739353 Mobile(+44) (0)7775 908612 bill.doss...@pb.com mailto:bill.doss...@pb.com mailto:bill.doss...@pb.com www.pb.com http://www.pb.com http://www.pb.com/ *In Engineering?**Raise a ticket via Remedy Anywhere [HERE
Re: [ovirt-users] Fwd: RE: FC QLogic question
Hey Bill and Daniel, it would indeed be interesting - mandatory - to know if it's qla* or qxl (qxl is IMHO related to graphics). Once we know that I'll see if the bug below fixes it. I'd need your help to verify that we are also including all relevant dependencies. Thanks - fabian - Original Message - Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one lsscsi does list all of the LUNs... so that's working! But nothing in ovirt engine when I try to add a storage domain. Thanks again, I guess I will try and get a bit further with the installs I have... Bill -Original Message- From: Daniel Helgenberger [mailto:daniel.helgenber...@m-box.de] Sent: 11 September 2014 11:53 To: Bill Dossett Subject: Re: [ovirt-users] FC QLogic question On 11.09.2014 12:17, Bill Dossett wrote: Thanks for responding Daniel. Never mind! I was wondering and thought the message may be irrelevant. But yes, it doesn't appear to be working. When I try to add a FC storage domain, no hosts are showing in the drop down box of hosts. I have loaded the latests ovirtnode (3.4.something...) on two hosts now and neither are working. Lets first try to isolate your FC issure. So you are using the latest ovirt-node? EL6? Keep in mind, the kernel is standard Linux and therefore should support Qlogic adapters out of the box (IMHO they are the best supported HBAs in Linux). Please try a: # lsmod | grep qxl to see whatever the qxl module is loaded. Keep in mind you can run into problems with other conrollers in your host, though I only had this with LSI SAS / FC controllers so far. Further check on the FC switch if the N port is logged in and the zoning is correct. Finally, check with # lsscsi if you can see the luns. Also having a lot of issues with configuring the network. I configure an interface I can ping it, but it shows as unconfigured... on one host I have added it to the manager and I see the interfaces anyway, but on the second host I configure the eth0, I can ping it, it added to the ovirt engine but it shows no NICs. Should I be using an earlier version maybe? thanks again and best regards. *Bill Dossett* *Virtualisation Architect* *Pitney Bowes - Tech Central*** 6 Hercules Way, 600 Leavesden Park Watford, WD25 7GS Phone(+44) (0)1923 279353 | VOIP: 54739353 Mobile(+44) (0)7775 908612 bill.doss...@pb.com mailto:bill.doss...@pb.com www.pb.com http://www.pb.com/ *In Engineering?**Raise a ticket via Remedy Anywhere [HERE http://teamsite.inside.pb.com/sites/TCPPM/GES/Lists/Create%20Incident /newIncident.aspx]takes less than a minute* *ManageIQ User Guide available* [*HERE https://dl.dropbox.com/u/6219441/ManageIQ%20-%20User%20Manual%20PB%20 v4.pdf*] *Check out the ManageIQ Blog for recent enhancements and information [HERE] http://teamsite.inside.pb.com/sites/TCPPM/GES/manageiq/default.aspx* On Sep 11, 2014, at 10:52 AM, Daniel Helgenberger daniel.helgenber...@m-box.de mailto:daniel.helgenber...@m-box.de wrote: Hello Bill, what are your problems acually? Is fibre channel not working (as expected)? The log snippet you provided is completly normal and the messages can safely be ignored. This is also unrelated to oVirt but rather CentOS (or the distro you are running) I suppose. Cheers! On 11.09.2014 10:51, Bill Dossett wrote: Hi, I seem to have run into a problem setting up ovirt with Fibre Channel storage. I have a 2MBs SAN and qlogic cards in the nodes. I used this same setup with RHEV about 6 months ago and it just worked, storage was there no problems. Now booting with ovirtnode distro, I am seeing Sep 9 08:01:27 localhost kernel: qla2xxx [:0c:00.0]-0091:1: Firmware images can be retrieved from: http://ldriver.qlogic.com/firmware/. Sep 9 08:01:27 localhost kernel: scsi1 : qla2xxx Sep 9 08:01:27 localhost kernel: qla2xxx [:0c:00.0]-00fb:1: QLogic QLE2460 - QLogic 4GB FC Single-Port PCI-E HBA for IBM System x. Sep 9 08:01:27 localhost kernel: qla2xxx [:0c:00.0]-00fc:1: ISP2432: PCIe (2.5GT/s x4) @ :0c:00.0 hdma+ host#=1 fw=4.00.26 (2). Sep 9 08:01:27 localhost kernel: ata_piix :00:1f.1: PCI INT A - GSI 16 (level, low) - IRQ 16 I have googled the firmware error and it says some cards need a firmware flash and some don't. These are pretty standard HBAs so I would think they would be supported out of the box, but maybe not. Looking at trying to add the firmware looked kind of complex - if I need to do this, does anyone know of a nice and easy howto guide on this. I'm just getting started with ovirt and appreciate any help Thanks *Bill
Re: [ovirt-users] Fwd: RE: FC QLogic question
Hey Daniel, - Original Message - On 11.09.2014 14:09, Fabian Deutsch wrote: Hey Bill and Daniel, it would indeed be interesting - mandatory - to know if it's qla* or qxl (qxl is IMHO related to graphics). Once we know that I'll see if the bug below fixes it. Hello Fabian, indeed this was my mistake; while typing the mail I was reading on the spice website - so qla is the correct module of course! I already replied this to Bill but forgot to cc the list also ... Also, for the node iso it would make little sense to have the qlx module loaded. For completeness, I attach my answer to Bill. Thanks :) So to summarize: The adapter is recognized on Node? But it does not show up in Engine? Thanks - fabian I'd need your help to verify that we are also including all relevant dependencies. Thanks - fabian - Original Message - On 11.09.2014 13:05, Bill Dossett wrote: Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one Indeed, sorry for that. lsscsi does list all of the LUNs... so that's working! Great! But nothing in ovirt engine when I try to add a storage domain. Please check back in ovirt engine, with Hosts / Hardware info. You should see the adapter there, along with WWN's ect. I am not quite sure, but AFAIK the node uses dm-multipath by default (for iscsi multipathing). So what does: # multipath -ll list? Thanks again, I guess I will try and get a bit further with the installs I have... Bill Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one lsscsi does list all of the LUNs... so that's working! But nothing in ovirt engine when I try to add a storage domain. Thanks again, I guess I will try and get a bit further with the installs I have... Bill -Original Message- From: Daniel Helgenberger [mailto:daniel.helgenber...@m-box.de] Sent: 11 September 2014 11:53 To: Bill Dossett Subject: Re: [ovirt-users] FC QLogic question On 11.09.2014 12:17, Bill Dossett wrote: Thanks for responding Daniel. Never mind! I was wondering and thought the message may be irrelevant. But yes, it doesn't appear to be working. When I try to add a FC storage domain, no hosts are showing in the drop down box of hosts. I have loaded the latests ovirtnode (3.4.something...) on two hosts now and neither are working. Lets first try to isolate your FC issure. So you are using the latest ovirt-node? EL6? Keep in mind, the kernel is standard Linux and therefore should support Qlogic adapters out of the box (IMHO they are the best supported HBAs in Linux). Please try a: # lsmod | grep qxl to see whatever the qxl module is loaded. Keep in mind you can run into problems with other conrollers in your host, though I only had this with LSI SAS / FC controllers so far. Further check on the FC switch if the N port is logged in and the zoning is correct. Finally, check with # lsscsi if you can see the luns. Also having a lot of issues with configuring the network. I configure an interface I can ping it, but it shows as unconfigured... on one host I have added it to the manager and I see the interfaces anyway, but on the second host I configure the eth0, I can ping it, it added to the ovirt engine but it shows no NICs. Should I be using an earlier version maybe? thanks again and best regards. *Bill Dossett* *Virtualisation Architect* *Pitney Bowes - Tech Central*** 6 Hercules Way, 600 Leavesden Park Watford, WD25 7GS Phone(+44) (0)1923 279353 | VOIP: 54739353 Mobile(+44) (0)7775 908612 bill.doss...@pb.com mailto:bill.doss...@pb.com www.pb.com http://www.pb.com/ *In Engineering?**Raise a ticket via Remedy Anywhere [HERE http://teamsite.inside.pb.com/sites/TCPPM/GES/Lists/Create%20Incident /newIncident.aspx]takes less than a minute* *ManageIQ User Guide available* [*HERE https://dl.dropbox.com/u/6219441/ManageIQ%20-%20User%20Manual%20PB%20 v4.pdf*] *Check out the ManageIQ Blog for recent enhancements and information [HERE] http://teamsite.inside.pb.com/sites/TCPPM/GES/manageiq/default.aspx* On Sep 11, 2014, at 10:52 AM, Daniel Helgenberger daniel.helgenber...@m-box.de mailto:daniel.helgenber...@m-box.de wrote: Hello Bill, what are your problems acually? Is fibre channel not working (as expected)? The log snippet you provided is completly normal and the messages can safely be ignored. This is also unrelated to oVirt but rather CentOS (or the distro you are running) I suppose. Cheers! On 11.09.2014 10:51, Bill Dossett wrote: Hi, I seem
Re: [ovirt-users] Fwd: RE: FC QLogic question
On 11.09.2014 14:23, Fabian Deutsch wrote: Hey Daniel, - Original Message - On 11.09.2014 14:09, Fabian Deutsch wrote: Hey Bill and Daniel, it would indeed be interesting - mandatory - to know if it's qla* or qxl (qxl is IMHO related to graphics). Once we know that I'll see if the bug below fixes it. Hello Fabian, indeed this was my mistake; while typing the mail I was reading on the spice website - so qla is the correct module of course! I already replied this to Bill but forgot to cc the list also ... Also, for the node iso it would make little sense to have the qlx module loaded. For completeness, I attach my answer to Bill. Thanks :) So to summarize: The adapter is recognized on Node? But it does not show up in Engine? Bill has to tell, but I think so. I suggested to him this might be because of dm-multipath, multipath.conf might blacklist here. My nodes are plain EL6 and I have multipath enabled - but I do not know if ovirt-node filters for dm-mp devnodes only by default (this would make sense because of iscsi-multipathing in ovirt 3.4). Maybe you can help out on this topic? Thanks! Thanks - fabian I'd need your help to verify that we are also including all relevant dependencies. Thanks - fabian - Original Message - On 11.09.2014 13:05, Bill Dossett wrote: Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one Indeed, sorry for that. lsscsi does list all of the LUNs... so that's working! Great! But nothing in ovirt engine when I try to add a storage domain. Please check back in ovirt engine, with Hosts / Hardware info. You should see the adapter there, along with WWN's ect. I am not quite sure, but AFAIK the node uses dm-multipath by default (for iscsi multipathing). So what does: # multipath -ll list? Thanks again, I guess I will try and get a bit further with the installs I have... Bill Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one lsscsi does list all of the LUNs... so that's working! But nothing in ovirt engine when I try to add a storage domain. Thanks again, I guess I will try and get a bit further with the installs I have... Bill -Original Message- From: Daniel Helgenberger [mailto:daniel.helgenber...@m-box.de] Sent: 11 September 2014 11:53 To: Bill Dossett Subject: Re: [ovirt-users] FC QLogic question On 11.09.2014 12:17, Bill Dossett wrote: Thanks for responding Daniel. Never mind! I was wondering and thought the message may be irrelevant. But yes, it doesn't appear to be working. When I try to add a FC storage domain, no hosts are showing in the drop down box of hosts. I have loaded the latests ovirtnode (3.4.something...) on two hosts now and neither are working. Lets first try to isolate your FC issure. So you are using the latest ovirt-node? EL6? Keep in mind, the kernel is standard Linux and therefore should support Qlogic adapters out of the box (IMHO they are the best supported HBAs in Linux). Please try a: # lsmod | grep qxl to see whatever the qxl module is loaded. Keep in mind you can run into problems with other conrollers in your host, though I only had this with LSI SAS / FC controllers so far. Further check on the FC switch if the N port is logged in and the zoning is correct. Finally, check with # lsscsi if you can see the luns. Also having a lot of issues with configuring the network. I configure an interface I can ping it, but it shows as unconfigured... on one host I have added it to the manager and I see the interfaces anyway, but on the second host I configure the eth0, I can ping it, it added to the ovirt engine but it shows no NICs. Should I be using an earlier version maybe? thanks again and best regards. *Bill Dossett* *Virtualisation Architect* *Pitney Bowes - Tech Central*** 6 Hercules Way, 600 Leavesden Park Watford, WD25 7GS Phone(+44) (0)1923 279353 | VOIP: 54739353 Mobile(+44) (0)7775 908612 bill.doss...@pb.com mailto:bill.doss...@pb.com www.pb.com http://www.pb.com/ *In Engineering?**Raise a ticket via Remedy Anywhere [HERE http://teamsite.inside.pb.com/sites/TCPPM/GES/Lists/Create%20Incident /newIncident.aspx]takes less than a minute* *ManageIQ User Guide available* [*HERE https://dl.dropbox.com/u/6219441/ManageIQ%20-%20User%20Manual%20PB%20 v4.pdf*] *Check out the ManageIQ Blog for recent enhancements and information [HERE] http://teamsite.inside.pb.com/sites/TCPPM/GES/manageiq/default.aspx* On Sep 11, 2014, at 10:52 AM, Daniel Helgenberger daniel.helgenber...@m-box.de mailto:daniel.helgenber...@m-box.de wrote: Hello Bill, what are your problems acually? Is fibre channel not working
Re: [ovirt-users] Fwd: RE: FC QLogic question
- Original Message - On 11.09.2014 14:23, Fabian Deutsch wrote: Hey Daniel, - Original Message - On 11.09.2014 14:09, Fabian Deutsch wrote: Hey Bill and Daniel, it would indeed be interesting - mandatory - to know if it's qla* or qxl (qxl is IMHO related to graphics). Once we know that I'll see if the bug below fixes it. Hello Fabian, indeed this was my mistake; while typing the mail I was reading on the spice website - so qla is the correct module of course! I already replied this to Bill but forgot to cc the list also ... Also, for the node iso it would make little sense to have the qlx module loaded. For completeness, I attach my answer to Bill. Thanks :) So to summarize: The adapter is recognized on Node? But it does not show up in Engine? Bill has to tell, but I think so. I suggested to him this might be Ah, I mixed you an Bill up in the thread. because of dm-multipath, multipath.conf might blacklist here. My nodes are plain EL6 and I have multipath enabled - but I do not know if ovirt-node filters for dm-mp devnodes only by default (this would make sense because of iscsi-multipathing in ovirt 3.4). Well, we did some multipath changes lately. Let#s see what Bill says: Bill, what node are you using? What does multipath -l say? How does /dev/mapper/* and /dev/multipath.conf look like? And is the disk appearing on Node and/or Engine? Thanks! fabian Maybe you can help out on this topic? Thanks! Thanks - fabian I'd need your help to verify that we are also including all relevant dependencies. Thanks - fabian - Original Message - On 11.09.2014 13:05, Bill Dossett wrote: Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one Indeed, sorry for that. lsscsi does list all of the LUNs... so that's working! Great! But nothing in ovirt engine when I try to add a storage domain. Please check back in ovirt engine, with Hosts / Hardware info. You should see the adapter there, along with WWN's ect. I am not quite sure, but AFAIK the node uses dm-multipath by default (for iscsi multipathing). So what does: # multipath -ll list? Thanks again, I guess I will try and get a bit further with the installs I have... Bill Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one lsscsi does list all of the LUNs... so that's working! But nothing in ovirt engine when I try to add a storage domain. Thanks again, I guess I will try and get a bit further with the installs I have... Bill -Original Message- From: Daniel Helgenberger [mailto:daniel.helgenber...@m-box.de] Sent: 11 September 2014 11:53 To: Bill Dossett Subject: Re: [ovirt-users] FC QLogic question On 11.09.2014 12:17, Bill Dossett wrote: Thanks for responding Daniel. Never mind! I was wondering and thought the message may be irrelevant. But yes, it doesn't appear to be working. When I try to add a FC storage domain, no hosts are showing in the drop down box of hosts. I have loaded the latests ovirtnode (3.4.something...) on two hosts now and neither are working. Lets first try to isolate your FC issure. So you are using the latest ovirt-node? EL6? Keep in mind, the kernel is standard Linux and therefore should support Qlogic adapters out of the box (IMHO they are the best supported HBAs in Linux). Please try a: # lsmod | grep qxl to see whatever the qxl module is loaded. Keep in mind you can run into problems with other conrollers in your host, though I only had this with LSI SAS / FC controllers so far. Further check on the FC switch if the N port is logged in and the zoning is correct. Finally, check with # lsscsi if you can see the luns. Also having a lot of issues with configuring the network. I configure an interface I can ping it, but it shows as unconfigured... on one host I have added it to the manager and I see the interfaces anyway, but on the second host I configure the eth0, I can ping it, it added to the ovirt engine but it shows no NICs. Should I be using an earlier version maybe? thanks again and best regards. *Bill Dossett* *Virtualisation Architect* *Pitney Bowes - Tech Central*** 6 Hercules Way, 600 Leavesden Park Watford, WD25 7GS Phone(+44) (0)1923 279353 | VOIP: 54739353 Mobile(+44) (0)7775 908612 bill.doss...@pb.com mailto:bill.doss...@pb.com www.pb.com http://www.pb.com/ *In Engineering?**Raise a ticket via Remedy Anywhere [HERE http://teamsite.inside.pb.com/sites/TCPPM/GES/Lists/Create%20Incident /newIncident.aspx]takes less
Re: [ovirt-users] Fwd: RE: FC QLogic question
Hi Sven! Thanks for the heads up on the BZ. One question remains; if the firmware is missing in oVirt node then how come Bill does the module loaded and all the LUNs with lsscsi? Cheers, -- Daniel Helgenberger m box bewegtbild GmbH P: +49/30/2408781-22 F: +49/30/2408781-10 ACKERSTR. 19 D-10115 BERLIN www.m-box.de www.monkeymen.tv Am 11.09.2014 um 17:29 schrieb Sven Kieske s.kie...@mittwald.de: Sorry for top posting, but this thread goes way too long: please take a look at this BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1063001 I think this is your issue! I already posted to the original thread but somehow nobody seems to read my mail. On 11/09/14 14:34, Daniel Helgenberger wrote: On 11.09.2014 14:23, Fabian Deutsch wrote: Hey Daniel, - Original Message - On 11.09.2014 14:09, Fabian Deutsch wrote: Hey Bill and Daniel, it would indeed be interesting - mandatory - to know if it's qla* or qxl (qxl is IMHO related to graphics). Once we know that I'll see if the bug below fixes it. Hello Fabian, indeed this was my mistake; while typing the mail I was reading on the spice website - so qla is the correct module of course! I already replied this to Bill but forgot to cc the list also ... Also, for the node iso it would make little sense to have the qlx module loaded. For completeness, I attach my answer to Bill. Thanks :) So to summarize: The adapter is recognized on Node? But it does not show up in Engine? Bill has to tell, but I think so. I suggested to him this might be because of dm-multipath, multipath.conf might blacklist here. My nodes are plain EL6 and I have multipath enabled - but I do not know if ovirt-node filters for dm-mp devnodes only by default (this would make sense because of iscsi-multipathing in ovirt 3.4). Maybe you can help out on this topic? Thanks! Thanks - fabian I'd need your help to verify that we are also including all relevant dependencies. Thanks - fabian - Original Message - On 11.09.2014 13:05, Bill Dossett wrote: Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one Indeed, sorry for that. lsscsi does list all of the LUNs... so that's working! Great! But nothing in ovirt engine when I try to add a storage domain. Please check back in ovirt engine, with Hosts / Hardware info. You should see the adapter there, along with WWN's ect. I am not quite sure, but AFAIK the node uses dm-multipath by default (for iscsi multipathing). So what does: # multipath -ll list? Thanks again, I guess I will try and get a bit further with the installs I have... Bill Hi Daniel, I did a lsmod | grep qxl and it returned nothing... did you perhaps mean qla? [root@ovirt141 ~]# lsmod | grep qla qla2xxx 466573 56 I think that is the one lsscsi does list all of the LUNs... so that's working! But nothing in ovirt engine when I try to add a storage domain. Thanks again, I guess I will try and get a bit further with the installs I have... Bill -Original Message- From: Daniel Helgenberger [mailto:daniel.helgenber...@m-box.de] Sent: 11 September 2014 11:53 To: Bill Dossett Subject: Re: [ovirt-users] FC QLogic question On 11.09.2014 12:17, Bill Dossett wrote: Thanks for responding Daniel. Never mind! I was wondering and thought the message may be irrelevant. But yes, it doesn't appear to be working. When I try to add a FC storage domain, no hosts are showing in the drop down box of hosts. I have loaded the latests ovirtnode (3.4.something...) on two hosts now and neither are working. Lets first try to isolate your FC issure. So you are using the latest ovirt-node? EL6? Keep in mind, the kernel is standard Linux and therefore should support Qlogic adapters out of the box (IMHO they are the best supported HBAs in Linux). Please try a: # lsmod | grep qxl to see whatever the qxl module is loaded. Keep in mind you can run into problems with other conrollers in your host, though I only had this with LSI SAS / FC controllers so far. Further check on the FC switch if the N port is logged in and the zoning is correct. Finally, check with # lsscsi if you can see the luns. Also having a lot of issues with configuring the network. I configure an interface I can ping it, but it shows as unconfigured... on one host I have added it to the manager and I see the interfaces anyway, but on the second host I configure the eth0, I can ping it, it added to the ovirt engine but it shows no NICs. Should I be using an earlier version maybe? thanks again and best regards. *Bill Dossett* *Virtualisation Architect* *Pitney Bowes - Tech Central*** 6 Hercules Way, 600 Leavesden Park Watford, WD25 7GS Phone(+44) (0)1923
Re: [ovirt-users] Fwd: RE: FC QLogic question
I really don't know, maybe it got silently fixed and the BZ was not updated. Just another reason _not_ to use node (sorry fabian) imho. On 11/09/14 18:16, Daniel Helgenberger wrote: Hi Sven! Thanks for the heads up on the BZ. One question remains; if the firmware is missing in oVirt node then how come Bill does the module loaded and all the LUNs with lsscsi? Cheers, -- Mit freundlichen Grüßen / Regards Sven Kieske Systemadministrator Mittwald CM Service GmbH Co. KG Königsberger Straße 6 32339 Espelkamp T: +49-5772-293-100 F: +49-5772-293-333 https://www.mittwald.de Geschäftsführer: Robert Meyer St.Nr.: 331/5721/1033, USt-IdNr.: DE814773217, HRA 6640, AG Bad Oeynhausen Komplementärin: Robert Meyer Verwaltungs GmbH, HRB 13260, AG Bad Oeynhausen ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] Fwd: RE: FC QLogic question
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 11.09.2014 19:20, Douglas Schilling Landgraf wrote: I do believe we are improving all components of oVirt release after release, including oVirt Node. Discouraging people to use it (or any other component) won't help the project. We are here to listen the feedback from users to improve the software and sometimes adding new features based in their requests and suggestions. I know you are a very active contributor in this project and thanks for that, if you still have issues in the node please keep the bugzillas coming. Sorry for getting Off-Topic: Well yeah, I see lots of improvement, also in the node project. But when people ask about what do you recommend for production I really can't call for node (yet). Most deployments tend to need customization specific to the site where the deployment takes place and in the past it was very hard to alter node for your personal needs (that's also not the projects main goal imho). I don't want to discourage anyone from testing node, but I won't recommend it for production, as I wouldn't recommend fedora (and breakage has happened for people using fedora, just search this mailing list). I really appreciate all the work that is done and I see many improvements, but when it comes to giving advice, some developers are a little naive, I can even make a list: someone has a problem in production: yeah this is fixed in $BETA (or $ALPHA), just try it out (without even mentioning that it's a beta release!) you could easily backport the fix yourself, it's just a one liner you need to alter config param $X (not mentioning this param gets overwritten on upgrades) and so on. I know these are all just good intentions, but I think some people seem to think those people asking are all just running test environments, which is plainly wrong, or I don't know what they think, maybe that it's already robust enough. the project has matured so much, you should keep in mind people might use it to run their business/store sensitive data/host very critical machines on top of ovirt. this is great, but this also comes with some kind of responsibility on the side of a developer. At least that's my opinion when I develop stuff (not such huge projects ;) ) I know there are labels on every open source project and in all software licenses like use at your own risk. But there is some moral responsibility for stuff you do you can't talk away with some legal paper saying your not responsible for what you do (code) or advice you give. Maybe you should just ask yourself sometimes: would I do myself what I just advocated if I have really sensible data in my environment? Anyway, keep up the good work! I know it's all good intentions and of course every developer wants his new features getting used. :) Thanks Sven -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.22 (MingW32) iQGcBAEBAgAGBQJUEiDjAAoJEAq0kGAWDrqlNzsL/1EWhurbdEgt1jhBb/stoNhH KnOANZyYCYUoUttGxG3h3KxM0wRA1mfpSO+yOtVYd/bfclu0M6dy5oRpRw4Q3m2U VGVKqyDKsuQSaYX9CGOY90sYyO8DCt+KV04IN/XntsSpudF5Y+V7EqgBAaGqSS3N ETNL9Jru91pdnJ+ThXtdEsp05DXg7pYYKTRFJ3mCEQOsD2AkcUlz91tQ2h7tC6ZF KPjdRLK2iwQR/ssKduPdirlJ/jxW1K2JaBWUJx9OLLFLpm7D6Ea87beIQwGqX27S jSk1lDm3MGJbK2cuKxZ8rDUnoJFa0DFkUi1pxp6l9IheOirXrvHcW57WkdUkZTb8 7fAeDMpFG3fny0+duAANneMI05NtA9hPl1Vnn6QBnp+nb7GET5JhcEpdgPe17h+W hCBdgfN1OveLscN8E8hGrV9YKKUxkJkFAcXhgxOrTW23z6rwwYDG46t630xTSgIp g401Zs7wyHZVb2FEZRZCuSBK07vcJI/b1nnG7OpS8A== =A8rd -END PGP SIGNATURE- ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users