[OmniOS-discuss] Moving a zfs pool from server to another with less memory.
Hi. Are there any special configuration or setup that I need to take consider ? I'm currently moving a 14TB ZFS pool from a out-dated OpenIndiana server with 16GB in RAM. The new server will only provide 8GB of RAM, and will be running OmniOS. The new server has a maximum of 8GB of RAM, so there is no option extending the RAM. I have also in plans to extend the size of the pool to at least 20TB. It only serves as a archive server. not much of client activity, only rsync and archival stuff. Any help and or information much appreciated. Thanks allot. Best regards, Svavar Reykjavik - Iceland ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss
Re: [OmniOS-discuss] Moving a zfs pool from server to another with less memory.
On 2013-10-30 16:44, Svavar Örn Eysteinsson wrote: Hi. Are there any special configuration or setup that I need to take consider ? I'm currently moving a 14TB ZFS pool from a out-dated OpenIndiana server with 16GB in RAM. The new server will only provide 8GB of RAM, and will be running OmniOS. The new server has a maximum of 8GB of RAM, so there is no option extending the RAM. I have also in plans to extend the size of the pool to at least 20TB. It only serves as a archive server. not much of client activity, only rsync and archival stuff. First of all, probably, make sure that your RAM is ECC (my limited NAS at home was based on desktop hardware which only used non-ECC memory which could be the cause of some of my problems - with no good way to check). Also, this amount of memory is very small for dedup for example. Otherwise, especially for an archival system with little random IO (sans scrubbing) it might not matter much... Basically, this all revolves around the working set (or hot data) size, so that for the data you mostly use your data or metadata is speedily available in RAM. There are rough estimates like you need 1GB RAM per TB of storage, but exact numbers depend on the system's usage and data layout (i.e. average number of blocks per TB). HTH, //Jim ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss
Re: [OmniOS-discuss] Re-installing OmniOS after Crash, Errors with pkg [Subject Edited]
On Wed, Oct 30, 2013 at 18:16:40 +0530, Sam M wrote: Hi Eric, The issue I'm having is that I had Bloody up and running and then my HD crashed. While on bloody, I upgraded my 8 TB data zpool. Now, the same zpool is not accessible when I'm running the release version of OmniOS. The issue I mentioned in my original email seems minor but it's causing other issues. I understand bloody is not supposed to be stable, but I can access the data zpool in readonly mode. So I was hoping for a way to fix the pkg issue, or I guess I'll wait for the updates to finish... If you install pkg@0.5.11,5.11-0.151007:20131027T234759Z (an unsigned version that supports signing), you should be able to use that as an interstitial version from which to do the rest of the upgrade. Let me know if this doesn't work for you. -- Chris Nehren Site Reliability Engineer, OmniTI ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss
[OmniOS-discuss] multipath problem when replacing a failed SAS drive
Hello, I'm running OmniOS r151006p on the following system: - Supermicro X8DT6 board, Xeon E5606 CPU, 48GB ram - Supermicro SC847 chassis, 36 drive bays, SAS expanders, LSI 9211-8i controller - 34 x Toshiba 3T SAS drives MG03SCA300 in one pool w/ 16 mirrored sets + 2 hot spares 'mpathadm list lu' showed all drives as having two paths to the controller. Yesterday, one of the drives failed and was replaced. The new drive is only showing one path in mpathadm, and errors have started showing up periodically in /var/adm/messages: # mpathadm list lu /dev/rdsk/c1t539478CA7150d0 mpath-support: libmpscsi_vhci.so /dev/rdsk/c1t539478CA7150d0s2 Total Path Count: 1 Operational Path Count: 1 Oct 30 09:30:22 hagler scsi: [ID 243001 kern.warning] WARNING: /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler mptsas_handle_event_sync: IOCStatus=0x8000, IOCLogInfo=0x31120101 Oct 30 09:30:22 hagler scsi: [ID 243001 kern.warning] WARNING: /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler mptsas_handle_event: IOCStatus=0x8000, IOCLogInfo=0x31120101 Oct 30 09:30:22 hagler scsi: [ID 365881 kern.info] /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler Log info 0x31120101 received for target 89. Oct 30 09:30:22 hagler scsi_status=0x0, ioc_status=0x804b, scsi_state=0xc Oct 30 09:30:22 hagler scsi: [ID 243001 kern.warning] WARNING: /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler mptsas_handle_event_sync: IOCStatus=0x8000, IOCLogInfo=0x31120101 Oct 30 09:30:22 hagler scsi: [ID 243001 kern.warning] WARNING: /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler mptsas_handle_event: IOCStatus=0x8000, IOCLogInfo=0x31120101 Oct 30 09:30:22 hagler scsi: [ID 365881 kern.info] /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler Log info 0x31120101 received for target 89. Oct 30 09:30:22 hagler scsi_status=0x0, ioc_status=0x804b, scsi_state=0xc Oct 30 09:30:22 hagler scsi: [ID 243001 kern.warning] WARNING: /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler mptsas_handle_event_sync: IOCStatus=0x8000, IOCLogInfo=0x31120101 Oct 30 09:30:22 hagler scsi: [ID 365881 kern.info] /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler Log info 0x31120101 received for target 89. Oct 30 09:30:22 hagler scsi_status=0x0, ioc_status=0x804b, scsi_state=0xc Oct 30 09:30:22 hagler scsi: [ID 243001 kern.warning] WARNING: /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler mptsas_handle_event: IOCStatus=0x8000, IOCLogInfo=0x31120101 Oct 30 09:30:22 hagler scsi: [ID 243001 kern.warning] WARNING: /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler mptsas_handle_event_sync: IOCStatus=0x8000, IOCLogInfo=0x31120101 Oct 30 09:30:22 hagler scsi: [ID 365881 kern.info] /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler Log info 0x31120101 received for target 89. Oct 30 09:30:22 hagler scsi_status=0x0, ioc_status=0x804b, scsi_state=0xc Oct 30 09:30:22 hagler scsi: [ID 243001 kern.warning] WARNING: /pci@0,0/pci8086,3410@9/pci1000,3020@0 (mpt_sas0): Oct 30 09:30:22 hagler mptsas_handle_event: IOCStatus=0x8000, IOCLogInfo=0x31120101 The error messages refer to target 89, which I can confirm corresponds to the missing path for my replacement drive using lsiutil: # lsiutil -p 1 16 LSI Logic MPT Configuration Utility, Version 1.63, June 4, 2009 1 MPT Port found Port Name Chip Vendor/Type/RevMPT Rev Firmware Rev IOC 1. mpt_sas0 LSI Logic SAS2008 03 200 0d000100 0 SAS2008's links are 6.0 G, 6.0 G, 6.0 G, 6.0 G, 6.0 G, 6.0 G, 6.0 G, 6.0 G B___T SASAddress PhyNum Handle Parent Type [ ... cut ... ] 0 89 539478ca715217 00590032 SAS Target 0 90 539478ca715317 005a000a SAS Target [ ... cut ... ] When I ask lsiutil to rescan the bus, I see the following error when it gets to target 89: # lsiutil -p 1 8 LSI Logic MPT Configuration Utility, Version 1.63, June 4, 2009 1 MPT Port found Port Name Chip Vendor/Type/RevMPT Rev Firmware Rev IOC 1. mpt_sas0 LSI Logic SAS2008 03 200 0d000100 0 SAS2008's links are 6.0 G, 6.0 G, 6.0 G, 6.0 G, 6.0 G, 6.0 G, 6.0 G, 6.0 G B___T___L Type Vendor Product Rev [ ... cut ... ] ScsiIo to Bus 0 Target 89 failed, IOCStatus = 004b (IOC Terminated) 0 90 0 Disk TOSHIBA MG03SCA300 0108 539478ca7153 17 [ ... cut ... ] This problem has happened to me once before on a similar system. At that time, I tried reseating the drive, and tried several different replacement drives, all had the same issue. I even tried rebooting the system and that didn't help. Does anyone know how I can clear this issue up? I'd be happy to provide any additional
[OmniOS-discuss] hw recomendation for JBOD box
We are looking at buying a system consisting of a 1U server and 1 or more 3U JBOD boxes to run and OmniOS ZFS server. Any HW recommendations for the JBOD box, Controller, Disks, what are you using ? cheers tobi -- Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland http://it.oetiker.ch t...@oetiker.ch ++41 62 775 9902 / sb: -9900 ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss
Re: [OmniOS-discuss] hw recomendation for JBOD box
+-- | On 2013-10-30 14:29:06, Eric Sproul wrote: | | Tobi, | I'd recommend reviewing the parts lists that Joyent publishes: | https://github.com/joyent/manufacturing | | See the parts_database.ods file. It's mostly Supermicro kit, and what | I find very helpful is the lists of BIOS and firmware revisions within | those parts. I've also had very good experience with Fujitsu RX200 and RX300s (using IT-flashed non-Fujitsu non-RAID LSI HBAs) and their JX40 JBODs. (We have ~40 or so of those systems, and half a dozen JX40s.) iRMC Enterprise is actually comparatively nice to use, too. Cheers. -- bdha cyberpunk is dead. long live cyberpunk. ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss
Re: [OmniOS-discuss] Physical slot based disk names for LSI SAS on OmniOS?
On Wed, Oct 30, 2013 at 4:25 PM, Chris Siebenmann c...@cs.toronto.edu wrote: This is a long shot and I suspect that the answer is no, but: in OmniOS, is it possible somehow to have disk device names for disks behind LSI SAS controllers that are based on the physical slot ('phy') that the disk is found in instead of the disk's reported WNN or serial number? Chris, I'm not 100% sure, but I think the integration of https://www.illumos.org/issues/4018 will give you what you want. This is currently available in r151007 but will be part of the upcoming r151008 release. Issues 4016-4019 are all related, in fact, and are all in r151007 at this time. Eric ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss
Re: [OmniOS-discuss] hw recomendation for JBOD box
On Wed, 30 Oct 2013 23:05:11 +0100 (CET) Tobias Oetiker t...@oetiker.ch wrote: http://www.supermicro.com/products/system/4u/6047/ssg-6047r-e1r36l.cfm filled with UltraStar 7k3000 and a ZeusRam 8GB as Zil and 256 GB ram Nice rick;-) nice :-) no jbod though as far as I can see ... so maybe jbod is not such a good idea ... Point 6: Onboard LSI 2308 IT mode. For LSI HBA IT mode means jbod mode. And no, the only good thing for ZFS is jbod mode! The non-IT mode HBA's where you create a RAID0 for each disk is simply horrible. -- Hilsen/Regards Michael Rasmussen Get my public GnuPG keys: michael at rasmussen dot cc http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xD3C9A00E mir at datanom dot net http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE501F51C mir at miras dot org http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE3E80917 -- I've got a very bad feeling about this. -- Han Solo signature.asc Description: PGP signature ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss
Re: [OmniOS-discuss] hw recomendation for JBOD box
Tobi! @ the discussion about JBOD and expanders: For me, JBOD is the best solution, at least for spinning disks. Where you would place your cache is another question. But for the JBOD, you would need to ask yourself wether you would like expanders or not. That depends of course of your environmental needs, the prestanda needed, your workload, etc. JBOD, for me, is the best solution because of the flexibility. It's easy to move it to another storage head. There was a discussion on this list a couple of month ago, where someone mentioned expanderless JBOD's, I believe it was Supermicro. I don't recall who it was that mentioned them, perhaps it was Richard Elling? With these you can get all the bandwith you like, so it's like internal disks. I would like someone that has experience of these to comment on this. Otherwise, normally perhaps, you wouldn't need direct links to all of your drives, especially if you got a good cache implementation. I would also consider, if I was in your situation, a 2U storage head server, instead of a 1U, because of two things: 1: The cache. You would like your cache to be able to expand. 2: Expansion even here: You would like to have good expansion possibilities in the chassi and the mo'bo'(slots), for beeing able to connect to more JBOD's. In a 1U chassi your expansion possibilities are very limited. Rgrds Johan -OmniOS-discuss omnios-discuss-boun...@lists.omniti.com skrev: - Till: Eric Sproul espr...@omniti.com Från: Tobias Oetiker Sänt av: OmniOS-discuss Datum: 2013.10.30 23:06 Kopia: omnios-discuss omnios-discuss@lists.omniti.com Ärende: Re: [OmniOS-discuss] hw recomendation for JBOD box Hi Eric, Today Eric Sproul wrote: Tobi, I'd recommend reviewing the parts lists that Joyent publishes: https://github.com/joyent/manufacturing so that would be this http://www.supermicro.com/products/system/4u/6047/ssg-6047r-e1r36l.cfm filled with UltraStar 7k3000 and a ZeusRam 8GB as Zil and 256 GB ram nice :-) no jbod though as far as I can see ... so maybe jbod is not such a good idea ... cheers tobi -- Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland http://it.oetiker.ch t...@oetiker.ch ++41 62 775 9902 / sb: -9900 ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss
Re: [OmniOS-discuss] hw recomendation for JBOD box
Yesterday Michael Rasmussen wrote: On Wed, 30 Oct 2013 23:05:11 +0100 (CET) Tobias Oetiker t...@oetiker.ch wrote: http://www.supermicro.com/products/system/4u/6047/ssg-6047r-e1r36l.cfm filled with UltraStar 7k3000 and a ZeusRam 8GB as Zil and 256 GB ram Nice rick;-) nice :-) no jbod though as far as I can see ... so maybe jbod is not such a good idea ... Point 6: Onboard LSI 2308 IT mode. For LSI HBA IT mode means jbod mode. And no, the only good thing for ZFS is jbod mode! The non-IT mode HBA's where you create a RAID0 for each disk is simply horrible. sure, jbod mode, I meant having an external jbod box attached cheers tobi -- Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland http://it.oetiker.ch t...@oetiker.ch ++41 62 775 9902 / sb: -9900 ___ OmniOS-discuss mailing list OmniOS-discuss@lists.omniti.com http://lists.omniti.com/mailman/listinfo/omnios-discuss