Re: [OpenIndiana-discuss] Migrating root zpool from old disks to new one. only one system available
2012-06-12 15:40, J. V. пишет: On 12/06/12 09:16 AM, Rich wrote: Won't zpool replace fail b/c the new disks require ashift=12 and his existing pool devices have ashift=9? This should work fine: HP Microserver, upgrading a mirrored pool from 2TB HDs to 3TB HDs. At one point, the pool had one 2TB HD with 512 sectors and the other with 4k sectors (both in a mirror). Seems there was a bit of a performance hit, but everything worked fine. When I upgraded from 2TB to 3TB, the 3TB HDs were both 4k sectors: Did the 3TB disk also report to the OS that it uses 4k sectors? What ashift value is used by the pool, ultimately? Example: # zdb -l /dev/rdsk/c2t0d0s0 | grep ashift ashift: 9 ashift: 9 ashift: 9 ashift: 9 If your 4KB disks use ashift=9 it is possible to get problems worse than decreased performance ;) Note that the ashift is set on a top-level vdev (the second-tier component of a pool) and the value is saved into/read from its vdevs (edge-tier components), and a pool can mix top-level vdevs with different ashifts, i.e. if it is a raid10 type of setup. PS: Do you use an HP N40L? Did you test if more than 8Gb RAM can fit into it (using DDR3 ECC sticks larger than 4Gb), or does the server's BIOS forbid that? HTH, //Jim Klimov ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] Migrating root zpool from old disks to new one. only one system available
I did check the ashift after the upgrade and both 3TB HDs report ashift=12. I am actually using an HP N35L (1.3 Ghz vs. 1.5 GHz in the N40L) and I have 8 GB of non-ECC DDR3 so I could not answer the memory question. These little machines are really great (If they came with SATA 3, they would be the ultimate small server): I do have an N40L with ESXi 5 running my firewall/router (Untangle 9.2) and Windows Home Server 2011 (For the Media Server, Backup and WSUS services).. The OpenIndiana server has 6 HDs (loaded modified BIOS so the extra internal and external SATA ports run at full speeds): 4 x 3TB Seagate ST3000DM001 (Newer Seagates: Fastest 7200 rpm HD right now ... ~140 MB on SATA 2, ~160 on SATA 3). 2 x OCZ 60 GB SSD (Agility 3). Mounted where the CD-ROM would go with a Vantec 5 1/4, 2 x 2.5 in. (Used an eSATA to SATA cable for one of the SSDs). Config: Partitioned the SSDs in 3 slices: rpool (16GB), zil1 (4GB), arc1 (~40GB) rpool (16GB), zil2 (4GB), arc2 (~40GB) rpool = 2 x 16GB mirrored Storage pool 1 = 2 x 3TB mirrored + zil1 + arc1 Storage pool 2 = 2 x 3TB mirrored + zil2 + arc2 The mirrored rpool saved my butt last month when I was playing around and over-wrote the boot rpool (gets confusing between slices and partitions with my SSD config). Took me whole 20 min. to fix it: Exchange SSD places, boot from good rpool, wait 5 min. for resilver, re-install grub in damaged rpool, poweroff, exchange SSDs again). No other dual cpu machine that I have uses less power: The whole server only uses 23, 24 Watts when idle (2 SSDs active, 4 HDs powered-down). 35 W with 1 pool active (2 SSDs, 2 HDs active) ~50 W with all HDs active and getting hammered. Once I added zil + arc to the pools, the little thing runs at wire speeds: Could not get any better! It took me months to start gathering the hardware and fully learn ZFS, but I am really glad to be back running a Solaris derivative: I started in college administering Ultrix, Sun-OS, AIX, NeXTStep and others. Then later I went to the dark side to learn all about MS, and finally feel back at home with OpenIndiana! Thanks to the IllumOS people and contributors here and everywhere else! (Sorry for the long post. Hope my full config can help/inspire others). Jose V. On Tue, Jun 12, 2012 at 8:00 AM, Jim Klimov jimkli...@cos.ru wrote: 2012-06-12 15:40, J. V. пишет: On 12/06/12 09:16 AM, Rich wrote: Won't zpool replace fail b/c the new disks require ashift=12 and his existing pool devices have ashift=9? This should work fine: HP Microserver, upgrading a mirrored pool from 2TB HDs to 3TB HDs. At one point, the pool had one 2TB HD with 512 sectors and the other with 4k sectors (both in a mirror). Seems there was a bit of a performance hit, but everything worked fine. When I upgraded from 2TB to 3TB, the 3TB HDs were both 4k sectors: Did the 3TB disk also report to the OS that it uses 4k sectors? What ashift value is used by the pool, ultimately? Example: # zdb -l /dev/rdsk/c2t0d0s0 | grep ashift ashift: 9 ashift: 9 ashift: 9 ashift: 9 If your 4KB disks use ashift=9 it is possible to get problems worse than decreased performance ;) Note that the ashift is set on a top-level vdev (the second-tier component of a pool) and the value is saved into/read from its vdevs (edge-tier components), and a pool can mix top-level vdevs with different ashifts, i.e. if it is a raid10 type of setup. PS: Do you use an HP N40L? Did you test if more than 8Gb RAM can fit into it (using DDR3 ECC sticks larger than 4Gb), or does the server's BIOS forbid that? HTH, //Jim Klimov ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] Migrating root zpool from old disks to new one. only one system available
Yes it will. The only way to do this is to create a secondary pool and send/receive your root pool to the new pool. - George On Jun 11, 2012, at 7:16 PM, Rich wrote: Won't zpool replace fail b/c the new disks require ashift=12 and his existing pool devices have ashift=9? - Rich On Mon, Jun 11, 2012 at 7:12 PM, James C. McPherson james.c.mcpher...@gmail.com wrote: On 12/06/12 08:39 AM, Hans J. Albertsson wrote: Suppose: I have a system with but two disks. They're fairly small: 300GB, and use 512B sectors. These two disks are a mirror zpool, creqated from the entire disks. There are about 20 or so filesystems in there. The system has room for only two disks. I'd like to replace these two small disks with two 2TB disks using 4kB blocks. So: Is there a writeup on how to connect one of these new disks to the existing machine, using an external esata cabinet. set up a new zpool on this new disk, and transfer all the root pool data to the new single disk zpool. then set up the new zpool to be bootable. And last: taking the old disks out of the machine, place the new single disk, and another, empty, similar 2TB disk in the machine, and boot from the single new one as the new root zpool. And then add the second new disk as a mirror.. effectively running the old system exactly as it was w/o reinstalling anything significant, but with much roomier disks. Note: in this case there is no way to get another system to do it on. And a third disk can only be connected using an external cabinet and Esata or USB. Hi Hans, I suggest this: #1 set autoexpand=on for rpool #2 connect your esata enclosure #3 once you've got them sliced up as desired (I suggest slice 0 should cover all except cylinders 0 and 1), run installgrub on both new disks, to the mbr #4 zpool replace one of your rpool disks #5 zpool replace the other rpool disk #6 poweroff #7 do the physical replacement #8 poweron James C. McPherson -- Solaris kernel software engineer, system admin and troubleshooter http://www.jmcp.homeunix.com/blog Find me on LinkedIn @ http://www.linkedin.com/in/jamescmcpherson ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] HDD Upgrade problem
On Jun 12, 2012, at 12:58 AM, Rich wrote: On Tue, Jun 12, 2012 at 12:50 AM, Richard Elling richard.ell...@richardelling.com wrote: On Jun 11, 2012, at 6:08 PM, Bob Friesenhahn wrote: On Mon, 11 Jun 2012, Jim Klimov wrote: ashift=12 (2^12 = 4096). For disks which do not lie, it works properly out of the box. The patched zpool binary forced ashift=12 at the user's discretion. It seems like new pools should provide the option to be created with ashift=12 even if none of the original disks need it. The reason is to allow adding 4K sector disks later. I don't disagree, in principle. However, history has shown that people don't plan very well (eg, reason #1 dedup exists). Is there some other way to be more clever? -- richard I believe that the consensus the last time this came up was agreement on that point, but contention over how the semantics of specifying it should work? ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss Illumos has a way to override the physical block size of a given disk by using the sd.conf file. Here's an example: sd-config-list = DGC RAID, physical-block-size:4096, NETAPP LUN, physical-block-size:4096; By adding the VID PID values into the sd-config-list along with the 'physical-block-size' parameter you can override the value that the disk will use for it's block size. ZFS will pick this up and use the correct ashift value. - George ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] HDD Upgrade problem
On Tue, 12 Jun 2012, George Wilson wrote: Illumos has a way to override the physical block size of a given disk by using the sd.conf file. Here's an example: sd-config-list = DGC RAID, physical-block-size:4096, NETAPP LUN, physical-block-size:4096; By adding the VID PID values into the sd-config-list along with the 'physical-block-size' parameter you can override the value that the disk will use for it's block size. ZFS will pick this up and use the correct ashift value. This can obviously work once the OS is installed and it is possible to edit system files. Is there a way to apply it to the root pool when the system is installed? Bob -- Bob Friesenhahn bfrie...@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer,http://www.GraphicsMagick.org/ ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] HDD Upgrade problem
On Jun 12, 2012, at 11:00 AM, Bob Friesenhahn wrote: On Tue, 12 Jun 2012, George Wilson wrote: Illumos has a way to override the physical block size of a given disk by using the sd.conf file. Here's an example: sd-config-list = DGC RAID, physical-block-size:4096, NETAPP LUN, physical-block-size:4096; By adding the VID PID values into the sd-config-list along with the 'physical-block-size' parameter you can override the value that the disk will use for it's block size. ZFS will pick this up and use the correct ashift value. This can obviously work once the OS is installed and it is possible to edit system files. Is there a way to apply it to the root pool when the system is installed? Bob -- Bob Friesenhahn bfrie...@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer,http://www.GraphicsMagick.org/ ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss I have not tried this but you might be able to do change sd.conf and then run 'update_drv -f sd' prior to starting the installation. - George ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] Migrating root zpool from old disks to, new one. only one system available
Actually I found the following blog quite illuminating. http://www.big-bubbles.fluff.org/blogs/bubbles/blog/2012/02/04/placeholder-migrating-a-zfs-root-pool/ It would have to be worked thru for a number of specific cases. But it is basically sound. A bird's eye view might be: format the new disk to have a partition s0 starting from cyl 1 and a boot partition 8 on cyl 0 set up a zpool (using the hacked version with --blocksize if you'r not certain the disk reports a truthful 4k block size) on s0 create a new BE on that zpool as a clone of your current BE. Add swap and dump in the new zpool too zfs send/recv using -R anything that wasn't brought over by the beadm cloning installgrub -R the new device, and edit vfstab and do a dumpadm to point swap and dump as seen from the new environment to the new ones. Name will be not rpool, and you can do another round as above if you care. Read that blog while doing it: this bird's eye isn't anywhere near complete. On 2012-06-12 09:00, openindiana-discuss-requ...@openindiana.org wrote: Message: 1 Date: Tue, 12 Jun 2012 00:39:32 +0200 From: Hans J. Albertssonhans.j.alberts...@branneriet.se To:openindiana-discuss@openindiana.org Subject: [OpenIndiana-discuss] Migrating root zpool from old disks to new one. only one system available Message-ID:4fd673a4.5030...@branneriet.se Content-Type: text/plain; charset=ISO-8859-1; format=flowed Suppose: I have a system with but two disks. They're fairly small: 300GB, and use 512B sectors. These two disks are a mirror zpool, creqated from the entire disks. There are about 20 or so filesystems in there. The system has room for only two disks. I'd like to replace these two small disks with two 2TB disks using 4kB blocks. So: Is there a writeup on how to connect one of these new disks to the existing machine, using an external esata cabinet. set up a new zpool on this new disk, and transfer all the root pool data to the new single disk zpool. then set up the new zpool to be bootable. And last: taking the old disks out of the machine, place the new single disk, and another, empty, similar 2TB disk in the machine, and boot from the single new one as the new root zpool. And then add the second new disk as a mirror.. effectively running the old system exactly as it was w/o reinstalling anything significant, but with much roomier disks. Note: in this case there is no way to get another system to do it on. And a third disk can only be connected using an external cabinet and Esata or USB. ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] HDD Upgrade problem
2012-06-12 19:00, Bob Friesenhahn wrote: On Tue, 12 Jun 2012, George Wilson wrote: Illumos has a way to override the physical block size of a given disk by using the sd.conf file. Here's an example: sd-config-list = DGC RAID, physical-block-size:4096, NETAPP LUN, physical-block-size:4096; By adding the VID PID values into the sd-config-list along with the 'physical-block-size' parameter you can override the value that the disk will use for it's block size. ZFS will pick this up and use the correct ashift value. This can obviously work once the OS is installed and it is possible to edit system files. Is there a way to apply it to the root pool when the system is (being) installed? First of all, I believe this snippet belongs on a Wiki page, and I'll try to make one to sum up all knowledge and FUD we have about The AShift Problem ;) At least, it would be easier to point people to this page as a common answer ;) Second, I believe the sd.conf fix won't apply to drives used in IDE mode via BIOS, and users might have that (although it is an item best fixed before installation). Third, perhaps the sd.conf settings can be re-read by reloading the module and/or using add_drv or something like that? From a different thread (on disabling power-saving on Dell HBAs): ...enter into sd.conf and reload the driver via update_drv -vf sd. Also regarding the install-time setting - if the installation is done using the wizard, then the zpool binary at the moment of installation should use the ashift=12 you need, without extra command-line parameters (as to not go hacking inside the wizard). Earlier I did this by using a modified binary from that blog entry instead of the system-provided one, now I guess this can be done better with sd.conf :) HTH, //Jim Klimov ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] HDD Upgrade problem
2012-06-12 19:22, Jim Klimov wrote: First of all, I believe this snippet belongs on a Wiki page, and I'll try to make one to sum up all knowledge and FUD we have about The AShift Problem ;) At least, it would be easier to point people to this page as a common answer ;) FWIW, here is the first draft: http://wiki.illumos.org/display/illumos/The+AShift+Value+and+Advanced+Format+disks+and+ZFS Comments/additions welcome, //Jim ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] HDD Upgrade problem
That is absolutely stunning. Many thanks for that. I guess I'm going to have to bite the bullet, wait until I've got another two 3tb drives and create a new pool. Thanks to all for the feedback. Your patience with me is much appreciated. On 12/06/12 17:44, Jim Klimov wrote: 2012-06-12 19:22, Jim Klimov wrote: First of all, I believe this snippet belongs on a Wiki page, and I'll try to make one to sum up all knowledge and FUD we have about The AShift Problem ;) At least, it would be easier to point people to this page as a common answer ;) FWIW, here is the first draft: http://wiki.illumos.org/display/illumos/The+AShift+Value+and+Advanced+Format+disks+and+ZFS Comments/additions welcome, //Jim ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] Access to ZFS viz CIFS from windows regularly hangs.
On Tue, 12 Jun 2012, John McEntee wrote: I am having problems with a openindiana storage server I have built am I am trying to track down the cause to fix it. The current symptoms are seen from all windows clients (both 7 and XP) that will report an error stating. Path File is not accessible. The specified network name is no longer available. The first thing to verify is your network and network interface. Run continuous traffic and see if there are any hickups. You can use /usr/sbin/ping for testing with larger packets. Also check the log files under /var/adm and /var/log. Also check output of 'fmadm -ev' and 'fmadm faulty'. Bob -- Bob Friesenhahn bfrie...@simple.dallas.tx.us, http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer,http://www.GraphicsMagick.org/ ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] Access to ZFS viz CIFS from windows regularly hangs.
I had similar issues before I enabled TLER, and disabled the head parking on my WD Green drives. A quick Google shows some evidence of similar features on the 3TB Hitachis. On Tue, Jun 12, 2012 at 9:55 PM, Bob Friesenhahn bfrie...@simple.dallas.tx.us wrote: On Tue, 12 Jun 2012, John McEntee wrote: I am having problems with a openindiana storage server I have built am I am trying to track down the cause to fix it. The current symptoms are seen from all windows clients (both 7 and XP) that will report an error stating. Path File is not accessible. The specified network name is no longer available. The first thing to verify is your network and network interface. Run continuous traffic and see if there are any hickups. You can use /usr/sbin/ping for testing with larger packets. Also check the log files under /var/adm and /var/log. Also check output of 'fmadm -ev' and 'fmadm faulty'. Bob -- Bob Friesenhahn bfrie...@simple.dallas.tx.us, http://www.simplesystems.org/** users/bfriesen/ http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer,http://www.GraphicsMagick.org/ __**_ OpenIndiana-discuss mailing list OpenIndiana-discuss@**openindiana.orgOpenIndiana-discuss@openindiana.org http://openindiana.org/**mailman/listinfo/openindiana-**discusshttp://openindiana.org/mailman/listinfo/openindiana-discuss -- Seconds to the drop, but it seems like hours. http://www.eff.org/ http://www.eff.org/http://creativecommons.org/ ___ OpenIndiana-discuss mailing list OpenIndiana-discuss@openindiana.org http://openindiana.org/mailman/listinfo/openindiana-discuss
Re: [OpenIndiana-discuss] Access to ZFS viz CIFS from windows regularlyhangs.
Does the suspend event only occur on SMB clients or does it impact the other storage clients when triggered by the Windows clients? Any domain controller event errors? dmsg output? fmdump -eV output? uname -a output? Have you attempted a packet capture of the event? snoop -o smb-client.cap clientip -Original Message- From: John McEntee [mailto:jmcen...@stirling-dynamics.com] Sent: Tuesday, June 12, 2012 8:52 AM To: Discussion list for OpenIndiana Subject: [OpenIndiana-discuss] Access to ZFS viz CIFS from windows regularlyhangs. I am having problems with a openindiana storage server I have built am I am trying to track down the cause to fix it. The current symptoms are seen from all windows clients (both 7 and XP) that will report an error stating. Path File is not accessible. The specified network name is no longer available. Another symptom is windows explorer hangs and the user has to wait for it to some back. Just waiting a while ( a few minutes) and the box comes back. I currently think the root cause is in openindiana somewhere but am at a bit of a loss. I have tried many things and have still not fixed it. I think the box is lightly loaded for the hardware spec but kernel load increases to 40% when a zfssnap is taking place. Hardware spec. 2 x Xeon E6520 cpus 48 GB RAM Intel HC5520 motherboard 3 x LSI SAS 9211-8i cards Currently on openindiana 148 The box is joined to a windows 2003 domain. Zpool tank is 3 way mirror of 7 x 3TB hitachi disk (using 21 disks in total, zpool size of 19 TB, ) with 2 x SSD 8GB ZIL on each and 140GB L2ARC on each, default checksum, no dedup and no compression. Server operates as a windows home directory for 58 users (some laptops users so just a backup location), a main shared drive for the company of 120 users. It is also a nfs server to a vmware vsphere 4 server hosting 10 virtual machines. There are only 8 active production file systems, and 12 backup file systems from other hosts (done out of hours). Zpool iostat peaks at about 35 MB for the pool mostly around the 0 to 7 MB level. Turning of time-sliderd does not stop the problem. (backups run out of hours) A dtrace -n 'sched:::off-cpu { @[execname]=count()}' Used to give a sched count in the 6 to 7 fiqures over 3 seconds, but turing apci off with #eeprom acpi-user-options=0x8 Reduced this to 5 figures. What can I do to identify the problem to be able to fix it? Thanks John Other information: dtrace -n 'sched:::off-cpu { @[execname]=count()}' dtrace: description 'sched:::off-cpu ' matched 3 probes ^C gconfd-2 2 idmapd2 inetd 2 nscd 2 sendmail 2 svc.startd2 gnome-power-mana 3 fmd 4 sshd 4 devfsadm 6 fsflush 7 nfsmapid 7 ntpd 7 dtrace 13 Xorg 17 gdm-simple-greet 17 svc.configd 71 smbd113 time-sliderd138 zpool-rpool 597 nfsd918 zpool-tank 1968 sched 80542 # echo hz/D | sudo mdb -k hz: hz: 100 # echo apic_timer::print apic_timer_t | sudo mdb -k { mode = 0 apic_timer_enable_ops = oneshot_timer_enable apic_timer_disable_ops = oneshot_timer_disable apic_timer_reprogram_ops = oneshot_timer_reprogram } ___ The contents of this e-mail and any attachment(s) are strictly confidential and are solely for the person(s) at the e-mail address(es) above. If you are not an addressee, you may not disclose, distribute, copy or use this e-mail, and we request that you send an e-mail to ad...@stirling-dynamics.com and delete this e-mail. Stirling Dynamics Ltd. accepts