[zfs-discuss] Problem with snapshot
Hi I am facing some problems after rolling back the snapshots created on pool. Environment: bash-3.00# uname -a SunOS hostname 5.10 Generic_118833-17 sun4u sparc SUNW,Sun-Blade-100 ZFS version: bash-3.00# zpool upgrade This system is currently running ZFS version 2. All pools are formatted using this version. I have a zpool called testpol with 10G This is the initial pool status of the pool bash-3.00# zpool list NAMESIZEUSED AVAILCAP HEALTH ALTROOT testpol9.94G 90K 9.94G 0% ONLINE - bash-3.00# zfs list NAME USED AVAIL REFER MOUNTPOINT testpol 84K 9.78G 24.5K /testpol Now i run the following commands bash-3.00# mkfile 10m /testpol/10megfile bash-3.00# zfs create testpol/fs1 bash-3.00# mkfile 20m /testpol/fs1/20megfile [b]bash-3.00# zfs snapshot test...@snap[/b] bash-3.00# zfs create testpol/fs2 bash-3.00# mkfile 30m /testpol/fs2/30megfile bash-3.00# mkfile 15m /testpol/15megfile Output of zfs list command after running the above commands bash-3.00# zfs list(shows that all the above commands were successfully executed) NAME USED AVAIL REFER MOUNTPOINT testpol 75.2M 9.71G 25.0M /testpol test...@snap 23.5K - 10.0M - testpol/fs1 20.0M 9.71G 20.0M /testpol/fs1 testpol/fs2 30.0M 9.71G 30.0M /testpol/fs2 The following are the file/file system entries under /testpol bash-3.00# ls -lR /testpol /testpol: total 51222 -rw--T 1 root root 10485760 Jan 29 13:32 10megfile -rw--T 1 root root 15728640 Jan 29 13:34 15megfile drwxr-xr-x 2 root sys3 Jan 29 13:33 fs1 drwxr-xr-x 2 root sys3 Jan 29 13:34 fs2 /testpol/fs1: total 40977 -rw--T 1 root root 20971520 Jan 29 13:33 20megfile /testpol/fs2: total 61461 -rw--T 1 root root 31457280 Jan 29 13:34 30megfile Everything shows up correctly until i rollback to the snapshot test...@snap bash-3.00# zfs rollback test...@snap bash-3.00# zfs list NAME USED AVAIL REFER MOUNTPOINT testpol 60.2M 9.72G 10.0M /testpol test...@snap 0 - 10.0M - testpol/fs1 20.0M 9.72G 20.0M /testpol/fs1 testpol/fs2 30.0M 9.72G 30.0M /testpol/fs2 bash-3.00# ls -lR /testpol/ /testpol/: total 20490 -rw--T 1 root root 10485760 Jan 29 13:32 10megfile drwxr-xr-x 2 root root 2 Jan 29 13:32 fs1 [b]fs1 is treated as a normal directory. rm fs1 will succeed, which would fail in case of a file system[/b] /testpol/fs1: total 0 [b]fs1 is empty.[/b] As expected fs2 (which was created after snapshot test...@snap was taken) is not listed under directories. Issues after rolling back: 1. Before snapshot was taken fs1 contained 20megfile which not present after the snapshot is rolled back. 2. Though file system fs2 is not present on the disk, zfs list shows fs2 3. The size of the file system fs1 is incorrect 4. After performing the rollback operation fs1 is not treated as a file system bash-3.00# mkfile 45m /testpol/fs1/45megfile bash-3.00# zfs list NAME USED AVAIL REFER MOUNTPOINT testpol105M 9.68G 55.0M /testpol test...@snap 23.5K - 10.0M - testpol/fs1 20.0M 9.68G 20.0M /testpol/fs1 testpol/fs2 30.0M 9.68G 30.0M /testpol/fs2 You could see 45m got added to /testpol not fs1 Did i do something that i shouldn't be doing? Can anyone please explain me what is wrong with this behavior? -Abishek -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Add SSD drive as L2ARC(?) cache to existing ZFS raid?
Thanx for your answers guys. :o) I dont contemplating trying this for my ZFS raid, as the SSD drives are expensive right now. I just want to be able to answer questions when I convert Windows/Linux to Solaris. And therefore collect info. Has anyone tried this on a blogg? Would be cool to blog about. Before and After. BTW, did I tell you that Solaris rocks? :o) -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] need to add space to zfs pool that's part of SNDR replication
BJ, The means to specify this is sndradm -nE ..., when 'E' is equal enabled. Got it. Nothing on the disk, nothing to replicate (yet). :-) The manner in which SNDR can guarantee that two or more volumes are write-order consistent, as they are replicated is place them in the same I/O consistency group. Ok, so my sndradm -nE command with g [same name as first data drive group] simply ADDs a set of drives to the group, it doesn't stop or replace the replication on the first set of drives, and in fact in keeping the same group name I even keep the two sets of drives in each server in sync. THEN I run my zfs attach command on the non-bitmap slice to my existing pool. Do I have that all right? Yes. Thanks! -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss Jim Dunham Engineering Manager Storage Platform Software Group Sun Microsystems, Inc. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] strange performance drop of solaris 10/zfs
Hi We have been using a Solaris 10 system (Sun-Fire-V245) for a while as our primary file server. This is based on Solaris 10 06/06, plus patches up to approx May 2007. It is a production machine, and until about a week ago has had few problems. Attached to the V245 is a SCSI RAID array, which presents one LUN to the OS. On this lun is a zpool (tank), and within that 300+ zfs file systems (one per user for automounted home directories). The system is connected to our LAN via gigabit Ethernet,. most of our NFS clients have just 100FD network connection. In recent days performance of the file server seems to have gone off a cliff. I don't know how to troubleshoot what might be wrong? Typical zpool iostat 120 output is shown below. If I run truss -D df I see each call to statvfs64(/tank/bla) takes 2-3 seconds. The RAID itself is healthy, and all disks are reporting as OK. I have tried to establish if some client or clients are thrashing the server via nfslogd, but without seeing anything obvious. Is there some kind of per-zfs-filesystem iostat? End users are reporting just saving small files can take 5-30 seconds? prstat/top shows no process using significant CPU load. The system has 8GB of RAM, vmstat shows nothing interesting. I have another V245, with the same SCSI/RAID/zfs setup, and a similar (though a bit less) load of data and users where this problem is NOT apparent there? Suggestions? Kevin Thu Jan 29 11:32:29 CET 2009 capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank2.09T 640G 10 66 825K 1.89M tank2.09T 640G 39 5 4.80M 126K tank2.09T 640G 38 8 4.73M 191K tank2.09T 640G 40 5 4.79M 126K tank2.09T 640G 39 5 4.73M 170K tank2.09T 640G 40 3 4.88M 43.8K tank2.09T 640G 40 3 4.87M 54.7K tank2.09T 640G 39 4 4.81M 111K tank2.09T 640G 39 9 4.78M 134K tank2.09T 640G 37 5 4.61M 313K tank2.09T 640G 39 3 4.89M 32.8K tank2.09T 640G 35 7 4.31M 629K tank2.09T 640G 28 13 3.47M 1.43M tank2.09T 640G 5 51 433K 4.27M tank2.09T 640G 6 51 450K 4.23M tank2.09T 639G 5 52 543K 4.23M tank2.09T 640G 26 57 3.00M 1.15M tank2.09T 640G 39 6 4.82M 107K tank2.09T 640G 39 3 4.80M 119K tank2.09T 640G 38 8 4.64M 295K tank2.09T 640G 40 7 4.82M 102K tank2.09T 640G 43 5 4.79M 103K tank2.09T 640G 39 4 4.73M 193K tank2.09T 640G 39 5 4.87M 62.1K tank2.09T 640G 40 3 4.88M 49.3K tank2.09T 640G 40 3 4.80M 122K tank2.09T 640G 42 4 4.83M 82.0K tank2.09T 640G 40 3 4.89M 42.0K ... ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] ZFS with Rational (ClearCase VOB) Supported ???
Hi, Is anyone using ZFS with IBM now Rational (ClearCase VOB)? I am looking for how to take the VOB backup using the ZFS snapshot and restoring it? Is this a supported config from vendor side? Any help would be appreciated. Rgds Vikash ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] strange performance drop of solaris 10/zfs
On Thu, Jan 29, 2009 at 6:13 AM, Kevin Maguire k.c.f.magu...@gmail.com wrote: I have tried to establish if some client or clients are thrashing the server via nfslogd, but without seeing anything obvious. Is there some kind of per-zfs-filesystem iostat? The following should work in bash or ksh, so long as the list of zfs mount points does not overflow the maximum command line length. $ fsstat $(zfs list -H -o mountpoint | nawk '$1 !~ /^(\/|-|legacy)$/') 5 -- Mike Gerdts http://mgerdts.blogspot.com/ ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Add SSD drive as L2ARC(?) cache to existing ZFS raid?
Imagine 10 SATA discs in raidz2 and one or two SSD drives as a cache. Each Vista client reaches ~90MB/sec to the server, using Solaris CIFS and iSCSI. So you want to use iSCSI with this. (iSCSI allows ZFS to export a file system as a native SCSI disc to a desktop PC. The desktop PC can mount this iSCSI disk as a native SCSI disk and format it with NTFS - on top of ZFS with snapshots, etc. This is done from desktop PC bios). Now, you install WinXP on the iSCSI ZFS volume, and clone it with a snapshot. Then you can boot from the clone, with the iSCSI volume on a desktop PC. Thus, your desktop PC doesnt need any hard drive at all. It uses the iSCSI volume on the ZFS server as a native SCSI disk, which has WinXP installed. This way, you can deploy lots of desktop PC in an instant. Using the cloned WinXP snapshot. And if there are problems e.g virus, just destroy the clone and create a new one in one second. Has anyone does this? Does the SSD provide extra speed? Any stories to share? (Ive read this iSCSI suggestion on a blogg with black background color, it's not my idea). If I add a SSD disk as a cache, can I remove it? No? Will there be problems if I remove it? Can I exchange it with a bigger? (Trying to convert these Windows people) -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS extended ACL
On Wed, Jan 28, 2009 at 11:16 PM, Christine Tran christine.t...@gmail.com wrote: On Wed, Jan 28, 2009 at 11:07 PM, Christine Tran christine.t...@gmail.com wrote: What is wrong with this? # chmod -R A+user:webservd:add_file/write_data/execute:allow /var/apache chmod: invalid mode: `A+user:webservd:add_file/write_data/execute:allow' Try `chmod --help' for more information. Never mind. /usr/gnu/bin/chmod Can we lose GNU, gee louise it is OpenSolaris2008.11 isn't it. ls [-v|-V] is messed up as well. Blarhghgh! There was a very long discussion about this a couple of weeks ago on one of the lists. Apparently the decision was made to put the GNU utilities in default system wide path before the native Sun utilities in order to make it easier to attract Linux users by making the environment more familiar to them. It was apparently assumed that longtime Solaris users would quickly and easily figure out what the problem was and adjust the PATH to their liking. Whether or not this was a good idea remains to be determined. From my vantage, it makes about as much sense to make Solaris more Linux-y in order to attract more Linux users while annoying current Solaris users as it does to change the speedometers on cars in the US to metric in order to make it more comfortable for immigrants and tourists to drive. fpsm ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS extended ACL
There was a very long discussion about this a couple of weeks ago on one of the lists. Apparently the decision was made to put the GNU utilities in default system wide path before the native Sun utilities in order to make it easier to attract Linux users by making the environment more familiar to them. It was apparently assumed that longtime Solaris users would quickly and easily figure out what the problem was and adjust the PATH to their liking. Well, it's OpenSOLARIS, comes with nice OpenSOLARIS goodies. Oooh ZFS ACL! rub hands together Goody! chmod A+user... gets slapped ls -V gets slapped OpenSOLARIS sucks! It's a quibble, but the way things are, it pleases no one, I don't think the casual Linux user moseying over to OpenSolaris would like the scenario above. CT ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Add SSD drive as L2ARC(?) cache to existing ZFS raid?
Kebabber, You can't expose zfs filesystems over iSCSI. You only can expose ZFS volumes (raw volumes) over iscsi. PS: 2 weeks ago I did a few tests, using filebench. I saw little to no improvement using a 32GB Intel X25E SSD. Maybe this is because filebench is flushing the cache in between tests. I also compared iscsi boot time (using gpxe as boot loader) , We are using raidz storagepool (4disks). here again, adding the X25E as cache device did not speedup the boot proccess. So I did not see real improvement. PS: We have 2 master volumes (xp and vista) which we clone to provision additional guests. I'm now waiting for new SSD disks (STEC Zeus 18GB en STEC Mach 100GB.), since those are used in SUN 7000 product. I hope they perform better. Kristof -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Add SSD drive as L2ARC(?) cache to existing ZFS raid?
How were you running this test? were you running it locally on the machine, or were you running it over something like NFS? What is the rest of your storage like? just direct-attached (SAS or SATA, for example) disks, or are you using a higher-end RAID controller? -Greg kristof wrote: Kebabber, You can't expose zfs filesystems over iSCSI. You only can expose ZFS volumes (raw volumes) over iscsi. PS: 2 weeks ago I did a few tests, using filebench. I saw little to no improvement using a 32GB Intel X25E SSD. Maybe this is because filebench is flushing the cache in between tests. I also compared iscsi boot time (using gpxe as boot loader) , We are using raidz storagepool (4disks). here again, adding the X25E as cache device did not speedup the boot proccess. So I did not see real improvement. PS: We have 2 master volumes (xp and vista) which we clone to provision additional guests. I'm now waiting for new SSD disks (STEC Zeus 18GB en STEC Mach 100GB.), since those are used in SUN 7000 product. I hope they perform better. Kristof ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] Problems with '..' on ZFS pool
Hello. I have a really weird problem with a ZFS pool on one machine, and it's only with 1 pool on that machine (the other pool is fine). Any non-root users cannot access '..' on any directories where the pool is mounted, eg: /a1000 on a1000 read/write/setuid/devices/nonbmand/exec/xattr/noatime/dev=4010002 on Wed Jan 28 20:55:38 2009 /home on a1000/home read/write/setuid/devices/nonbmand/exec/xattr/noatime/dev=4010005 on Wed Jan 28 20:55:39 2009 $ ls -ld / drwxr-xr-x 28 root root1024 Jan 29 10:09 / $ ls -ld /home drwxr-xr-x 11 root sys 11 Jan 9 14:49 /home $ ls -ld /home/.. /home/..: Permission denied $ ls -ld /a1000/.. /a1000/..: Permission denied $ ls -V / total 1065 drwxr-xr-x 2 root sys2 Dec 1 14:39 a1000 owner@:--:--:deny owner@:rwxp---A-W-Co-:--:allow group@:-w-p--:--:deny group@:r-x---:--:allow everyone@:-w-p---A-W-Co-:--:deny everyone@:r-x---a-R-c--s:--:allow drwxr-xr-x 6 root sys6 Aug 20 11:47 appl owner@:--:--:deny owner@:rwxp---A-W-Co-:--:allow group@:-w-p--:--:deny group@:r-x---:--:allow everyone@:-w-p---A-W-Co-:--:deny everyone@:r-x---a-R-c--s:--:allow lrwxrwxrwx 1 root root 9 Jun 18 2008 bin - ./usr/bin drwxr-xr-x 3 root sys 512 Jan 28 18:49 boot 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x drwxr-xr-x 19 root sys 7680 Jan 28 20:54 dev 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x drwxr-xr-x 2 root sys 512 Jan 28 20:53 devices 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x drwxr-xr-x 80 root sys 4608 Jan 29 09:40 etc 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x drwxr-xr-x 2 root sys 512 Jun 18 2008 export 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x drwxr-xr-x 11 root sys 11 Jan 9 14:49 home owner@:--:--:deny owner@:rwxp---A-W-Co-:--:allow group@:-w-p--:--:deny group@:r-x---:--:allow everyone@:-w-p---A-W-Co-:--:deny everyone@:r-x---a-R-c--s:--:allow drwxr-xr-x 15 root sys 512 Jun 18 2008 kernel 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x drwxr-xr-x 7 root bin 5632 Jan 28 19:50 lib 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x drwx-- 2 root root8192 Jun 18 2008 lost+found 0:user::rwx 1:group::--- #effective:--- 2:mask:--- 3:other:--- drwxr-xr-x 2 root sys 512 Jun 18 2008 mnt 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x dr-xr-xr-x 2 root root 512 Jun 18 2008 net 0:user::r-x 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x -rw-r--r-- 1 root root 0 Jun 18 2008 noautoshutdown 0:user::rw- 1:group::r-- #effective:r-- 2:mask:r-- 3:other:r-- drwxr-xr-x 7 root sys7 Jan 28 15:50 opt owner@:--:--:deny owner@:rwxp---A-W-Co-:--:allow group@:-w-p--:--:deny group@:r-x---:--:allow everyone@:-w-p---A-W-Co-:--:deny everyone@:r-x---a-R-c--s:--:allow drwxr-xr-x 40 root sys 1536 Jun 18 2008 platform 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x drwxr-xr-x 2 root sys2 Jul 29 2008 pool owner@:--:--:deny owner@:rwxp---A-W-Co-:--:allow group@:-w-p--:--:deny group@:r-x---:--:allow everyone@:-w-p---A-W-Co-:--:deny everyone@:r-x---a-R-c--s:--:allow dr-xr-xr-x 76 root root 480032 Jan 29 10:23 proc 0:user::r-x 1:group::r-x #effective:r-x 2:mask:rwx 3:other:r-x drwxr-x--- 12 root root1024 Jan 29 10:09 root 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:--- drwxr-xr-x 2 root sys 1024 Jan 28 19:37 sbin 0:user::rwx 1:group::r-x #effective:r-x 2:mask:r-x 3:other:r-x -rw-rw-rw- 1 root root1576 Oct 15 12:40 sybinit.err 0:user::rw- 1:group::rw- #effective:rw- 2:mask:rw-
Re: [zfs-discuss] Problems with '..' on ZFS pool
Forgot to add that a truss shows: 14960: lstat64(/a1000/.., 0xFFBFF7E8)Err#13 EACCES [file_dac_search] ppriv shows the error in UFS: $ ppriv -e -D -s -file_dac_search ls -ld /a1000/.. ls[15022]: missing privilege file_dac_search (euid = 100, syscall = 216) needed at ufs_iaccess+0x110 /a1000/..: Permission denied However seeing as it only happens for mounts on that 1 ZFS pool, it being a UFS problem seems highly unlikely. -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] destroy means destroy, right?
On Wed, 28 Jan 2009 20:16:37 -0500 Christine Tran christine.t...@gmail.com wrote: Everybody respects rm -f *. +1 -- Dick Hoogendijk -- PGP/GnuPG key: 01D2433D + http://nagual.nl/ | SunOS sxce snv105 ++ + All that's really worth doing is what we do for others (Lewis Carrol) ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Problems with '..' on ZFS pool
Dustin Marquess wrote: Forgot to add that a truss shows: 14960: lstat64(/a1000/.., 0xFFBFF7E8)Err#13 EACCES [file_dac_search] ppriv shows the error in UFS: $ ppriv -e -D -s -file_dac_search ls -ld /a1000/.. ls[15022]: missing privilege file_dac_search (euid = 100, syscall = 216) needed at ufs_iaccess+0x110 /a1000/..: Permission denied However seeing as it only happens for mounts on that 1 ZFS pool, it being a UFS problem seems highly unlikely. unmount the file system and look at the permission on the UFS mountpoint directory /a1000. They will probably be 0700 or something similar. -Mark ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS extended ACL
Christine Tran wrote: There was a very long discussion about this a couple of weeks ago on one of the lists. Apparently the decision was made to put the GNU utilities in default system wide path before the native Sun utilities in order to make it easier to attract Linux users by making the environment more familiar to them. It was apparently assumed that longtime Solaris users would quickly and easily figure out what the problem was and adjust the PATH to their liking. Well, it's OpenSOLARIS, comes with nice OpenSOLARIS goodies. Oooh ZFS ACL! rub hands together Goody! chmod A+user... gets slapped ls -V gets slapped OpenSOLARIS sucks! It's a quibble, but the way things are, it pleases no one, I don't think the casual Linux user moseying over to OpenSolaris would like the scenario above. As a previous long-time linux user who came over for ZFS, I totally agree. I much preferred to learn the solaris way and do things right than try and think it was still linux. Now I'm comfortable working on both despite their differences, and I'm sure I can perform tasks a lot better for it. Matt ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Problems with '..' on ZFS pool
Bingo, they were 0750. Thanks so much, that was the one thing I didn't think of. I thought I was going crazy :). Thanks again! -Dustin -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] firewire card?
I am pretty sure that Oxford 911 is a family of parts. The current Oxford Firewire parts are the 934 and 936 families. It appears that the Oxford 911 was commonly used in drive enclosures. The most troublesome part in my experience is the Initio INIC-1430. It does not get along with scsa1394 at all. -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS extended ACL
Yeah, breaking functionality in one of the main reasons people are going to be trying OpenSolaris is just dumb... really, really dumb. One thing Linux, Windows, OS/X, etc all get right is that they're pretty easy to use right out of the box. They're all different, but they all do their own jobs pretty well. So what do Sun do, make OpenSolaris harder to use... -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS extended ACL
On 29-Jan-09, at 2:17 PM, Ross wrote: Yeah, breaking functionality in one of the main reasons people are going to be trying OpenSolaris is just dumb... really, really dumb. One thing Linux, Windows, OS/X, etc all get right is that they're pretty easy to use right out of the box. They're all different, but they all do their own jobs pretty well. So what do Sun do, make OpenSolaris harder to use... I've not used OpenSolaris (yet), but I did spend quite a bit of time in Solaris 10 after using Linux and OS X for a long time. Matt's approach does work. Learn the differences, don't resent that they're the same, and get comfortable in both. Given the massive success of GNU based systems (Linux, OS X, *BSD) one can hardly fault OpenSolaris for taking this direction (I assume this is part of Ian Murdock's brief to make it a more appealing O/S option). Maybe Sun needs to make this optional, leaving the traditional SYSV flavour for the kneejerk anti-GNU elements (a bizarre phenomenon given SunOS' pure open source genesis ... Berkeley, Bill Joy, BSD, yadda yadda). --Toby -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Import Issues Btw 101 104
I just noticed that my previous assessment was not quite accurate. It's even stranger. Let's try again. On S10/b101, I have two pools: zpool list NAME SIZE USED AVAILCAP HEALTH ALTROOT double 172G 116G 56.3G67% ONLINE - single 516G 61.3G 455G11% ONLINE - double is mirrored. single is not. single has the following datasets: zfs list -r single NAME USED AVAIL REFER MOUNTPOINT single 61.3G 447G 1.50K none single/backup 1.55G 447G 1.93M /export/storage single/backup/cari18K 447G18K /export/home/cari/storage single/backup/dant 1.54G 447G 1.54G /export/home/dant/storage single/misc 54.0G 447G 24.5K none single/misc/dant54.0G 447G 24.5K none single/misc/dant/music 38.1G 447G 38.1G /export/home/dant/Music single/misc/dant/vbox 15.9G 447G 15.9G /export/home/dant/VMs single/share 18K 447G18K /export/share single/software 5.73G 447G 5.73G /usr/local After doing a zpool import from b104, I have: zpool list NAME SIZE USED AVAILCAP HEALTH ALTROOT double 516G 61.3G 455G11% ONLINE - and zpool import -D shows me that single is unavailable because it's corrupted because zpool thinks it's on the wrong partition. Same as what I wrote before. What I didn't notice before was a subtle difference in the zpool list. It's more obvious if we look at the zfs list: zfs list NAME USED AVAIL REFER MOUNTPOINT double 61.3G 447G 1.50K none double/backup 1.55G 447G 1.93M /export/storage double/backup/cari18K 447G18K /export/home/cari/storage double/backup/dant 1.54G 447G 1.54G /export/home/dant/storage double/misc 54.0G 447G 24.5K none double/misc/dant54.0G 447G 24.5K none double/misc/dant/music 38.1G 447G 38.1G /export/home/dant/Music double/misc/dant/vbox 15.9G 447G 15.9G /export/home/dant/VMs double/share 18K 447G18K /export/share double/software 5.73G 447G 5.73G /usr/local Notice that those are the datasets from single on b101! So, not only does b104 have the location of the single pool wrong, but it thinks the double pool is where the single pool is. WTF? Daniel Daniel Templeton wrote: Hi! I have a system with S10, b101, and b104 installed in the same partition on disk 1. On disks 1 and 2 in different partitions, I also created ZFS pools from S10 to be imported by b101 and b104. Pool 1 is mirrored. Pool 2 is not. About every three builds, I replace the oldest build with the latest available and switch to that as the default OS. Up through 101 everything was fine. I just installed 104, however, and when I do the zpool import, the mirrored pool is picked up just fine, but the non-mirrored pool shows up as corrupted. zpool import -D shows me that zpool on b104 thinks that the pool is on c1t1d0p2, whereas S10, b101, and fdisk agree that it's actually on c1t1d0p3. How do I convince zpool on b104 where my non-mirrored pool really is? I'm a bit afraid to do an import -f because that pool is my home directory, and I'd really rather not screw it up. And I don't see where the -f will change zpool's mind about where the pool actually lives. Maybe import -c from the default location? Where is the default location? Any thoughts or suggestions? Thanks! Daniel ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] destroy means destroy, right?
Maybe add a timer or something? When doing a destroy, ZFS will keep everything for 1 minute or so, before overwriting. This way the disk won't get as fragmented. And if you had fat fingers and typed wrong, you have up to one minute to undo. That will catch 80% of the mistakes? -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] destroy means destroy, right?
I like that, although it's a bit of an intelligence insulter. Reminds me of the old pdp11 install ( http://charles.the-haleys.org/papers/setting_up_unix_V7.pdf ) -- This step makes an empty file system. 6.The next thing to do is to restore the data onto the new empty file system. To do this you respond to the ':' printed in the last step with (bring in the program restor) : tm(0,4) ('ht(0,4)' for TU16/TE16) tape? tm(0,5) (use 'ht(0,5)' for TU16/TE16) disk? rp(0,0)(use 'hp(0,0)' for RP04/5/6) Last chance before scribbling on disk. (you type return) (the tape moves, perhaps 5-10 minutes pass) end of tape Boot : You now have a UNIX root file system. On Thu, Jan 29, 2009 at 3:42 PM, Orvar Korvar knatte_fnatte_tja...@yahoo.com wrote: Maybe add a timer or something? When doing a destroy, ZFS will keep everything for 1 minute or so, before overwriting. This way the disk won't get as fragmented. And if you had fat fingers and typed wrong, you have up to one minute to undo. That will catch 80% of the mistakes? ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] write cache and cache flush
So, I'm still beating my head against the wall, trying to find our performance bottleneck with NFS on our Thors. We've got a couple Intel SSDs for the ZIL, using 2 SSDs as ZIL devices. Cache flushing is still enabled, as are the write caches on all 48 disk devices. What I'm thinking of doing is disabling all write caches, and disabling the cache flushing. What would this mean for the safety of data in the pool? And, would this even do anything to address the performance issue? -Greg ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS extended ACL
Given the massive success of GNU based systems (Linux, OS X, *BSD) Ouch! Neither OSX nor *BSD are GNU-based. They do ship with GNU-related things but that's been a long and hard battle. And the massive success has really only been Linux due to brilliant PR (and FUD about *BSD) and OS X due to Apple's commercial approach to BSD. Regards -- Volker -- Volker A. Brandt Consulting and Support for Sun Solaris Brandt Brandt Computer GmbH WWW: http://www.bb-c.de/ Am Wiesenpfad 6, 53340 Meckenheim Email: v...@bb-c.de Handelsregister: Amtsgericht Bonn, HRB 10513 Schuhgröße: 45 Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS extended ACL
Volker A. Brandt wrote: Given the massive success of GNU based systems (Linux, OS X, *BSD) Ouch! Neither OSX nor *BSD are GNU-based. Not here, please. This topic has been beaten to death on the discuss list, where it's topical. -- Ian. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] destroy means destroy, right?
For years, we resisted stopping rm -r / because people should know better, until *finally* someone said - you know what - that's just dumb. Then, just like that, it was fixed. Yes - This is Unix. Yes - Provide the gun and allow the user to point it. Just don't let it go off in their groin or when pointed at their foot, or provide at least some protection when they do. Having even limited amount of restore capability will provide the user with steel capped boots and a codpiece. It won't protect them from herpes or fungus but it might deflect the bullet. On 01/30/09 08:19, Jacob Ritorto wrote: I like that, although it's a bit of an intelligence insulter. Reminds me of the old pdp11 install ( http://charles.the-haleys.org/papers/setting_up_unix_V7.pdf ) -- This step makes an empty file system. 6.The next thing to do is to restore the data onto the new empty file system. To do this you respond to the ':' printed in the last step with (bring in the program restor) : tm(0,4) ('ht(0,4)' for TU16/TE16) tape? tm(0,5) (use 'ht(0,5)' for TU16/TE16) disk? rp(0,0)(use 'hp(0,0)' for RP04/5/6) Last chance before scribbling on disk. (you type return) (the tape moves, perhaps 5-10 minutes pass) end of tape Boot : You now have a UNIX root file system. On Thu, Jan 29, 2009 at 3:42 PM, Orvar Korvar knatte_fnatte_tja...@yahoo.com wrote: Maybe add a timer or something? When doing a destroy, ZFS will keep everything for 1 minute or so, before overwriting. This way the disk won't get as fragmented. And if you had fat fingers and typed wrong, you have up to one minute to undo. That will catch 80% of the mistakes? ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss -- // // Nathan Kroenert nathan.kroen...@sun.com // // Senior Systems Engineer Phone: +61 3 9869 6255 // // Global Systems Engineering Fax:+61 3 9869 6288 // // Level 7, 476 St. Kilda Road // // Melbourne 3004 VictoriaAustralia // // ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] Extended attributes in ZFS
Does ZFS currently support actual use of extended attributes? If so, where can I find some documentation that describes how to use them? -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Extended attributes in ZFS
On Jan 29, 2009, at 18:02, Peter Reiher wrote: Does ZFS currently support actual use of extended attributes? If so, where can I find some documentation that describes how to use them? Your best bet would probably be: http://search.sun.com/docs/index.jsp?qt=zfs+extended+attributes Is there anything in particular you're wondering about? ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Extended attributes in ZFS
On Thu, Jan 29, 2009 at 03:02:50PM -0800, Peter Reiher wrote: Does ZFS currently support actual use of extended attributes? If so, where can I find some documentation that describes how to use them? man runat.1 openat.2 etcetera Nico -- ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Extended attributes in ZFS
Hi Peter, Yes, ZFS supports extended attributes. The runat.1 and fsattr.5 man pages are good places to start. Cindy Peter Reiher wrote: Does ZFS currently support actual use of extended attributes? If so, where can I find some documentation that describes how to use them? ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] ZFS extended ACL
On 29-Jan-09, at 4:53 PM, Volker A. Brandt wrote: Given the massive success of GNU based systems (Linux, OS X, *BSD) Ouch! Neither OSX nor *BSD are GNU-based. I meant, extensive GNU userland (in OS X's case). (sorry Ian) --Toby They do ship with GNU-related things but that's been a long and hard battle. And the massive success has really only been Linux due to brilliant PR (and FUD about *BSD) and OS X due to Apple's commercial approach to BSD. Regards -- Volker -- -- -- Volker A. Brandt Consulting and Support for Sun Solaris Brandt Brandt Computer GmbH WWW: http://www.bb- c.de/ Am Wiesenpfad 6, 53340 Meckenheim Email: v...@bb- c.de Handelsregister: Amtsgericht Bonn, HRB 10513 Schuhgröße: 45 Geschäftsführer: Rainer J. H. Brandt und Volker A. Brandt ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Extended attributes in ZFS
Nicolas Williams nicolas.willi...@sun.com wrote: On Thu, Jan 29, 2009 at 03:02:50PM -0800, Peter Reiher wrote: Does ZFS currently support actual use of extended attributes? If so, where can I find some documentation that describes how to use them? man runat.1 openat.2 etcetera Nico, the term extended attributes is overloaded. What you are talking of is called extended attribute files on Solaris. We should first ask the OP what he has in mind while asking. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] New RAM disk from ACARD might be interesting
ACARD have launched a new RAM disk which can take up to 64 GB of ECC RAM while still looking like a standard SATA drive. If anyone remember the Gigabyte I-RAM this might be a new development in this area. Its called ACARD ANS-9010 and up... http://www.acard.com.tw/english/fb01-product.jsp?idno_no=270prod_no=ANS-9010type1_title=%20Solid%20State%20Drivetype1_idno=13 This might be interesting to use as a cheap log instead of SSD cards... This test compares it with both Intel SSD (consumer and pro): http://www.techreport.com/articles.x/16255/1 However the test is more from a homeuser point of view... Anyone got the money and time to test it ;) -J -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] New RAM disk from ACARD might be interesting
As it presents as standard SATA, there should be no reason for this not to work... It has battery backup, and CF for backup / restore from DDR2 in the event of power loss... Pretty cool. (Would have preferred a super-cap, but oh, well... ;) Should make an excellent ZIL *and* L2ARC style device... Seems a little pricey for what it is though. It's going onto my list of what I'd buy if I had the money... ;) Nathan. On 01/30/09 12:10, Janåke Rönnblom wrote: ACARD have launched a new RAM disk which can take up to 64 GB of ECC RAM while still looking like a standard SATA drive. If anyone remember the Gigabyte I-RAM this might be a new development in this area. Its called ACARD ANS-9010 and up... http://www.acard.com.tw/english/fb01-product.jsp?idno_no=270prod_no=ANS-9010type1_title=%20Solid%20State%20Drivetype1_idno=13 This might be interesting to use as a cheap log instead of SSD cards... This test compares it with both Intel SSD (consumer and pro): http://www.techreport.com/articles.x/16255/1 However the test is more from a homeuser point of view... Anyone got the money and time to test it ;) -J -- // // Nathan Kroenert nathan.kroen...@sun.com // // Senior Systems Engineer Phone: +61 3 9869 6255 // // Global Systems Engineering Fax:+61 3 9869 6288 // // Level 7, 476 St. Kilda Road // // Melbourne 3004 VictoriaAustralia // // ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Problem with snapshot
Snapshots are not on a per-pool basis but a per-file-system basis. Thus, when you took a snapshot of testpol, you didn't actually snapshot the pool; rather, you took a snapshot of the top level file system (which has an implicit name matching that of the pool). Thus, you haven't actually affected file systems fs1 or fs2 at all. However, apparently you were able to roll back the file system, which either unmounted or broke the mounts to fs1 and fs2. This probably shouldn't have been allowed. (I wonder what would happen with an explicit non-ZFS mount to a ZFS directory which is removed by a rollback?) Your fs1 and fs2 file systems still exist, but they're not attached to their old names any more. Maybe they got unmounted. You could probably mount them, either on the fs1 directory and on a new fs2 directory if you create one, or at a different point in your file system hierarchy. Anton -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Problems with '..' on ZFS pool
That bug has been in Solaris forever. :-( -- This message posted from opensolaris.org ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] New RAM disk from ACARD might be interesting
On Thu, Jan 29, 2009 at 21:11, Nathan Kroenert nathan.kroen...@sun.com wrote: Seems a little pricey for what it is though. For what it's worth, there's also a 9010B model that has only one sata port and room for six dimms instead of eight at $250 instead of $400. That might fit in your budget a little easier... I'm considering one for a log device. I wish someone else could test it first and report problems, but someone's gotta take the jump first. It looks like this device (the 9010, that is) is also being marketed as the HyperDrive V at the same price point. Will ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] write cache and cache flush
Multiple Thors (more than 2?), with performance problems. Maybe it's the common demnominator - the network. Can you run local ZFS IO loads and determine if performance is expected when NFS and the network are out of the picture? Thanks, /jim Greg Mason wrote: So, I'm still beating my head against the wall, trying to find our performance bottleneck with NFS on our Thors. We've got a couple Intel SSDs for the ZIL, using 2 SSDs as ZIL devices. Cache flushing is still enabled, as are the write caches on all 48 disk devices. What I'm thinking of doing is disabling all write caches, and disabling the cache flushing. What would this mean for the safety of data in the pool? And, would this even do anything to address the performance issue? -Greg ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] New RAM disk from ACARD might be interesting
You could be the first... Man up! ;) Nathan. Will Murnane wrote: On Thu, Jan 29, 2009 at 21:11, Nathan Kroenert nathan.kroen...@sun.com wrote: Seems a little pricey for what it is though. For what it's worth, there's also a 9010B model that has only one sata port and room for six dimms instead of eight at $250 instead of $400. That might fit in your budget a little easier... I'm considering one for a log device. I wish someone else could test it first and report problems, but someone's gotta take the jump first. It looks like this device (the 9010, that is) is also being marketed as the HyperDrive V at the same price point. Will ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] New RAM disk from ACARD might be interesting
On Thu, Jan 29, 2009 at 22:44, Nathan Kroenert nathan.kroen...@sun.com wrote: You could be the first... Man up! ;) *sigh* The 9010b is ordered. Ground shipping, unfortunately, but eventually I'll post my impressions of it. Will Nathan. Will Murnane wrote: On Thu, Jan 29, 2009 at 21:11, Nathan Kroenert nathan.kroen...@sun.com wrote: Seems a little pricey for what it is though. For what it's worth, there's also a 9010B model that has only one sata port and room for six dimms instead of eight at $250 instead of $400. That might fit in your budget a little easier... I'm considering one for a log device. I wish someone else could test it first and report problems, but someone's gotta take the jump first. It looks like this device (the 9010, that is) is also being marketed as the HyperDrive V at the same price point. Will ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] write cache and cache flush
This problem only manifests itself when dealing with many small files over NFS. There is no throughput problem with the network. I've run tests with the write cache disabled on all disks, and the cache flush disabled. I'm using two Intel SSDs for ZIL devices. This setup is faster than using the two Intel SSDs with write caches enabled on all disks, and with the cache flush enabled. My test would run around 3.5 to 4 minutes, now it is completing in abound 2.5 minutes. I still think this is a bit slow, but I still have quite a bit of testing to perform. I'll keep the list updated with my findings. I've already established both via this list and through other research that ZFS has performance issues over NFS when dealing with many small files. This seems to maybe be an issue with NFS itself, where NVRAM-backed storage is needed for decent performance with small files. Typically such an NVRAM cache is supplied by a hardware raid controller in a disk shelf. I find it very hard to explain to a user why an upgrade is a step down in performance. For the users these Thors are going to serve, such a drastic performance hit is a deal breaker... I've done my homework on this issue, I've ruled out the network as an issue, as well as the NFS clients. I've narrowed my particular performance issue down to the ZIL, and how well ZFS plays with NFS. -Greg Jim Mauro wrote: Multiple Thors (more than 2?), with performance problems. Maybe it's the common demnominator - the network. Can you run local ZFS IO loads and determine if performance is expected when NFS and the network are out of the picture? Thanks, /jim ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] strange performance drop of solaris 10/zfs
Kevin, Looking at the stats I think the tank pool is about 80% full. And at this point you are possibly hitting the bug : 6596237 - Stop looking and start ganging Also, there is another ZIL related bug which worsens the case by fragmenting the space : 6683293 concurrent O_DSYNC writes to a fileset can be much improved over NFS You could compare the disk usage of the other machine that you have. Also, it would be useful to know what patch levels you are running. We do have IDRs for the bug#6596237 and the other bug has been fixed in the official patches. Hope that helps. Thanks and regards, Sanjeev. On Thu, Jan 29, 2009 at 01:13:29PM +0100, Kevin Maguire wrote: Hi We have been using a Solaris 10 system (Sun-Fire-V245) for a while as our primary file server. This is based on Solaris 10 06/06, plus patches up to approx May 2007. It is a production machine, and until about a week ago has had few problems. Attached to the V245 is a SCSI RAID array, which presents one LUN to the OS. On this lun is a zpool (tank), and within that 300+ zfs file systems (one per user for automounted home directories). The system is connected to our LAN via gigabit Ethernet,. most of our NFS clients have just 100FD network connection. In recent days performance of the file server seems to have gone off a cliff. I don't know how to troubleshoot what might be wrong? Typical zpool iostat 120 output is shown below. If I run truss -D df I see each call to statvfs64(/tank/bla) takes 2-3 seconds. The RAID itself is healthy, and all disks are reporting as OK. I have tried to establish if some client or clients are thrashing the server via nfslogd, but without seeing anything obvious. Is there some kind of per-zfs-filesystem iostat? End users are reporting just saving small files can take 5-30 seconds? prstat/top shows no process using significant CPU load. The system has 8GB of RAM, vmstat shows nothing interesting. I have another V245, with the same SCSI/RAID/zfs setup, and a similar (though a bit less) load of data and users where this problem is NOT apparent there? Suggestions? Kevin Thu Jan 29 11:32:29 CET 2009 capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank2.09T 640G 10 66 825K 1.89M tank2.09T 640G 39 5 4.80M 126K tank2.09T 640G 38 8 4.73M 191K tank2.09T 640G 40 5 4.79M 126K tank2.09T 640G 39 5 4.73M 170K tank2.09T 640G 40 3 4.88M 43.8K tank2.09T 640G 40 3 4.87M 54.7K tank2.09T 640G 39 4 4.81M 111K tank2.09T 640G 39 9 4.78M 134K tank2.09T 640G 37 5 4.61M 313K tank2.09T 640G 39 3 4.89M 32.8K tank2.09T 640G 35 7 4.31M 629K tank2.09T 640G 28 13 3.47M 1.43M tank2.09T 640G 5 51 433K 4.27M tank2.09T 640G 6 51 450K 4.23M tank2.09T 639G 5 52 543K 4.23M tank2.09T 640G 26 57 3.00M 1.15M tank2.09T 640G 39 6 4.82M 107K tank2.09T 640G 39 3 4.80M 119K tank2.09T 640G 38 8 4.64M 295K tank2.09T 640G 40 7 4.82M 102K tank2.09T 640G 43 5 4.79M 103K tank2.09T 640G 39 4 4.73M 193K tank2.09T 640G 39 5 4.87M 62.1K tank2.09T 640G 40 3 4.88M 49.3K tank2.09T 640G 40 3 4.80M 122K tank2.09T 640G 42 4 4.83M 82.0K tank2.09T 640G 40 3 4.89M 42.0K ... ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] write cache and cache flush
On 01/29/09 21:32, Greg Mason wrote: This problem only manifests itself when dealing with many small files over NFS. There is no throughput problem with the network. I've run tests with the write cache disabled on all disks, and the cache flush disabled. I'm using two Intel SSDs for ZIL devices. This setup is faster than using the two Intel SSDs with write caches enabled on all disks, and with the cache flush enabled. My test would run around 3.5 to 4 minutes, now it is completing in abound 2.5 minutes. I still think this is a bit slow, but I still have quite a bit of testing to perform. I'll keep the list updated with my findings. I've already established both via this list and through other research that ZFS has performance issues over NFS when dealing with many small files. This seems to maybe be an issue with NFS itself, where NVRAM-backed storage is needed for decent performance with small files. Typically such an NVRAM cache is supplied by a hardware raid controller in a disk shelf. I find it very hard to explain to a user why an upgrade is a step down in performance. For the users these Thors are going to serve, such a drastic performance hit is a deal breaker... Perhaps I missed something, but what was your previous setup? I.e. what did you upgrade from? Neil. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Issue with drive replacement
On Wed, 28 Jan 2009, Cuyler Dingwell wrote: In the process of replacing a raidz1 of four 500GB drives with four 1.5TB drives on the third one I ran into an interesting issue. The process was to remove the old drive, put the new drive in and let it rebuild. The problem was the third drive I put in had a hardware fault. That caused both drives (c4t2d0) to show as FAULTED. I couldn't put a new 1.5TB drive in as a replacement - it'd still show as a faulted drive. I couldn't remove the faulted since you can't remove a drive without enough replicas. You also can't do anything to a pool in the process of replacing. The remedy was to put the original drive back in and let it resilver. Once complete, a new 1.5TB drive was put in and the process was able to complete. If I didn't have the original drive (or it was broken) I think I would have been in a tough spot. Has anyone else experienced this - and if so, is there a way to force the replacement of drive that failed during resilvering? Not to my knowledge. You have to first cancel the replacement, and to do that you need the actual device (or something claiming it is) to be present. At least, I couldn't figure out how... ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss