Hi This weekend I upgraded my blade 1500 to b46 from b39, when importing my pools, one raidz and another a single slice with -f because I forgot to export the drives as usual, zpool dropped a core, file somewhere after the pools were imported and before the filessysttems were mounted. I ran zfs mount -a, and all zfs filesystems are now mounted. If you are interested in it. I got a backtrace below. bash-3.00# pstack core.zpool.import core ''core.zpool.import'' of 644: zpool import -f pool ff348adc zfs_prop_get (0, d, ffbfaef0, 400, 0, 0) + 14 ff34ed88 dataset_compare (c3f90, c3f9c, ffbfa2f0, 472a4, 14, ff396000) + 30 ff12f798 qsort (c3f90, c3fa4, 4, ff34ed58, c3f90, c3fa4) + 44c ff34eebc zpool_mount_datasets (ffffffff, 0, 0, 468b8, 471f4, 73388) + b4 00014854 do_import (d9e98, eb9c8, 0, 20800, 1, 1c400) + 15c 00014ed4 zpool_do_import (0, 1, 0, 75fe8, 1c800, 0) + 5dc 00018898 main (4, ffbff69c, 20860, 20800, ffbff7aa, 20400) + 148 00012858 _start (0, 0, 0, 0, 0, 0) + 108 bash-3.00# Let me know if anyone has a core file fetish ;-) and wants to see it. James Dickens uadmin.blogspot.com
On 8/26/06, James Dickens <jamesd.wi at gmail.com> wrote:> bash-3.00# pstack core.zpool.import > core ''core.zpool.import'' of 644: zpool import -f pool > ff348adc zfs_prop_get (0, d, ffbfaef0, 400, 0, 0) + 14 > ff34ed88 dataset_compare (c3f90, c3f9c, ffbfa2f0, 472a4, 14, ff396000) + 30 > ff12f798 qsort (c3f90, c3fa4, 4, ff34ed58, c3f90, c3fa4) + 44c > ff34eebc zpool_mount_datasets (ffffffff, 0, 0, 468b8, 471f4, 73388) + b4 > 00014854 do_import (d9e98, eb9c8, 0, 20800, 1, 1c400) + 15c > 00014ed4 zpool_do_import (0, 1, 0, 75fe8, 1c800, 0) + 5dc > 00018898 main (4, ffbff69c, 20860, 20800, ffbff7aa, 20400) + 148 > 00012858 _start (0, 0, 0, 0, 0, 0) + 108 > bash-3.00#FWIW, I saw the same backtrace on build 46 doing some wierd stuff documented at http://mgerdts.blogspot.com/. At the time I was booted from cdrom media importing a pool that I had previously exported. Mike -- Mike Gerdts http://mgerdts.blogspot.com/
On 8/26/06, Mike Gerdts <mgerdts at gmail.com> wrote:> FWIW, I saw the same backtrace on build 46 doing some wierd stuff > documented at http://mgerdts.blogspot.com/. At the time I was booted > from cdrom media importing a pool that I had previously exported.I got thinking... how can I outdo the "ME TOO!" post. Ahhh... I could reply to my own message. That will get me a free AOL floppy for sure! This problem with zpool seems to be logged as a bug already. http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6461922 Mike -- Mike Gerdts http://mgerdts.blogspot.com/
Yes, and the fix is in the last codereview I sent out. Hopefully someone from the team can review it and it will make build 48. In the meantime, ''zfs mount -a'' after importing the pool will get you around this problem. Sorry bout that. - Eric On Sat, Aug 26, 2006 at 08:46:50PM -0500, Mike Gerdts wrote:> On 8/26/06, Mike Gerdts <mgerdts at gmail.com> wrote: > >FWIW, I saw the same backtrace on build 46 doing some wierd stuff > >documented at http://mgerdts.blogspot.com/. At the time I was booted > >from cdrom media importing a pool that I had previously exported. > > I got thinking... how can I outdo the "ME TOO!" post. Ahhh... I could > reply to my own message. That will get me a free AOL floppy for sure! > > This problem with zpool seems to be logged as a bug already. > > http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6461922 > > Mike > > -- > Mike Gerdts > http://mgerdts.blogspot.com/ > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss-- Eric Schrock, Solaris Kernel Development http://blogs.sun.com/eschrock