Tobias Hintze
2008-Jun-04 12:12 UTC
[zfs-discuss] panic on `zfs export` with UNAVAIL disks
hi list,
initial situation:
SunOS alusol 5.11 snv_86 i86pc i386 i86pc
SunOS Release 5.11 Version snv_86 64-bit
3 USB HDDs on 1 USB hub:
zpool status:
state: ONLINE
NAME STATE READ WRITE CKSUM
usbpool ONLINE 0 0 0
mirror ONLINE 0 0 0
c7t0d0p0 ONLINE 0 0 0
c8t0d0p0 ONLINE 0 0 0
c9t0d0p0 ONLINE 0 0 0
i shut the machine down and replugged the USB HDDs to not use the USB
hub but to connect directly to the machine.
after booting the machine i have the following zpool status:
state: DEGRADED
NAME STATE READ WRITE CKSUM
usbpool DEGRADED 0 0 0
mirror UNAVAIL 0 0 0 insufficient replicas
c8t0d0p0 UNAVAIL 0 0 0 cannot open
c7t0d0p0 UNAVAIL 0 0 0 cannot open
c10t0d0p0 ONLINE 0 0 0
i should propably have done a zfs export before the replugging, but i
expected that zfs would somehow be able to find the disks.
to fix the problem i tried to `zfs export` from this status.
this resulted in the kernel panic:
=============Jun 4 11:31:47 alusol unix: [ID 836849 kern.notice]
Jun 4 11:31:47 alusol ^Mpanic[cpu0]/thread=ffffff0004b1ac80:
Jun 4 11:31:47 alusol genunix: [ID 603766 kern.notice] assertion failed:
vdev_config_sync(rvd->vdev_child, rvd->vdev_children, txg) == 0 (0x5 ==
0x0), file: ../../common/fs/zfs/spa.c, line: 4095
Jun 4 11:31:47 alusol unix: [ID 100000 kern.notice]
Jun 4 11:31:47 alusol genunix: [ID 655072 kern.notice] ffffff0004b1ab30
genunix:assfail3+b9 ()
Jun 4 11:31:47 alusol genunix: [ID 655072 kern.notice] ffffff0004b1abd0
zfs:spa_sync+5d2 ()
Jun 4 11:31:47 alusol genunix: [ID 655072 kern.notice] ffffff0004b1ac60
zfs:txg_sync_thread+19a ()
Jun 4 11:31:47 alusol genunix: [ID 655072 kern.notice] ffffff0004b1ac70
unix:thread_start+8 ()
=============
after being back up i could do `zfs import` and have the pool back:
state: ONLINE
NAME STATE READ WRITE CKSUM
usbpool ONLINE 0 0 0
mirror ONLINE 0 0 0
c12t0d0p0 ONLINE 0 0 0
c13t0d0p0 ONLINE 0 0 0
c10t0d0p0 ONLINE 0 0 0
errors: 4 data errors, use ''-v'' for a list
with the 4 errors:
errors: Permanent errors have been detected in the following files:
<metadata>:<0x3>
<metadata>:<0xa>
<metadata>:<0x13>
<metadata>:<0x14>
so the questions:
1) was there something else i could have done instead of the
`zfs export` + panic + `zfs import` from that situation
(after replugging and booting)?
2) how do i get rid of the metadata errors?
thanks,
--
Tobias Hintze // threllis GmbH
Tobias Hintze
2008-Jun-04 15:47 UTC
[zfs-discuss] panic on `zfs export` with UNAVAIL disks
On Wed, Jun 04, 2008 at 02:12:10PM +0200, Tobias Hintze wrote: [...]> > 2) how do i get rid of the metadata errors? >scrubbing fixed the metadata errors. th --> > thanks, > -- > Tobias Hintze // threllis GmbH-- Tobias Hintze // threllis GmbH http://threllis.de/impressum/
Richard Elling
2008-Jun-04 16:31 UTC
[zfs-discuss] panic on `zfs export` with UNAVAIL disks
Tobias Hintze wrote:> hi list, > > initial situation: > > SunOS alusol 5.11 snv_86 i86pc i386 i86pc > SunOS Release 5.11 Version snv_86 64-bit > > 3 USB HDDs on 1 USB hub: > > zpool status: > > state: ONLINE > NAME STATE READ WRITE CKSUM > usbpool ONLINE 0 0 0 > mirror ONLINE 0 0 0 > c7t0d0p0 ONLINE 0 0 0 > c8t0d0p0 ONLINE 0 0 0 > c9t0d0p0 ONLINE 0 0 0 > > i shut the machine down and replugged the USB HDDs to not use the USB > hub but to connect directly to the machine. > > after booting the machine i have the following zpool status: > > state: DEGRADED > NAME STATE READ WRITE CKSUM > usbpool DEGRADED 0 0 0 > mirror UNAVAIL 0 0 0 insufficient replicas > c8t0d0p0 UNAVAIL 0 0 0 cannot open > c7t0d0p0 UNAVAIL 0 0 0 cannot open > c10t0d0p0 ONLINE 0 0 0 > > i should propably have done a zfs export before the replugging, but i > expected that zfs would somehow be able to find the disks. > > > to fix the problem i tried to `zfs export` from this status. > this resulted in the kernel panic: > =============> Jun 4 11:31:47 alusol unix: [ID 836849 kern.notice] > Jun 4 11:31:47 alusol ^Mpanic[cpu0]/thread=ffffff0004b1ac80: > Jun 4 11:31:47 alusol genunix: [ID 603766 kern.notice] assertion failed: vdev_config_sync(rvd->vdev_child, rvd->vdev_children, txg) == 0 (0x5 == 0x0), file: ../../common/fs/zfs/spa.c, line: 4095 >This looks like CR6616599 which is fixed in NV b90. http://bugs.opensolaris.org/view_bug.do?bug_id=6616599 -- richard