[zfs-discuss] replace bad hard with new hard, on ZoL

[s***@gmail.com]

Hi Dears,
I install centos6.5 final and last zfs on VM ware workstation,

I set autoreplace=on and in zed.rc , uncomment :
ZED_SPARE_ON_IO_ERRORS=1 and ZED_SPARE_ON_IO_ERRORS=1

and reboot linux and create pool and add a spare hdd for it, and i remove
hdd in pool , but it dont replace with spare hdd,

can you help me for simple way for solved it?

Thanks my friends,

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[Durval Menezes]

Hello Saeed,

1) Are you sure zed is running?
2) What does "zpool history" shows?
3) Anything from zed in syslog?

Cheers,
--
Durval.

Post by s***@gmail.com
Hi Dears,
I install centos6.5 final and last zfs on VM ware workstation,
ZED_SPARE_ON_IO_ERRORS=1 and ZED_SPARE_ON_IO_ERRORS=1
and reboot linux and create pool and add a spare hdd for it, and i remove
hdd in pool , but it dont replace with spare hdd,
can you help me for simple way for solved it?
Thanks my friends,
To unsubscribe from this group and stop receiving emails from it, send an

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[s***@gmail.com]

Hi ,
1. I do ps -A and zed process in running ,
2. zpool history:
zpool create pool mirror sdb sdc spare sdd
3. dear where is syslog?
thanks ,

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[Durval Menezes]

Hello Gregor,

Post by s***@gmail.com
Hi Dears,
I install centos6.5 final and last zfs on VM ware workstation,
ZED_SPARE_ON_IO_ERRORS=1 and ZED_SPARE_ON_IO_ERRORS=1
and reboot linux and create pool and add a spare hdd for it, and i
remove hdd in pool , but it dont replace with spare hdd,
can you help me for simple way for solved it?

IIRC in current ZED removing a drive dosn't count as I/O error.

I didn't know that, but it does not sound very sensible. Is this intended?

Cheers,
--
Durval.

Maybe if you write some garbage over one of the drives (so it destroys
data on it) ZED will trigger.
Gregor
To unsubscribe from this group and stop receiving emails from it, send an

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[s***@gmail.com]

ok , i copy many files in it, and show in zpool status command :
state: DEGRADED
scan: none requested

but not auto replace with spare.

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[Durval Menezes]

Hello Saeed,

Post by s***@gmail.com
state: DEGRADED
scan: none requested
but not auto replace with spare.

The way I understand it, Gregor suggested (1) returning the removed
hard-disk to the pool, waiting for it to resilver, then writing garbage
*directly* on it (using dd onto the physical device, for example), then
accessing files from it (to force ZFS to detect the mashed disk).

If zed doesn't consider a disk removal as an event, it certainly won't
consider your reading files from the degraded pool as one...

Cheers,

[s***@gmail.com]

dears,do this stages is not complete for run autoreplace ?

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[Durval Menezes]

Hello Saeed,

Post by s***@gmail.com
dears,do this stages is not complete for run autoreplace ?

Huh? Can't understand what you mean. Can you please rephrase?

Cheers,

[s***@gmail.com]

I'm sorry,
I say , this stages for enable autoreplace is correct?
To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[Durval Menezes]

Hi Saeed,

Post by s***@gmail.com
I'm sorry,
I say , this stages for enable autoreplace is correct?

You probably mean to ask whether the steps you took to enable it are
correct, right?

If that's your question, I'd say yes. But see Gregor's reply, perhaps zed
simply doesn't do (yet?) what you want it to.

Cheers,

[s***@gmail.com]

yes my question was same,

when I remove a hdd from Virtual machine , and use zpool scrub pool ,

but it not replace with spare disk,

dear , i do :
ps -A
and show zed in process ,
do another way is for enable zed ?

thanks,

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[s***@gmail.com]

Dears,
Do it implement for ZoL , that it can replace automatically , or it is a
bug for ZoL?

Thanks,

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[Michael Kjörling]

Post by s***@gmail.com
Do it implement for ZoL , that it can replace automatically , or it is a
bug for ZoL?

Have a look at the discussion starting at [1] or [2]. (Same content,
slightly different formats.)

[1]: https://groups.google.com/a/zfsonlinux.org/forum/?_escaped_fragment_=topic/zfs-discuss/D5tL0JaqMA8#!topic/zfs-discuss/D5tL0JaqMA8

[2]: http://comments.gmane.org/gmane.linux.file-systems.zfs.user/17627

[s***@gmail.com]

thanks dear,
but I dont find any solution,
where i download script that said write and it is ready?

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[s***@gmail.com]

hi dears,
thanks for your helps,
I delete a hdd from VM and see logs, in log show that zed invoking
io-spare.sh but exit=2 ,
I see io-spare , it write exit = 2 meaning hot spareing disabled,
can you help me for

*how enabled hot spareing ,*

I set autoreplace for pool = on,
and in zed.rc un comment
ZED_SPARE_ON_IO_ERRORS=1 and ZED_SPARE_ON_IO_ERRORS=1

thanks for your help,

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[Luke Olson]

When I tried hot spares with ZFS on Linux 0.6.2 it didn't work and was
being developed. I haven't tried it with 0.6.3 but I don't think it's fully
implemented yet. There are pieces of it but not everything that is needed
is ready. Quote below as well in case the link or lines change.

https://github.com/zfsonlinux/zfs/blob/master/cmd/zed/zed.d/io-spare.sh#L19-21

# Full support for autoreplace is planned, but it requires that the full

# ZFS Diagnosis Engine be ported. In the meanwhile this script provides
# the majority of the expected hot spare functionality.

I setup a Python script to email me when there are problems with drives and
then I manually replace them. After doing it this way for a while I much
prefer it as I can troubleshoot things thoroughly before making changes to
the pool.

Luke

hi dears,
thanks for your helps,
I delete a hdd from VM and see logs, in log show that zed invoking
io-spare.sh but exit=2 ,
I see io-spare , it write exit = 2 meaning hot spareing disabled,
can you help me for
*how enabled hot spareing ,*
I set autoreplace for pool = on,
and in zed.rc un comment
ZED_SPARE_ON_IO_ERRORS=1 and ZED_SPARE_ON_IO_ERRORS=1
thanks for your help,
To unsubscribe from this group and stop receiving emails from it, send an

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[s***@gmail.com]

ok,
I download zfs-master and install it that include below script,
but for autoreplace show exit=2 in io-spare.sh file,
dears , for "...but it requires that the full ZFS Diagnosis Engine be
ported..."
how must do?

thanks,

Post by s***@gmail.com
hi dears,
thanks for your helps,
I delete a hdd from VM and see logs, in log show that zed invoking
io-spare.sh but exit=2 ,
I see io-spare , it write exit = 2 meaning hot spareing disabled,
can you help me for
*how enabled hot spareing ,*
I set autoreplace for pool = on,
and in zed.rc un comment
ZED_SPARE_ON_IO_ERRORS=1 and ZED_SPARE_ON_IO_ERRORS=1
thanks for your help,

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[Luke Olson]

It's not ready yet. Being developed. Incomplete.

Luke

Post by s***@gmail.com
ok,
I download zfs-master and install it that include below script,
but for autoreplace show exit=2 in io-spare.sh file,
dears , for "...but it requires that the full ZFS Diagnosis Engine be
ported..."
how must do?
thanks,

Post by s***@gmail.com
hi dears,
thanks for your helps,
I delete a hdd from VM and see logs, in log show that zed invoking
io-spare.sh but exit=2 ,
I see io-spare , it write exit = 2 meaning hot spareing disabled,
can you help me for
*how enabled hot spareing ,*
I set autoreplace for pool = on,
and in zed.rc un comment
ZED_SPARE_ON_IO_ERRORS=1 and ZED_SPARE_ON_IO_ERRORS=1
thanks for your help,

To unsubscribe from this group and stop receiving emails from it, send an

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.

[s***@gmail.com]

hi and thanks,
what are your opinion ?
from anyfriends who use it and get answer?

Post by Luke Olson
It's not ready yet. Being developed. Incomplete.
Luke

To unsubscribe from this group and stop receiving emails from it, send an email to zfs-discuss+***@zfsonlinux.org.