zfs discuss - Jan 2010 - (snv_129, snv_130) can''t import zfs pool

Hi 

(snv_130) created zfs pool storage (a mirror of two whole disks)

zfs created storage/iscsivol,  made some tests, wrote some GBs

zfs created storage/mynas filesystem
(sharesmb
dedup=on
compression=on)

FILLED the storage/mynas


tried to ZFS DESTROY my storage/iscsivol, but the system has HUNG...
this system now tends to boot to maintenance mode due to the boot-archive
corruption

The pool can''t be imported -f by the recent EON storage (snv_129), it
hangs also and don''t return to the CLI


Any help is appreciated
-- 
This message posted from opensolaris.org

On Jan 1, 2010, at 4:57 AM, LevT wrote:
> Hi
>
> (snv_130) created zfs pool storage (a mirror of two whole disks)
>
> zfs created storage/iscsivol,  made some tests, wrote some GBs
>
> zfs created storage/mynas filesystem
> (sharesmb
> dedup=on
> compression=on)
>
> FILLED the storage/mynas
>
>
> tried to ZFS DESTROY my storage/iscsivol, but the system has HUNG...
dedup is still new and several people have reported that destroying  
deduped
datasets can take a long time. Plenty of memory or cache devices seems  
to help,
as does having high IOPS drives in the main pool.  Otherwise, you''ll  
have to
wait for it to finish.
> this system now tends to boot to maintenance mode due to the boot- 
> archive corruption
This is unrelated to the above problem.  More likely this occurred  
when you
gave up and forced a restart. Follow the standard instructions for  
rebuilding
the boot archive.
  -- richard
> The pool can''t be imported -f by the recent EON storage (snv_129),
> it hangs also and don''t return to the CLI
>
>
> Any help is appreciated
> -- 
> This message posted from opensolaris.org
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss

You might want to checkout another thread that me and some of the others started
on this topic. some of the guys in that thread got their pool back but I
haven''t been able to.  I have SSDs for my log and cache and it
hasn''t helped me because my system hangs hard on import the way you are
describing. Thus far I still haven''t been able to regain my pool after
switching to a totally different system with more memory, but some of the other
guys have.
-- 
This message posted from opensolaris.org

switched to another system, RAM 4Gb -> 16Gb

the importing process lasts about 18hrs now
the system is responsive

if developers want it I may provide ssh access
I have no critical data there, it is an acceptance test only
-- 
This message posted from opensolaris.org

Bump.

Any devs want to take him up on his offer, obviously this is effecting a few
users and judging from the view counts of the other threads about this problem,
many more. This would probably effect the 7000 series as well.

Thanks.
-- 
This message posted from opensolaris.org

LevT wrote:
> switched to another system, RAM 4Gb -> 16Gb
> 
> the importing process lasts about 18hrs now
> the system is responsive
> 
> if developers want it I may provide ssh access
> I have no critical data there, it is an acceptance test only
If it is still relevant, feel free to contact me offline to setup ssh 
access.

regards,
victor

Just curious if anything has happened here.

I had a similar issue that was solved by upgrading from 4GB to 8GB of RAM.

I now have the issue again, and my box hard locks when doing the import after
about 30 minutes. (This time not using de-dup, but was using iscsi). I debated
on upgrading to 16GB of RAM, but can''t justify the cost.

Hoping there is some sort of bug found and fixed in a future release so that I
may get my 4.5 TB pool back.
-- 
This message posted from opensolaris.org

I''d like to thank Tim and Cindy at Sun for providing me with a new zfs
binary file that fixed my issue. I was able to get my zpool back! Hurray!

Thank You.
-- 
This message posted from opensolaris.org

Thanks Jack,

I was just a listener in this case. Tim did all the work. :-)

Cindy

On 01/23/10 21:49, Jack Kielsmeier wrote:
> I''d like to thank Tim and Cindy at Sun for providing me with a new
zfs binary file that fixed my issue. I was able to get my zpool back! Hurray!
> 
> Thank You.