zfs discuss - Jan 2010 - Hang zpool detach while scrub is running..

I know, we should have done zpool scrub -s first.. but.. sigh..

bitshop at zfs:/opt/StorMan# zpool status -v tankmir1
  pool: tankmir1
 state: ONLINE
 scrub: scrub in progress for 0h16m, 0.14% done, 187h17m to go
config:

        NAME        STATE     READ WRITE CKSUM
        tankmir1    ONLINE       0     0     0
          mirror-0  ONLINE       0     0     0
            c7t6d0  ONLINE       0     0     0
            c7t7d0  ONLINE       0     0     0

errors: No known data errors
bitshop at zfs:/opt/StorMan# zpool detach tankmir1 c7t7d0

(hung).

At this point a new SSH session to the server hangs during login, I was
connected via KVM over IP to the console and don''t seem to have any
problem with that session (although I''m not trying to log off and back
in).

iostat shows all activity on c7t6d0 (as expected), however IO is extremely slow
(< 1 megabyte / second and 100% busy). We''ve been fighting a slow
i/o problem with Seagate ES2 drives, some needed firmware flashing which we
didn''t catch before they were in a pool, so remove, flash, reinstall,
resilver, etc. is a long process.

Anything I can try to dump that''s not overly intrusive? The system
seems to be working still for iSCSI and CIFS which is it''s purpose, so
a reboot isn''t planned unless this hangs in more ways.  Hopefully it
will respond in a while.

snv129 installed.


Steve Radich - Founder and Principal of Business Information Technology Shop -
bitshop.com
Developer Resources Site: ASPDeveloper.Net  - VirtualServerFAQ.com 
  LinkedIn Public Profile: linkedin.com/in/steveradich
-- 
This message posted from opensolaris.org

It came back right after I posted (had waited SEVERAL minutes). 

Perhaps related to dedup on slow downs (?).
-- 
This message posted from opensolaris.org