Hi,
On 12/02/2015 08:12 PM, Gregory Farnum wrote:
On Wed, Dec 2, 2015 at 11:11 AM, Major Csaba <major.csaba@xxxxxxxxxxx> wrote:
Hi,
[ sorry, I accidentaly left out the list address ]
This is the content of the LOG file in the directory
/var/lib/ceph/osd/ceph-7/current/omap:
2015/12/02-18:48:12.241386 7f805fc27900 Recovering log #26281
2015/12/02-18:48:12.242455 7f805fc27900 Level-0 table #26283: started
2015/12/02-18:48:12.274615 7f805fc27900 Level-0 table #26283: 32841 bytes OK
2015/12/02-18:48:12.352606 7f805fc27900 Delete type=2 #26282
2015/12/02-18:48:12.353143 7f805fc27900 Delete type=2 #26284
2015/12/02-18:48:12.353657 7f805fc27900 Delete type=2 #26285
2015/12/02-18:48:12.354145 7f805fc27900 Delete type=3 #26279
2015/12/02-18:48:12.354183 7f805fc27900 Delete type=0 #26281
2015/12/02-18:48:12.354343 7f80541a4700 Compacting 14@0 + 5@1 files
2015/12/02-18:48:12.497244 7f80541a4700 Generated table #26285: 116539 keys,
2134818 bytes
2015/12/02-18:48:12.653044 7f80541a4700 Generated table #26286: 173429 keys,
2130210 bytes
2015/12/02-18:48:12.819800 7f80541a4700 Generated table #26287: 172490 keys,
2130006 bytes
2015/12/02-18:48:12.894106 7f80541a4700 compacted to: files[ 14 5 8 0 0 0 0
]
2015/12/02-18:48:12.894115 7f80541a4700 Compaction error: Corruption:
corrupted compressed block contents
I don't know if it's relevant, but seems wrong.
Uh, yeah. Seems something has gone terribly wrong in or under leveldb.
I'm not sure if there's any good way to repair that, or if you just
need to figure out how it died (to prevent future recurrences) and
rebuild the OSD. :/
-Greg
We had some RAM error earlier on this node, we replaced the HW but we
kept the disks with their content. So, I can imagine the content is
corrupted earlier by the bad RAM module. So, I'll rebuild this OSD.
Thanks for your help.
Regards,
Csaba
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
