Hi all, there are some "new" error messages on our MDT, haven''t seen these before and according to Google nobody else has... The usual question: what does it mean? Something to worry about? > Jun 7 06:23:53 lxmds kernel: [4565451.097596] LustreError: 9998:0:(obd.h:1372:lsm_op_find()) Cannot recognize lsm_magic 0 > Jun 7 06:23:53 lxmds kernel: [4565451.181790] LustreError: 9998:0:(lov_pack.c:278:lov_verify_lmm()) bad disk LOV MAGIC: 0x00000000; dumping LMM (size=0): > Jun 7 06:23:53 lxmds kernel: [4565451.266723] LustreError: 9998:0:(lov_pack.c:287:lov_verify_lmm()) > Jun 7 06:23:53 lxmds kernel: [4565451.350904] LustreError: 9998:0:(obd.h:1372:lsm_op_find()) Cannot recognize lsm_magic 0 > Jun 7 06:23:53 lxmds kernel: [4565451.434894] LustreError: 9998:0:(lov_pack.c:278:lov_verify_lmm()) bad disk LOV MAGIC: 0x00000000; dumping LMM (size=0): > Jun 7 06:23:53 lxmds kernel: [4565451.519946] LustreError: 9998:0:(lov_pack.c:287:lov_verify_lmm()) Regards, Thomas
Hello! On Jun 7, 2011, at 7:49 AM, Thomas Roth wrote:> there are some "new" error messages on our MDT, haven''t seen these > before and according to Google nobody else has... > The usual question: what does it mean? Something to worry about? >> Jun 7 06:23:53 lxmds kernel: [4565451.097596] LustreError: > 9998:0:(obd.h:1372:lsm_op_find()) Cannot recognize lsm_magic 0 >> Jun 7 06:23:53 lxmds kernel: [4565451.181790] LustreError: > 9998:0:(lov_pack.c:278:lov_verify_lmm()) bad disk LOV MAGIC: 0x00000000; > dumping LMM (size=0):Bug 23615 in hte bugzilla was about it, but I don''t remember the details of it and the bug is "private". No patches landed as the result of that bug, though. If Oracle or the filing party of the bug can step forward and remind us what was in that bug.... Bye, Oleg
Hi. bz 23615 was from Cray. I''m checking to see whether we can open that bug up or not. It''s a little fuzzy to me now, but it may have been related to bz 19934. If I recall correctly, there was some other root cause that was causing stale unlinked orphans to reside on the MDT and after we removed the offending objects from the /PENDING directory, the problem didn''t reoccur. Thanks, -Cory On 6/7/2011 6:52 PM, Oleg Drokin wrote:> Hello! > > On Jun 7, 2011, at 7:49 AM, Thomas Roth wrote: >> there are some "new" error messages on our MDT, haven''t seen these >> before and according to Google nobody else has... >> The usual question: what does it mean? Something to worry about? >>> Jun 7 06:23:53 lxmds kernel: [4565451.097596] LustreError: >> 9998:0:(obd.h:1372:lsm_op_find()) Cannot recognize lsm_magic 0 >>> Jun 7 06:23:53 lxmds kernel: [4565451.181790] LustreError: >> 9998:0:(lov_pack.c:278:lov_verify_lmm()) bad disk LOV MAGIC: 0x00000000; >> dumping LMM (size=0): > > Bug 23615 in hte bugzilla was about it, but I don''t remember the details of it > and the bug is "private". > No patches landed as the result of that bug, though. > If Oracle or the filing party of the bug can step forward and remind us > what was in that bug.... > > Bye, > Oleg > _______________________________________________ > Lustre-discuss mailing list > Lustre-discuss at lists.lustre.org > http://lists.lustre.org/mailman/listinfo/lustre-discuss