[Drbd-dev] DRBD8: drbd nodes deadlock in WFBitMapT

[Montrose, Ernest]

Hi all,
This is another hard to reproduce one but the proofs are in the logs
that the problem is alive and well.
I am hoping for at least some clues that may help reproduce
this...Essentially after one node in 
Primary state is powered down(Not a graceful shutdown)  both nodes ends
up  In WFBitMapT for a drbd volume.
 
Here are some logs:
On one node========================


Drbd2: Writing metadata to superblock now.
......
....... This nodes is powered of and came back with:

Mar 23 13:16:53 jerry kernel: drbd2: rct = 0 in
/test_logs/builds/SuperNova/trunk/070323/platform/drbd/src/drbd/drbd_rec
eiver.c:1878
Mar 23 13:16:55 [  OK  ]
Mar 23 13:16:54 jerry kernel: drbd2: drbd_sync_handshake:
Mar 23 13:16:54 jerry kernel: drbd2: self
F71E503A8179BC5D:0000000000000000:3F430D4E1D59C3EA:6B0B1DA5CB20689C
Mar 23 13:16:54 jerry kernel: drbd2: peer
F71E503A8179BC5C:0000000000000000:3F430D4E1D59C3EA:6B0B1DA5CB20689C
Mar 23 13:16:55 jerry kernel: drbd2: uuid_compare()=0 by rule 4
Mar 23 13:16:55 jerry kernel: drbd2: No resync, but bits in bitmap!

......

Mar 23 13:17:00 jerry kernel: drbd2: drbd_sync_handshake:
Mar 23 13:17:00 jerry kernel: drbd2: self
F71E503A8179BC5D:0000000000000000:3F430D4E1D59C3EA:6B0B1DA5CB20689C
Mar 23 13:17:00 jerry kernel: drbd2: peer
0000000000000000:0000000000000000:F71E503A8179BC5C:3F430D4E1D59C3EA
Mar 23 13:17:00 jerry kernel: drbd2: uuid_compare()=-2 by rule 6
Mar 23 13:17:00 jerry kernel: drbd2: Writing meta data super block now.
Mar 23 13:17:01 jerry kernel: drbd2: writing of bitmap took 11 jiffies
Mar 23 13:17:01 jerry kernel: drbd2: 12 GB marked out-of-sync by on disk
bit-map.
Mar 23 13:17:02 jerry kernel: drbd2: 13336132 KB now marked out-of-sync
by on disk bit-map.
Mar 23 13:17:02 jerry kernel: drbd2: Writing meta data super block now.
Mar 23 13:17:02 jerry kernel: drbd2:  uuid[History_start] now
F71E503A8179BC5D
Mar 23 13:17:02 jerry kernel: drbd2:  uuid[Current] now 0000000000000000
Mar 23 13:17:03 jerry kernel: drbd2: conn( Connected -> WFBitMapT )
Mar 23 13:17:03 jerry kernel: drbd2: Writing meta data super block now.

On theh other node=============================

Mar 23 13:16:48 ben kernel: drbd2: aftr_isp( 0 -> 1 )
Mar 23 13:16:48 ben kernel: drbd2: Handshake successful: DRBD Network
Protocol version 86
Mar 23 13:16:48 ben kernel: drbd2: peer( Unknown -> Secondary ) conn(
WFReportParams -> Connected ) pdsk( DUnknown -> UpToDate ) peer_isp( 0
-> 1 )
Mar 23 13:16:48 ben kernel: drbd2: Writing meta data super block now.

....

Mar 23 13:16:49 ben kernel: drbd2: rct = 2 in
/test_logs/builds/SuperNova/trunk/070323/platform/drbd/src/drbd/drbd_rec
eiver.c:1878
Mar 23 13:16:49 ben kernel: drbd2: drbd_sync_handshake:
Mar 23 13:16:49 ben kernel: drbd2: self
F71E503A8179BC5C:0000000000000000:3F430D4E1D59C3EA:6B0B1DA5CB20689C
Mar 23 13:16:49 ben kernel: drbd2: peer
F71E503A8179BC5D:0000000000000000:3F430D4E1D59C3EA:6B0B1DA5CB20689C
Mar 23 13:16:49 ben kernel: drbd2: uuid_compare()=-1 by rule 4
Mar 23 13:16:49 ben kernel: drbd2:  uuid[History_start] now
F71E503A8179BC5C
Mar 23 13:16:49 ben kernel: drbd2:  uuid[Current] now 0000000000000000
Mar 23 13:16:49 ben kernel: drbd2: conn( Connected -> WFBitMapT )
Mar 23 13:16:49 ben kernel: drbd2: Writing meta data super block now.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.linbit.com/pipermail/drbd-dev/attachments/20070330/588748b1/attachment.htm