[LU-9828] LBUG ASSERTION( desc->bd_nob_transferred == 0 ) failed:

• Type: Bug
• Resolution: Fixed
• Priority: Major
• Fix Version/s: Lustre 2.10.1, Lustre 2.11.0
• Severity: 3
• Assignee: Amir Shehata (Inactive)
• Reporter: Minh Diep

Description

One of clients crashed due to the following LBUG.

LustreError: 11818:0:(events.c:201:client_bulk_callback()) event type 2, status -103, desc ffff880827971600
LustreError: 11840:0:(niobuf.c:329:ptlrpc_register_bulk()) ASSERTION( desc->bd_nob_transferred == 0 ) failed:
LustreError: 11818:0:(events.c:201:client_bulk_callback()) event type 2, status -103, desc ffff880d40623400
Lustre: yshare1-OST0023-osc-ffff882049a1c800: Connection to yshare1-OST0023 (at 172.28.8.204@o2ib1) was lost; in progress operations using this service will wait for recovery to complete
Lustre: Skipped 21 previous similar messages
LNet: 11818:0:(o2iblnd_cb.c:1364:kiblnd_reconnect_peer()) Abort reconnection of 172.28.8.204@o2ib1: connected
LNet: 11818:0:(o2iblnd_cb.c:1364:kiblnd_reconnect_peer()) Skipped 1 previous similar message
LustreError: 11840:0:(niobuf.c:329:ptlrpc_register_bulk()) LBUG
Pid: 11840, comm: ptlrpcd_01_01

Call Trace:
 [<ffffffffa0967895>] libcfs_debug_dumpstack+0x55/0x80 [libcfs]
 [<ffffffffa0967e97>] lbug_with_loc+0x47/0xb0 [libcfs]
 [<ffffffffa0cae07c>] ptlrpc_register_bulk+0xfc/0x9c0 [ptlrpc]
 [<ffffffffa0985c74>] ? cfs_percpt_unlock+0x24/0xb0 [libcfs]
 [<ffffffffa0a1b7b4>] ? LNetMDUnlink+0xd4/0x160 [lnet]
 [<ffffffffa0cb5c64>] ? lustre_msg_get_flags+0x34/0xb0 [ptlrpc]
 [<ffffffffa0caf5af>] ptl_send_rpc+0x1af/0xea0 [ptlrpc]
 [<ffffffffa0ce6804>] ? sptlrpc_req_refresh_ctx+0x154/0x910 [ptlrpc]
 [<ffffffffa0ca90b2>] ptlrpc_check_set+0x1462/0x1bf0 [ptlrpc]
 [<ffffffffa0cd6d83>] ptlrpcd_check+0x3d3/0x610 [ptlrpc]
 [<ffffffffa0cd7232>] ptlrpcd+0x272/0x4f0 [ptlrpc]
 [<ffffffff8106c500>] ? default_wake_function+0x0/0x20
 [<ffffffffa0cd6fc0>] ? ptlrpcd+0x0/0x4f0 [ptlrpc]
 [<ffffffff810a640e>] kthread+0x9e/0xc0
 [<ffffffff8100c28a>] child_rip+0xa/0x20
 [<ffffffff810a6370>] ? kthread+0x0/0xc0
 [<ffffffff8100c280>] ? child_rip+0x0/0x20

 

Issue Links

• is related to: LU-10799 ASSERTION(desc->bd_nob_transferred == 0)

Comments

[Peter Jones]

Thanks askulysh. For future reference we can just update the commit message without losing positive testing and reviews so making these corrections does not require abandoning patches.

[Andriy Skulysh]

Opened LU-10799

[Peter Jones]

Cory

A new ticket linked to this one please. It causes no end of confusion when patches are tagged onto long-closed tickets.

Peter

[Cory Spitz]

This issue is marked RESOLVED, yet https://review.whamcloud.com/#/c/30368 is still linked to here. Should we get a new ticket or should this issue be reopened?

[Gerrit Updater]

Andriy Skulysh ([e-mail]) uploaded a new patch: https://review.whamcloud.com/30368
Subject: LU-9828 ptlrpc: ASSERTION(desc->bd_nob_transferred == 0)
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: d4d5658f3cdad115c06a998e7fa91a2bd89e33dd

[Andriy Skulysh]

The assertion failure can happen only during resend vs reply race. It is better to skip reply and restore the assertion. I'll commit the patch.

[Gerrit Updater]

John L. Hammond ([e-mail]) merged in patch https://review.whamcloud.com/28759/
Subject: LU-9828 ptlrpc: Do not assert when bd_nob_transferred != 0
Project: fs/lustre-release
Branch: b2_10
Current Patch Set:
Commit: 39a275578e5d77d14f5b50b3c2a3fc924081e03c

[Gerrit Updater]

Minh Diep ([e-mail]) uploaded a new patch: https://review.whamcloud.com/28759
Subject: LU-9828 ptlrpc: Do not assert when bd_nob_transferred != 0
Project: fs/lustre-release
Branch: b2_10
Current Patch Set: 1
Commit: 7289ae9b0767ac65323bad97471b15f735154024

[Peter Jones]

Landed for 2.11

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/28491/
Subject: LU-9828 ptlrpc: Do not assert when bd_nob_transferred != 0
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: e6490ea6cf0b793c0b47f17ac5a5fa3a2a136e0d