[LU-385] __ldlm_handle2lock()) ASSERTION(handle) Created: 01/Jun/11  Updated: 13/Jun/11  Resolved: 13/Jun/11

Status: Resolved
Project: Lustre
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Bug Priority: Major
Reporter: Christopher Morrone Assignee: Lai Siyao
Resolution: Fixed Votes: 0
Labels: None
Environment:

CHAOS4.4 (RHEL5), lustre 1.8.5.0-3chaos, https://github.com/chaos/lustre/tree/1.8.5.0-3chaos

Severity: 2
Rank (Obsolete): 10234

 Description   

An MDS of ours hit an assertion in __ldlm_handle2lock(), ASSERTION(handle). Below are the lines that precede the assertion, and the backtrace that appeared on the console (copied by hand). No further information available at this time.

LustreError: 8537:0:(ldlm_lock.c:430:__ldlm_handle2lock()) ASSERTION(handle)
failed
LustreError: 8550:0:(mds_open.c:1177:mds_open()) parent 824180809/556408324
lookup/take lock error -13
LustreError: 8541:0:(mds_open.c:1177:mds_open()) parent 824180841/556408354
lookup/take lock error -13
LustreError: 8541:0:(mds_open.c:1177:mds_open()) Skipped 5 previous similar
messages
LustreError: 8537:0:(ldlm_lock.c:430:__ldlm_handle2lock()) LBUG

Task 8537 was apparently ll_mdt_12.

Backtrace is:

__ldlm_handle2lock
ldiskfs_permission
permission
ldlm_lock_decref
mds_verify_child
default_wake_function
__wake_up_common
mds_get_parent_child_locked
ksocknal_queue_tx_locked
mds_open
lustre_msg_bug
mntput_no_expire
mds_reint_rec
mds_update_unpack
mds_reint
fixup_handle_for_resent_req
mds_intent_policy
ldlm_lock_enqueue
ldlm_resource_putref
ldlm_lock_create
lustre_hash_add
ldlm_server_blocking_ast
ldlm_handle_enqueue
mds_handle
smp_send_reschedule
lustre_msg_get_conn_cnt
try_to_wake_up
lustre_msg_get_opc
ptlrpc_server_handle_request
ptlrpc_wait_event
ptlrpc_main

 Comments   
Comment by Peter Jones [ 01/Jun/11 ]

Lai

Could you please look into this one?

Thanks

Peter

Comment by Liang Zhen (Inactive) [ 01/Jun/11 ]

I think it's just a instance of https://bugzilla.lustre.org/show_bug.cgi?id=24360 and the fix (by Niu) is already in upstream
Liang

Comment by Lai Siyao [ 01/Jun/11 ]

Hi Chris, could you verify the fix is in your code?

Liang, thanks, the backtrace and messages look the same.

Comment by Christopher Morrone [ 02/Jun/11 ]

No we don't have that. We're are using 1.8.5, and that landed in 1.8.6.

I'll add that patch to our branch.

Comment by Peter Jones [ 13/Jun/11 ]

Let's close this ticket for now and reopen if the issue reoccurs with the patch applied

Generated at Sat Feb 10 01:06:30 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.