[LU-5578] LBUG : (mdt_lib.c:543:mdt_dump_lmm()) ASSERTION( stripe_count <= (__s16)160 ) failed Created: 03/Sep/14  Updated: 03/Sep/14  Resolved: 03/Sep/14

Status: Closed
Project: Lustre
Component/s: None
Affects Version/s: Lustre 2.1.6
Fix Version/s: None

Type: Bug Priority: Critical
Reporter: Bruno Travouillon (Inactive) Assignee: Jian Yu
Resolution: Duplicate Votes: 0
Labels: None
Environment:

Routed filesystem: 1 MDT + 55 OSTs
Servers w/ Lustre 2.1.6 + Bull patches
Clients w/ Lustre 2.4.3 + Bull patches
Routers w/ Lustre 2.4.3 + Bull patches


Issue Links:
Duplicate
duplicates LU-2811 LBUG: stripe_count > LOV_MAX_STRIPE_C... Resolved
Severity: 3
Rank (Obsolete): 15561

 Description   

A user has been able to trigger a LBUG on the MDT of a routed filesystem with the following commands:

client2.4$ lfs setstripe -c 200 test-striping

The client is running Lustre 2.4.
The MDS is running Lustre 2.1.

Console traces:

kern emerg kernel [2283281.373220] LustreError: 4412:0:(mdt_lib.c:543:mdt_dump_lmm()) ASSERTION( stripe_count <= (__s16)160 ) failed:
kern emerg kernel [2283281.383897] LustreError: 4412:0:(mdt_lib.c:543:mdt_dump_lmm()) LBUG

[2283281.390506] Pid: 4412, comm: mdt_353
[2283281.395807]
[2283281.395808] Call Trace:
[2283281.403306] [<ffffffffa051a7f5>] libcfs_debug_dumpstack+0x55/0x80 [libcfs]
[2283281.411916] [<ffffffffa051ae07>] lbug_with_loc+0x47/0xb0 [libcfs]
[2283281.419816] [<ffffffffa0dabbd2>] mdt_dump_lmm+0x272/0x280 [mdt]
[2283281.427509] [<ffffffffa0da55f2>] mdt_getattr_internal+0x672/0xe90 [mdt]
[2283281.436180] [<ffffffffa0789c90>] ? lustre_swab_mdt_body+0x0/0x150 [ptlrpc]
[2283281.444886] [<ffffffffa0da6035>] mdt_getattr+0x225/0x920 [mdt]
[2283281.452621] [<ffffffffa0d9c762>] mdt_handle_common+0x932/0x1750 [mdt]
[2283281.460733] [<ffffffffa0d9d655>] mdt_regular_handle+0x15/0x20 [mdt]
[2283281.468705] [<ffffffffa07974f6>] ptlrpc_main+0xd16/0x1a80 [ptlrpc]
[2283281.476675] [<ffffffff810017cc>] ? __switch_to+0x1ac/0x320
[2283281.484190] [<ffffffffa07967e0>] ? ptlrpc_main+0x0/0x1a80 [ptlrpc]
[2283281.492041] [<ffffffff8100412a>] child_rip+0xa/0x20
[2283281.498695] [<ffffffffa07967e0>] ? ptlrpc_main+0x0/0x1a80 [ptlrpc]
[2283281.506842] [<ffffffffa07967e0>] ? ptlrpc_main+0x0/0x1a80 [ptlrpc]
[2283281.514923] [<ffffffff81004120>] ? child_rip+0x0/0x20

I can't say why the user was trying to stripe over 200 OSTs on a 55 OSTs FS, but the fact is that it led the MDS to panic.



 Comments   
Comment by Bruno Travouillon (Inactive) [ 03/Sep/14 ]

Duplicate of LU-2811 ?

Comment by Peter Jones [ 03/Sep/14 ]

Yu, Jian

Could you please advise as to whether this issue is indeed a duplicate of LU-2811?

Thanks

Peter

Comment by Bruno Faccini (Inactive) [ 03/Sep/14 ]

Yes it is very likely to be a dup !

Comment by Jian Yu [ 03/Sep/14 ]

Thanks Bruno. Yes, Peter, it's a duplicate. The patch for Lustre b2_1 branch is in http://review.whamcloud.com/9884.

Comment by Bruno Travouillon (Inactive) [ 03/Sep/14 ]

Thanks guys!

Generated at Sat Feb 10 01:52:42 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.