[Lustre-discuss] Crash and reboot accessing NFS exported Lustre filesystem.

Girish Shilamkar gshilamkar at ddn.com
Wed Aug 14 05:24:27 PDT 2013

Hello Phill,

This is a known issue https://jira.hpdd.intel.com/browse/LU-3717
inode is passed as NULL in ll_encode_fh() which results in this LBUG


On 14-Aug-2013, at 5:45 PM, Phill Harvey-Smith <p.harvey-smith at warwick.ac.uk<mailto:p.harvey-smith at warwick.ac.uk>> wrote:

Hi All,

We have one of our Lustre filesystems that we export via NFS to a
machine that we don't currently have the Lustre client modules built for.

The remote machine can mount the filesystems however as soon as it tries
to access a file on the NFS mounted filesystem, it causes the NFS server
to crash and reboot.

The NFS server is Ubuntu 12.04LTS x86_64, kernel 3.8.0-27-generic, and
Lustre version :

Lustre: Build Version: v2_4_90_0-g3917e62-CHANGED-3.8.0-27-generic

This was checked out from git yesterday, so is the newest 2.4 code.

The filesystems are exported as :

/home ,async,no_subtree_check)

On the target system I did the following :

mount test1:/home /mnt
ls /mnt

At this point it crashes the server, the relevant section of the
crashlog as captured from the console is :

[  472.439409] LustreError:
2131:0:(llite_internal.h:1175:ll_inode2fid()) ASSERTION( inode != ((void
*)0) ) failed:
[  472.449660] LustreError:
2131:0:(llite_internal.h:1175:ll_inode2fid()) LBUG
[  472.478560] Kernel panic - not syncing: LBUG

Does anyone know how I can fix this ?


Lustre-discuss mailing list
Lustre-discuss at lists.lustre.org<mailto:Lustre-discuss at lists.lustre.org>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20130814/84c40883/attachment.htm>

More information about the lustre-discuss mailing list