[pnfs] Connectathon lock test #12 fails with NFSERR_BADSESSION

William A. (Andy) Adamson andros at citi.umich.edu
Tue Sep 25 15:04:55 EDT 2007


I'm testing the 2.6.18.3 pnfs-git kernel with the following additional
patches from benny:

0001-pnfs-client-fix-error-path-in-filelayout_alloc_layou.patch
0002-pnfs-do-proc_commit_setup-unconditionally.patch
0003-pnfs-increase-decode_pnfs_layoutget_maxsz.patch
0004-pnfsd-add-debug-printouts-in-return_layout-path.patch
0005-pnfsd-fix-bug-in-return_layout-for-RETURN_FILE.patch
0006-pnfsd-refactor-return_layout.patch
0007-pnfsd-simulate-layoutreturn-on-nomatching_layouts-e.patch

I don't think the above patches are causing this error.

Lock test #12 tests blocking locks "Signalled process should release locks"

Wireshark shows (all with SEQUENCE) an OPEN, a PUTFH:LOCK (new lock) a
PUTFH:GETATTR and a PUTFH:LOCKU
The PUTFH:GETATTR and the PUTFH:LOCKU have the same op sequence

bd 31 f9 46  sessionid
02 00 00 00
bd 31 f9 46
02 00 00 00
00 00 46 21  sequence id
00 00 00 00  slot ID
00 00 00 10  slot ID

and the PUTFH:LOCKU fails with NFSERR_BADSESSION.

Any body else see this?

-->Andy
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://linux-nfs.org/pipermail/pnfs/attachments/20070925/f2a0ab29/attachment.html 


More information about the pNFS mailing list