[OpenAFS-devel] oops on AFS mount
Derek Atkins
warlord@MIT.EDU
13 May 2002 11:51:20 -0400
"John A. Goebel" <jgoebel@SLAC.Stanford.EDU> writes:
> ++ 12/05/02 23:01 -0400 - <Derek Atkins>:
> > Are there any contents in "./Bonnie.8803" or it is really an
> > empty directory? What do you get from:
> >
> > ls -la ./Bonnie.8803
>
> root@temp11 # ls
> ls: 0004563i3i: No such file or directory
> ls: 0004564LjtANNJK: No such file or directory
> ls: 0004565wG5X0IOO0gDN: No such file or directory
> ls: 0004566i: No such file or directory
> ls: 0004567awajNiu6qEhb: No such file or directory
>
> ...then a hang and a slow roll of ls:
>
> 0004568OM9k0d: No such file or directory
This really looks like the refresh bug that was corrected in 1.2.4.
Did this happen with OpenAFS 1.2.3 or the 1.2.4 release candidate?
> > Can you show me a consise example that exhibits the issue?
>
> The way I get a consistant problems is by running bonnie++ with the
> line:
>
> strace -o /tmp/bonnie.slog ./bonnie++ -x 1000 \
> -d /afs/test.slac.stanford.edu/ext 2/temp11 -u jgoebel >> /tmp/bonnie.dat
Ok, let me rephrase: can you show me the actual steps that cause
the problem? Find the portion of the bonnie tests and cut-and-paste
just that test that causes the failure. For example:
mkdir ...
touch ..
rm ...
rmdir .. <- crashes
I'd like to see a concise "this will cause the crash" process.
I don't have access to "bonnie++".
> > You also might want to wait for the 1.2.4 RPMS to be released
> > Really Soon Now. There were some late-2.4.x kernel changes which
> > needed a patch to OpenAFS, and that patch is in 1.2.4.
>
> I tested this on the release canidate 2 of 1.2.4 also, but this was
> before switching over to ext2. I can run the test today on that
> version.
Thanks. That would be useful.
> Thanks.
>
> John
-derek
--
Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
Member, MIT Student Information Processing Board (SIPB)
URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH
warlord@MIT.EDU PGP key available