Red Hat Bugzilla – Bug 127078
nfs o_direct hangs unexpectedly in read
Last modified: 2013-03-06 00:57:07 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4.2)
Description of problem:
While simpley 'cat' of files mounted with the 'noac' option, and nfs
loaded with the module parameter, nfs_uncache_io=1, some file reads
unexpected hang in read.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. mount -o noac dhcp64-180:/tmp /tmp/pro/mnt/tmp/
2. cd /tmp/pro/mnt/tmp/
3. cat war-e.43-e.44
Actual Results: hang in read
Expected Results: no hand
I'm attaching the file, 'war-e.43-e.44' that the read failed on in
case there is sumething special about it, i've also got alt-sysrq-t at
the time of the hang. This also does not appear to be a regression,
although initial reproduced on the U5 beta candidate e.46, i've also
reproduced thsi issue on e.40, and e.41.
Created attachment 101566 [details]
here is the file that we failed to read, and /var/log/messages from the client box w/alt-syrq-t output, the hanging process is the 'cat'
Created attachment 101598 [details]
A SysRq of the hang I'm seeing
rhel2.1 is only accepting security/data corruption fixes at this point. closing.