Red Hat Bugzilla – Bug 517162
cthon test5 failing on nfsv4 with rhel6 client vs. rhel4 server
Last modified: 2011-02-16 10:43:13 EST
Description of problem: rhts test /kernel/filesystems/nfs/connectathon/rhel4-nfs/nfsvers=4/base fails against rhel 6 alpha 1 kernel. Apperantly it gets into a situation where it can't write one of the required files. Version-Release number of selected component (if applicable): RHEL6.0-20090709.0 kernel 2.6.29.4-1.el6 How reproducible: always Steps to Reproduce: 1.queue up /kernel/filesystems/nfs/connectathon/ in rhts 2.observe that /kernel/filesystems/nfs/connectathon/rhel4-nfs/nfsvers=4/base fails 3. Actual results: Expected results: all tests pass Additional info: For logs, packet capture etc... http://rhts.redhat.com/cgi-bin/rhts/test_log.cgi?id=9174502 test log: Running as user "root" and group "root". This could be dangerous. Capturing on eth0 ===== Starting 'nfsvers=4' test 'base' ===== ----- Server load rhel4-nfs up 10 days, 9:10, load average: 1.90 2.42 1.36 ----- ----- start: Tue Jul 21 11:07:30 EDT 2009 ----- ./server -b -F nfs4 -p /export/home rhel4-nfs Start tests on path /mnt/rhel4-nfs/dell-pe800-01.test [y/n]? sh ./runtests -b -t /mnt/rhel4-nfs/dell-pe800-01.test Starting BASIC tests: test directory /mnt/rhel4-nfs/dell-pe800-01.test (arg: -t) ./test1: File and directory creation test created 155 files 62 directories 5 levels deep in 0.88 seconds ./test1 ok. ./test2: File and directory removal test removed 155 files 62 directories 5 levels deep in 0.46 seconds ./test2 ok. ./test3: lookups across mount point 500 getcwd and stat calls in 0.0 seconds ./test3 ok. ./test4: setattr, getattr, and lookup 1000 chmods and stats on 10 files in 2.5 seconds ./test4 ok. ./test5: read and write ./test5: (/mnt/rhel4-nfs/dell-pe800-01.test) can't create 'bigfile' : Input/output error basic tests failed Tests failed, leaving /mnt/rhel4-nfs mounted
This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux major release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Major release. This request is not yet committed for inclusion.
*** Bug 567333 has been marked as a duplicate of this bug. ***
On the open(), the RHEL6 client does a SETATTR call to take care of the O_TRUNC open flag. When it does this however, it sends a zero stateid, even though it has the file open and has gotten a valid stateid back. I suspect that RHEL4 handling this case incorrectly, but I'll need to verify it with the spec.
Created attachment 400268 [details] patch -- 2 patches to fix zero-stateid handling in nfsd setattr These two (really old) patches seem to fix the problem for me. They change the code to handle the zero-stateid differently. I need to look over them a bit more carefully but they seem sane.
It's also not clear to me whether rhel6 is doing the right thing. The file has been opened, so it seems like it should be using the stateid we got from that.
It's tough to tell from the spec whether RHEL6 is doing the right thing. I think that it's technically *not*, but most servers apparently don't really care. Fixing this in RHEL6 will be more of a VFS patch though. We'd probably have to fix may_open to take an "nd" arg, and then have it check to see whether there's an already instantiated filp in the open intent stuff, and then have it pass that to do_truncate. It's probably the right thing to do, but really isn't high priority.
For giggles, I went ahead and sent a patch upstream to fix this behavior in the client too. It turns out to be a pretty small patch there, but will probably be a bit larger for RHEL6. http://marc.info/?l=linux-kernel&m=126868917104621&w=2 ...if this ends up flying upstream, I'll clone this BZ for RHEL6 too and we'll fix it there as well.
RHEL6 bug 573995.
*** Bug 486768 has been marked as a duplicate of this bug. ***
This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux maintenance release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Update release for currently deployed products. This request is not yet committed for inclusion in an Update release.
Committed in 89.26.EL . RPMS are available at http://people.redhat.com/vgoyal/rhel4/
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2011-0263.html