[LU-5109] Failure on test suite parallel-scale-nfsv3 test_compilebench: IOError: [Errno 22] Invalid argument

• Status: Resolved
• Project: Lustre
• Affects Version/s: Lustre 2.6.0
• Type: Bug
• Priority: Blocker
• Reporter: Maloo
• Assignee: Lai Siyao
• Resolution: Duplicate

Environment

server: lustre-master build #2064

This error is seen in both master client and interop test with 2.5.1 client

• Severity: 3

Description

This issue was created by maloo for sarah <[e-mail]>

This issue relates to the following test suite run: http://maloo.whamcloud.com/test_sets/0b4b43ca-e50d-11e3-a294-52540035b04c.

The sub-test test_compilebench failed with the following error:

compilebench failed: 1

== parallel-scale-nfsv3 test compilebench: compilebench == 07:14:41 (1401113681)
OPTIONS:
cbench_DIR=/usr/bin
cbench_IDIRS=2
cbench_RUNS=2
client-24vm1
client-24vm2.lab.whamcloud.com
./compilebench -D /mnt/lustre/d0.compilebench -i 2         -r 2 --makej
using working directory /mnt/lustre/d0.compilebench, 2 intial dirs 2 runs
Traceback (most recent call last):
  File "./compilebench", line 567, in <module>
    dset = dataset(options.sources, rnd)
  File "./compilebench", line 319, in __init__
    self.unpatched = native_order(self.unpatched, "unpatched")
  File "./compilebench", line 97, in native_order
    run_directory(tmplist, dirname, "native %s" % tag)
  File "./compilebench", line 245, in run_directory
    fp.close()
IOError: [Errno 22] Invalid argument
 parallel-scale-nfsv3 test_compilebench: @@@@@@ FAIL: compilebench failed: 1 

Issue Links

• is related to: LU-3544 Writing to new files under NFS export...

Comments

[Andreas Dilger]

This may release to LU-3544, indicating that there are ORR-existing problem with the NFS code.

Is the parallel-scale-nfs test failing all the time, or was this just a one time failure?

[Lai Siyao]

It should be a different issue: LU-3544 happens on SLES11SP2, because NFS in new kernels packs '/' as file name, but this test is done on RHEL6. It's interesting that nfsv4 can pass, but v3 not. I'll look into the logs and do some test.

[Lai Siyao]

It failed most of the time against master branch, and passed 1 out of 3 rounds.

[Lai Siyao]

I reproduced it locally, and found that the name is "", it failed because of the patch http://review.whamcloud.com/#/c/10198/ for LU-4992 added stricter check on file name. And the patch for LU-3544 can't fully fix this because it still packs name for old servers (not supporting OBD_CONNECT_OPEN_BY_FID), I'll update the patch of LU-3544 to fix this along. This can be marked as duplicate of LU-3544.

[Lai Siyao]

file name is "", llite thinks this is invalid and returns -EINVAL. This will be fixed along in LU-3544.