[LU-3909] Interop 2.4.0<->2.5 failure on test suite sanity test_129: return code 28 received instead of expected 27 Created: 08/Sep/13  Updated: 14/Jul/15  Resolved: 05/Jan/14

Status: Resolved
Project: Lustre
Component/s: None
Affects Version/s: Lustre 2.5.0
Fix Version/s: Lustre 2.6.0, Lustre 2.5.1

Type: Bug Priority: Minor
Reporter: Maloo Assignee: Cliff White (Inactive)
Resolution: Fixed Votes: 0
Labels: patch
Environment:

client: 2.4.0
server: lustre-master build #1652


Issue Links:
Related
is related to LU-3906 Failure on test suite parallel-scale ... Resolved
is related to LU-3908 Interop 2.4.0<->2.5 failure on test s... Closed
Severity: 3
Rank (Obsolete): 10305

 Description   

This issue was created by maloo for sarah <sarah@whamcloud.com>

This issue relates to the following test suite run: http://maloo.whamcloud.com/test_sets/9fb64ab0-15be-11e3-a83f-52540035b04c.

The sub-test test_129 failed with the following error:

return code 28 received instead of expected 27

Cannot find useful log



 Comments   
Comment by Jodi Levi (Inactive) [ 09/Sep/13 ]

Both tests LU-3906 and this one fail because they run out of space.

Comment by Oleg Drokin [ 09/Sep/13 ]

This test failed because it run out of space before reaching limits.

Probably a test env issue?

Comment by Sarah Liu [ 11/Sep/13 ]

Also hit this error between 2.4.0 server and 2.5 client:
https://maloo.whamcloud.com/test_sets/f5eae562-19c4-11e3-bb73-52540035b04c

Comment by Vladimir Saveliev [ 02/Nov/13 ]

The problem seems to be caused by changes of max dir size limit logic. The below patch is to fix that. Please take a look.
http://review.whamcloud.com/8143

For correct work the following patch from https://jira.hpdd.intel.com/browse/LU-2479 is needed:
http://review.whamcloud.com/8137

Comment by Cliff White (Inactive) [ 02/Dec/13 ]

Patch needs to be rebased

Comment by Vladimir Saveliev [ 02/Dec/13 ]

new version is uploaded, please, take a look
http://review.whamcloud.com/8143

Comment by Vladimir Saveliev [ 04/Dec/13 ]

3rd version is uploaded.
'lctl get_param' is used to get mds blocksize as Andreas suggested.

Please, take a look
http://review.whamcloud.com/8143

Comment by Jian Yu [ 05/Jan/14 ]

Lustre client build: http://build.whamcloud.com/job/lustre-b2_4/70/ (2.4.2)
Lustre server build: http://build.whamcloud.com/job/lustre-b2_5/5/

The same failure occurred:
https://maloo.whamcloud.com/test_sets/5584c4fa-755c-11e3-95ae-52540035b04c

Comment by Jian Yu [ 05/Jan/14 ]

Lustre client build: http://build.whamcloud.com/job/lustre-b2_5/5/
Lustre server build: http://build.whamcloud.com/job/lustre-b2_4/70/ (2.4.2)

The same failure occurred:
https://maloo.whamcloud.com/test_sets/a8e2dbf8-74f5-11e3-95ae-52540035b04c

Here is the back-ported patch for Lustre b2_5 branch: http://review.whamcloud.com/#/c/8727/

Comment by Peter Jones [ 05/Jan/14 ]

Landed for 2.6. Should land to b2_5 shortly

Comment by Andreas Dilger [ 17/Jan/14 ]

I noticed that on my system the max_dir_size tunable is /sys/fs/ldiskfs/

{dev}/max_dir_size instead of /proc/fs/ldiskfs/{dev}

/max_dir_size. That will cause this test to fail on newer kernels I think?

Comment by Jian Yu [ 17/Apr/14 ]

That will cause this test to fail on newer kernels I think?

In set_dir_limits():

        local LDPROC=/proc/fs/ldiskfs
        #......

                do_facet $facet "test -e $LDPROC/$canondev/max_dir_size" ||
                                                LDPROC=/sys/fs/ldiskfs

If "/proc/fs/ldiskfs" does not exist, "/sys/fs/ldiskfs" will be used.

Generated at Sat Feb 10 01:37:58 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.