[LU-13578] sanity test_39r: atime on client != ost

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.14.0, Lustre 2.16.0, Lustre 2.15.0
• Affects Version/s: Lustre 2.14.0
• Severity: 3
• Assignee: John Hammond
• Reporter: Maloo

Description

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

This issue relates to the following test suite run: https://testing.whamcloud.com/test_sets/ec17386d-cefc-4749-9b51-1570a2c15729

test_39r failed with the following error:

atime on client 1589573245 != ost 0x5ebef67c

Very few additional information available:

OST atime:  atime: 0x5ebef67c:00000000 -- Fri May 15 20:07:24 2020
 sanity test_39r: @@@@@@ FAIL: atime on client 1589573245 != ost 0x5ebef67c 

VVVVVVV DO NOT REMOVE LINES BELOW, Added by Maloo for auto-association VVVVVVV
sanity test_39r - atime on client 1589573245 != ost 0x5ebef67c

Issue Links

• is duplicated by: LU-16421 sanity-flr test_61a: atime: old '1670819992' != new '1670819987'
• is duplicated by: LU-14091 sanity test_39r: 'atime on client 1603945630 != ost 0x5f9a449d'
• is related to: LU-17265 sanity test_39r: atime on client 1699192823 != ost 0x65479ff6

Comments

[Peter Jones]

Landed for 2.14

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/40973/
Subject: LU-13578 test: use a single read() in sanity test_39r
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 842f8158a533b21d8c6e401881db5ed7013b7890

[Andreas Dilger]

+1 on master https://testing.whamcloud.com/test_sets/7f189347-052c-4c16-897f-cf26a3c17114

[Gerrit Updater]

John L. Hammond ([e-mail]) uploaded a new patch: https://review.whamcloud.com/40973
Subject: LU-13578 test: use a single read() in sanity test_39r
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: ed1f8d0a2d1a2c28ab72434c6ca4c9d031d9f729

[John Hammond]

Copied over from LU-14091:

Note that 0x5f9a449d == 1603945629 and that since the test uses (( ... )) this is not due to a difference of base.

This may be from a final read() done by dd which returns no bytes and does not generate a BRW RPC to the OST. Even though it returns 0 bytes, it requested a non-zero number of bytes and is therefore required to update the file access time.

[Andreas Dilger]

+1 on master https://testing.whamcloud.com/test_sets/c9dca577-ca19-4ebb-8275-17d37fcdf09a

[Bruno Faccini (Inactive)]

+1 on recent master at https://testing.whamcloud.com/test_sets/999472ce-8888-4fcd-bc47-7210a11127f6

[Andreas Dilger]

Saw this again on master:

atime on client 1595930572 != ost 0x5f1ff7cb

Again the OST timestamp was out by 1s. It may be that there is some kind of race, or the 1s granularity of the clocks makes the test fail some small fraction of time?

This is failing about once every 3 days (9x in the past 4 weeks). Not critical, but it might be nice to understand it better:
https://testing.whamcloud.com/test_sets/40740ba8-bf93-4123-be3c-06965acb4214
https://testing.whamcloud.com/test_sets/c25091a4-bb82-476d-abba-77839eaa665a
https://testing.whamcloud.com/test_sets/bb7c9f0f-88bb-4743-8d09-ae869bcc8037
https://testing.whamcloud.com/test_sets/1bd52851-161f-4ced-818f-54e9f073298d
https://testing.whamcloud.com/test_sets/62c3553a-1bad-4c3c-847f-511531523230
https://testing.whamcloud.com/test_sets/a58fda9c-348e-48c6-9d9a-b4039e8cd333
https://testing.whamcloud.com/test_sets/aa7b32cb-737b-483b-90ee-563abb9a757b
https://testing.whamcloud.com/test_sets/259a006a-8e65-4ada-9a61-d27a067419b3
https://testing.whamcloud.com/test_sets/682470a0-36c2-4364-8f73-661e7f8e9f1d