Ceph » RADOS

There are many instances of this flag getting set in the test run intentionally, so it makes sense to whitelist.

/a/yuriw-2024-04-09_14:58:21-upgrade-wip-yuri4-testing-2024-04-08-1432-distro-default-smithi/7648856$ cat teuthology.log | grep "osd set" | grep "nodeep-scrub" 
2024-04-10T02:27:05.258 DEBUG:teuthology.orchestra.run.smithi062:> sudo /home/ubuntu/cephtest/cephadm --image quay.ceph.io/ceph-ci/ceph:reef shell --fsid c02b312e-f6dc-11ee-b64a-cb9ed24678a4 -- ceph osd set nodeep-scrub
2024-04-10T02:27:18.827 INFO:journalctl@ceph.mon.a.smithi062.stdout:Apr 10 02:27:18 smithi062 bash[80906]: audit 2024-04-10T02:27:18.270108+0000 mon.a (mon.0) 1416 : audit 1 from='client.? 172.21.15.62:0/1866068261' entity='client.admin' cmd=[{"prefix": "osd set", "key": "nodeep-scrub"}]: dispatch
2024-04-10T02:27:18.827 INFO:journalctl@ceph.mon.c.smithi062.stdout:Apr 10 02:27:18 smithi062 bash[83329]: audit 2024-04-10T02:27:18.270108+0000 mon.a (mon.0) 1416 : audit 1 from='client.? 172.21.15.62:0/1866068261' entity='client.admin' cmd=[{"prefix": "osd set", "key": "nodeep-scrub"}]: dispatch
2024-04-10T02:27:19.138 INFO:journalctl@ceph.mon.b.smithi129.stdout:Apr 10 02:27:18 smithi129 bash[67507]: audit 2024-04-10T02:27:18.270108+0000 mon.a (mon.0) 1416 : audit 1 from='client.? 172.21.15.62:0/1866068261' entity='client.admin' cmd=[{"prefix": "osd set", "key": "nodeep-scrub"}]: dispatch

/a/yuriw-2024-04-11_17:03:54-rados-wip-yuri6-testing-2024-04-02-1310-distro-default-smithi/7652474

  msgr-failures/few rados thrashers/mapgap thrashosds-health workloads/radosbench}
duration: 4465.95135641098
failure_reason: '"2024-04-12T01:10:00.000098+0000 mon.a (mon.0) 1265 : cluster [WRN]
  Health detail: HEALTH_WARN noscrub flag(s) set" in cluster log'
flavor: default
owner: scheduled_yuriw@teuthology
success: false

Unfortunately, noscrub and nodeep-scrub are not the only warnings we would need to mask for the thrashosds-health test as there appear to be many others. Looking into this further.

@Brad Hubbard perhaps we can address just the warnings reported in this Tracker, and address additional warnings in a Part 2.

Sure Laura

Trying to reproduce this on centos9 since nothing on centos8 seems to work any more.

scrub note: tracked as a part of the meta-ticket.

/a/yuriw-2024-08-28_23:20:46-upgrade-wip-yuri4-testing-2024-08-28-1359-distro-default-smithi/7879171

scrub note: in QA.

scrub note: in QA.

scrub note: still in QA.

This ticket has the fix like https://tracker.ceph.com/issues/66602.
Anyway, pinging Yuri.

scrub note: the PR went into QA!

In QA; awaiting approval

https://github.com/ceph/ceph/pull/59568 merged