https://jira.whamcloud.com This file is an XML representation of an issue en-us 9.4.14 940014 05-12-2023 https://jira.whamcloud.com/browse/LU-9548 Lustre <p>While running lfsck in dry-run mode on our servers we set lctl set_param debug=+lfsck in order to collect debug data to see what files will be impacted by a real lfsck run. The debug logs has zero lfsck debug tracing. This doesn't seem to be correct behavior. We would like to know what repairs would be done before a real lfsck run. </p> Any server that runs lfsck in dry-run mode. LU-9548

No debug info from lctl set_param debug=+lfsck shows up with lfsck dry-run

Bug Trivial Open Unresolved nasf James A Simmons Tue, 23 May 2017 16:19:35 +0000 Sat, 16 Jun 2018 14:28:49 +0000 Lustre 2.8.0 Lustre 2.9.0 Lustre 2.10.0 0 5 <p>Fan Yong</p> <p>Could you please advise with this one?</p> <p>Thanks</p> <p>Peter</p> <p>Does a tool exist that can create errors at the OSD level so lfsck can be tested?</p> <p>In theory, if there are inconsistent items detected by LFSCK, then "debug=+lfsck" will show related LFSCK actions, in spite of dryrun mode or not, unless the log is overwritten. To verify that, we can inject some failure stub as sanity-lfsck does. For example, sanity-lfsck test_14 will simulate dangling LOV EA reference case.</p> <p>James,</p> <p>Have you tried the LFSCK with some failure injection to simulate the inconsistent cases? Is there still no logs with dryrun mode for that?</p> <p>Thanks!</p> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=simmonsja" class="user-hover" rel="simmonsja">simmonsja</a>,</p> <p>Any further feedback for this ticket?</p> <p>Sorry I haven't got around to looking into this. Please keep the ticket open for now. I will talk to our admins about what they are looking for exactly.</p> Related LU-9187 Development Rank 1|hzzdgf: Rank (Obsolete) 9223372036854775807 Severity