Bug 2142270
Summary: | Locale change caused by RHEL upgrade results in database index corruption "get() returned more than one Modulemd -- it returned 2!" | |||
---|---|---|---|---|
Product: | Red Hat Enterprise Linux 7 | Reporter: | Anand Jambhulkar <ajambhul> | |
Component: | leapp-repository | Assignee: | Evgeni Golov <egolov> | |
Status: | CLOSED ERRATA | QA Contact: | Lukas Pramuk <lpramuk> | |
Severity: | urgent | Docs Contact: | ||
Priority: | urgent | |||
Version: | 7.9 | CC: | ahumbe, bdm, dalley, egolov, ehelms, ggainey, hyu, iballou, jbhatia, jbreitwe, jentrena, jpasqual, jsenkyri, lpramuk, momran, paji, pdwyer, pmoravec, pstodulk, risantam, sajha, saydas, wclark, wpinheir | |
Target Milestone: | rc | Keywords: | Triaged, Upgrades, WorkAround | |
Target Release: | --- | |||
Hardware: | x86_64 | |||
OS: | Linux | |||
Whiteboard: | ||||
Fixed In Version: | leapp-repository-0.17.0-8.el7_9 | Doc Type: | If docs needed, set a value | |
Doc Text: | Story Points: | --- | ||
Clone Of: | ||||
: | 2161413 2161929 (view as bug list) | Environment: | ||
Last Closed: | 2023-03-14 13:56:57 UTC | Type: | Bug | |
Regression: | --- | Mount Type: | --- | |
Documentation: | --- | CRM: | ||
Verified Versions: | Category: | --- | ||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | ||
Cloudforms Team: | --- | Target Upstream Version: | ||
Embargoed: | ||||
Bug Depends On: | ||||
Bug Blocks: | 2161929, 2167984 |
Description
Anand Jambhulkar
2022-11-12 13:33:29 UTC
We've been investigating, it's just being discussed on Matrix instead of Bugzilla. I will mark it as assigned. We have yet to be able to reproduce this failure, so if support is able to help us get a reliable reproducer that would be very helpful. I tried to setup a reproducer at my end but could not reproduce it. Please let me know how I can help you in setting up the reproducer? If you need any specific conf files/data from customer's Satellite, I can get it for you to debug the issue. Great finding! I can confirm the uniq index "rpm_modulemd_name_stream_version_context_arch_f6598e4e_uniq" violation on our Satellite (https://bugzilla.redhat.com/show_bug.cgi?id=2044631#c52) is timely aligned with / after the OS upgrade from RHEL7. Now the questions come: - how to prevent such index violation? - some customers already got into the broken index state - is the orphan cleanup a valid workaround / remedy every time? - can I close mine https://bugzilla.redhat.com/show_bug.cgi?id=2154289 as duplicate BZ? :) (In reply to Pavel Moravec from comment #24) > Great finding! I can confirm the uniq index > "rpm_modulemd_name_stream_version_context_arch_f6598e4e_uniq" violation on > our Satellite (https://bugzilla.redhat.com/show_bug.cgi?id=2044631#c52) is > timely aligned with / after the OS upgrade from RHEL7. > > Now the questions come: > - how to prevent such index violation? > - some customers already got into the broken index state - is the orphan > cleanup a valid workaround / remedy every time? > - can I close mine https://bugzilla.redhat.com/show_bug.cgi?id=2154289 as > duplicate BZ? :) Needinfo on Daniel to confirm the action plan. (In reply to Pavel Moravec from comment #24) > Great finding! I can confirm the uniq index > "rpm_modulemd_name_stream_version_context_arch_f6598e4e_uniq" violation on > our Satellite (https://bugzilla.redhat.com/show_bug.cgi?id=2044631#c52) is > timely aligned with / after the OS upgrade from RHEL7. > > Now the questions come: > - how to prevent such index violation? Self-answering this question by a proposal: during the Leapp upgrade of Satellite from RHEL7 to RHEL8, enter a new mandatory step to reindex the table (or whole/all DBs..? just to be sure @pmoravec Yes I think the other issue (#2154289) can be closed as a duplicate. I'll defer to Hao on the prevention / workaround since he knows way more on this topic than me :) I assume that it will require some kind of change to the upgrade procedure and that the component should change? *** Bug 2154289 has been marked as a duplicate of this bug. *** Moving this back to NEW and switching the component since it is an upgrade / installer issue rather than a code issue, and changing the title to better reflect what we know about the problem. I agree, this is https://wiki.postgresql.org/wiki/Locale_data_changes and we should run a re-index during the upgrade automatically. Running "runuser -u postgres -- reindexdb -a" should fix all indexes VERIFIED.
@Satellite 6.11.4.1 CDN
leapp-0.15.0-2.el7_9.noarch
leapp-upgrade-el7toel8-0.17.0-8.el7_9.noarch (brew)
by the following manual reproducer:
1) Have a Satellite 6.11.z (mine was 6.11.4.1) on RHEL7
2) Perform LEAPP Upgrade
3) Wait till leapp_resume.service is finished (=inactive)
4) Run DB re-index on RHEL8 Satellite to detect a possible corruption of unique indices
# runuser - postgres -c "reindexdb -a"
reindexdb: reindexing database "candlepin"
reindexdb: reindexing database "foreman"
reindexdb: reindexing database "postgres"
reindexdb: reindexing database "pulpcore"
reindexdb: reindexing database "template1"
>>> in-place upgraded Satellite manifests no unique indices corruption on RHEL8
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (leapp-repository bug fix update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2023:1228 The needinfo request[s] on this closed bug have been removed as they have been unresolved for 120 days |