Jira (PDB-4626) Investigate resource_events partitioning migration failure

18 views
Skip to first unread message

Austin Blatt (JIRA)

unread,
Jan 17, 2020, 12:46:05 PM1/17/20
to puppe...@googlegroups.com
Austin Blatt updated an issue
 
PuppetDB / Bug PDB-4626
Investigate resource_events partitioning migration failure
Change By: Austin Blatt
Summary: Investigate resource_events partitioning migration failure
Release Notes: Bug Fix
Add Comment Add Comment
 
This message was sent by Atlassian JIRA (v7.7.1#77002-sha1:e75ca93)
Atlassian logo

Austin Blatt (JIRA)

unread,
Jan 17, 2020, 12:47:03 PM1/17/20
to puppe...@googlegroups.com

Austin Blatt (JIRA)

unread,
Jan 17, 2020, 3:18:04 PM1/17/20
to puppe...@googlegroups.com
Austin Blatt commented on Bug PDB-4626
 
Re: Investigate resource_events partitioning migration failure

We are able to reproduce this failure by changing system time to a positive UTC offset

TZ=Europe/Paris lein test :only puppetlabs.puppetdb.scf.migrate-test

Austin Blatt (JIRA)

unread,
Jan 17, 2020, 3:29:04 PM1/17/20
to puppe...@googlegroups.com

Austin Blatt (JIRA)

unread,
Jan 17, 2020, 3:29:04 PM1/17/20
to puppe...@googlegroups.com

Austin Blatt (JIRA)

unread,
Jan 17, 2020, 3:31:03 PM1/17/20
to puppe...@googlegroups.com
Austin Blatt updated an issue
Change By: Austin Blatt
Release Notes Summary: A database migration would fail because it didn't properly handle timezones with positive UTC offsets

Austin Blatt (JIRA)

unread,
Jan 17, 2020, 3:31:03 PM1/17/20
to puppe...@googlegroups.com
Austin Blatt updated an issue
Change By: Austin Blatt
Method Found: Needs Assessment Customer Feedback

Michael Prokop (JIRA)

unread,
Jan 21, 2020, 11:18:04 AM1/21/20
to puppe...@googlegroups.com
Michael Prokop commented on Bug PDB-4626
 
Re: Investigate resource_events partitioning migration failure

FTR, a customer of mine is also affected by this:

mika@puppet ~ % sudo zless /var/log/puppetlabs/puppetdb/puppetdb-2020-01-15.0.log.gz[...]2020-01-15T06:55:45.426+01:00 ERROR [p.p.s.migrate] Caught SQLException during migration
java.sql.BatchUpdateException: Batch entry 0 INSERT INTO resource_events_20200103Z ( event_hash, report_id, certname_id, status, timestamp, resource_type, resource_title, property, new_value, old_value, message, file, line, containment_path, containing_class, corrective_change, name ) VALUES ( '\x501aa8aa6d7d2a66d7b93ca07a8fbe63e36c8d77', 429028, 14, 'skipped', '2020-01-03 06:58:06.805+01', 'Exec', '/usr/bin/gitlab-ctl reconfigure', NULL, 'null', 'null', NULL, '/etc/puppetlabs/code/environments/production/site/gitlab/manifests/server.pp', 13, '{Stage[main],Gitlab::Server,"Exec[/usr/bin/gitlab-ctl reconfigure]"}', 'Gitlab::Server', NULL, NULL ),( '\x403832e909d515158843a61847ea20ae92468e1a', 429028, 14, 'skipped', '2020-01-03 06:58:06.801+01', 'File', '/etc/gitlab/gitlab.rb', NULL, 'null', 'null', NULL, '/etc/puppetlabs/code/environments/production/site/gitlab/manifests/server.pp', 7, '{Stage[main],Gitlab::Server,File[/etc/gitlab/gitlab.rb]}', 'Gitlab::Server', NULL, NULL ) was aborted: ERROR: relation "resource_events_20200103z" does not exist
{{ Position: 13 Call getNextException to see other errors in the batch.}}

Downgrading puppetdb from 6.8.0-1stretch to puppetdb 6.7.3-1stretch fixed this for us in the meanwhile.

Environment:

  • Debian stretch
  • Postgres 9.6

Michael Prokop (JIRA)

unread,
Feb 6, 2020, 9:36:03 AM2/6/20
to puppe...@googlegroups.com

FTR, I've tested the upgrade to 6.8.1-1stretch today, but this still fails for us with the same error message, we had to revert/downgrade to 6.7.3-1stretch again. Is there anything special that needs to be taken care of to get this fixed?

Reply all
Reply to author
Forward
0 new messages