search

redhat-openstack / ansible-role-tripleo-collect-logs

A role designed to collect the critical logs, configuration and artifacts from a tripleo deployment
Apache License 2.0
7 stars 1 forks source link

We should collect a human-readable journal and not the binary journals #2

Closed trown closed 7 years ago

trown commented 8 years ago

The binary journal is not readable from the web, so we should instead collect jouralctl output to a file, and not collect the binary journals in /var/log/journal.

HarryRybacki-zz commented 8 years ago

@trown do you know if this is still an issue or if someone solved it? If not I'll jump on it soon.

trown commented 8 years ago

@HarryRybacki It does not look like there is anything collecting all journal output, there is collecting from failed services though: https://github.com/redhat-openstack/ansible-role-tripleo-collect-logs/blob/master/tasks/collect.yml#L31-L34

not sure if that is enough or not...

HarryRybacki-zz commented 8 years ago

@trown who would we ask to determine an answer to ^^? I'm happy to add the functionality if it's helpful for debugging builds.

trown commented 8 years ago

@HarryRybacki I wonder how many bytes that adds to the log collection for each job... IMO what we have is fine, as services are still logging to their own files in /var/log. When/if services start to only log to the journal, we would need to revisit this.

apevec commented 7 years ago

There is still useful system-wide logging in the journal and plain copying binary journal is not good enough, see https://github.com/redhat-openstack/ansible-role-tripleo-collect-logs/issues/7

Also looks like tripleo is reconfiguring journal to not log to syslog? /var/log/messages looks rather empty after journal reconfiguration in recent jobs so we really need full journal properly collected.

apevec commented 7 years ago

I propose to close this one as duplicate of issue #7