Issue #2786
closed
The content wsgi does not respect logging configuration
Status:
CLOSED - CURRENTRELEASE
Description
The ability to log to the console was added recently (https://pulp.plan.io/issues/2548) for the Pulp server. However, this logging configuration was not extended to the content wsgi application that serves up repositories which continues to log strictly to syslog.
- Related to Story #2548: As an administrator, I can configure how pulp logs added
@ehelms: I'm not sure what to do to resolve this. The logging is configured at the .conf file for httpd right? Like this [0]. I don't think we can configure the logger via Python code like we do with all the other processes.
We can confirm that httpd doesn't use the Pulp logger by having it raise an Exception in the Pulp logger and try to start httpd. That would probably be good.
[0]: https://github.com/pulp/pulp/blob/master/server/etc/httpd/conf.d/pulp_apache_24.conf
bmbouter
I'm not sure I entirely follow your concern or how this is different than the way the Pulp server (also run by httpd as you know) would be logging information. I can say that this fixed my issue but I am likely missing something subtle here.
- Status changed from NEW to POST
- Assignee set to ehelms@redhat.com
- Triaged changed from No to Yes
Rebased and polished: https://github.com/pulp/pulp/pull/3095
Logging within our WSGI apps is independent from httpd's logging.
Our REST API WSGI app takes particular measures to setup its logging. When the content app was added, it got a slightly different approach to configuring logging. That divide is I think what led to "console" logging getting added in one place and not the other.
- Status changed from POST to MODIFIED
- Platform Release set to 2.14.0
- Status changed from MODIFIED to 5
- Status changed from 5 to CLOSED - CURRENTRELEASE
Also available in: Atom
PDF
fixes #2786: Allow content application to use console logging