On Fri, Mar 22, 2013 at 10:12:46AM +0100, Duncan Mac-Vicar P. wrote: > > I gave it a second round to my patch that improves the channel > status. Right now part of the status is calculated with queries to > the job history, but the metadata generation leaves no traces on the > queue. I could enhance the schema to insert a flag when an exception > is thrown, but this is overkill, as I don't need this data to > survive reboots. If the latest status is lost is harmless.
Not sure what the data would be about. You have the channel repodata generation status, right? I believe we compare the modified timestamp of the channel (or channel package mapping?) to the mtime of the file on disk. Why isn't that enough? > In other "stacks" I would use something like redis for this. A > key/value store that can be configured in-memory or with very basic > persistence, and accessible from all the components (webapp, > taskomatic, python, etc). > > Do we have something like this already? Any other way of achieving the same? > Would adding something like redis to the infrastructure would be desired? I don't see benefit for that. We have simple key/value caching store in /var/cache/rhn/ -- the key is the filename, the value is its content. For persistent data we have the database. -- Jan Pazdziora Principal Software Engineer, Satellite Engineering, Red Hat _______________________________________________ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
