maas-devel team mailing list archive

[Mark Shuttleworth <mark@xxxxxxxxxx>]

On 18/07/14 16:32, Gavin Panella wrote:
> To stop the "Unable to get RPC connection" errors from appearing, shut
> the region web app down and truncate the eventloops table in
> PostgreSQL. Alternatively, truncate it and wait 60 seconds. Stale
> entries get pruned after 5 minutes so these errors should not persist
> in a longer-running system; it's a bug if they do. Stale entries in
> that table should only appear when Apache crashes (or the machine
> does); that they're appearing routinely might indicate that Apache is
> not shutting down its Python interpreters "nicely".

We can't rely on sharing this sort of anecdotal fix between people in
the know - is there a bug for this situation to be detected and fixed
programmatically?

Mark

Attachment: signature.asc
Description: OpenPGP digital signature