On 22 October 2015 at 22:53, Andrew Wilkins <[email protected]> wrote:
> On Thu, Oct 22, 2015 at 5:39 PM roger peppe <[email protected]> > wrote: > >> This seems reasonable to me. For myself, the time I'm looking at >> all-machines.log >> is often when the system is broken. I wonder whether a reasonable >> substitute might be a low level tool that can operate directly on the >> mongo >> database to fetch the logs, bypassing the juju system. >> > > +1 > I'd like to see rsyslog gone, and the new logging to be enabled by > default, but I think we do need some tooling to extract logs when things go > awry. > OK, we'll work on a utility which can be run on a Juju controller server to extract all the logs from the database, even when the Juju server isn't running. Note that there is also already a last resort logsink.log file generated on each Juju controller server which records the last 1GB of logs received by that server. If you grab this file from each server and merge them you end up with the all the recent logs.
-- Juju-dev mailing list [email protected] Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/juju-dev
