← Back to team overview

fuel-dev team mailing list archive

Re: Astute log improvement

 

Dmitry, hi!

About this case https://bugs.launchpad.net/fuel/+bug/1356954.
Short description: we marked node as error because lost access to it during
mcollective, but deploy other nodes.
Current mechanic: if this failed node (reason not important) have no
critical significance, we marked it as error and continue. If we deploy 20
nodes, we can easily survive fail (for any reason: mcollective, network,
puppet) of 2 computes nodes for example. I think this is a correct
behavior, but i can mistake.

If we add information why we marked this node as error (mcollective
connection problem) in brief, we can save a lot of time.


On Thu, Aug 28, 2014 at 2:43 PM, Dmitry Mescheryakov <
dmescheryakov@xxxxxxxxxxxx> wrote:

> Vladimir,
>
> Good idea! I can suggest another bit to improve: when astute fails to
> connect to mcollective on a node, just a number of debug/info messages
> appear in the log. At the same time that ruins the deployment and
> clearly should be logged as error. An example could be seen in logs
> provided in the following bug
> https://bugs.launchpad.net/fuel/+bug/1356954
>
> Thanks,
>
> Dmitry
>
>
> 2014-08-28 12:05 GMT+04:00 Vladimir Sharshov <vsharshov@xxxxxxxxxxxx>:
> > Hi!
> >
> > I believe many of us try to investigate Astute logs to find problem with
> > cluster.
> > And in mostly case it was very difficult. Please find time and look to
> this
> > document: https://etherpad.openstack.org/p/fuel-astute-log-improvement
> >
> > Your feedback will help us to improve this moment.
> >
> > Feel free to contact with me.
> >
> > IRC nickname in #fuel-dev: warpc
> > Skype: warpcv
> >
> > --
> > Mailing list: https://launchpad.net/~fuel-dev
> > Post to     : fuel-dev@xxxxxxxxxxxxxxxxxxx
> > Unsubscribe : https://launchpad.net/~fuel-dev
> > More help   : https://help.launchpad.net/ListHelp
> >
>

References