Re: [Kolla] FAILED - RETRYING: wait for MariaDB to be available via HAProxy (10 retries left).

Clark Boylan cboylan at
Tue Nov 26 16:56:12 UTC 2019

On Tue, Nov 26, 2019, at 7:04 AM, Mark Goddard wrote:
> On Tue, 26 Nov 2019 at 14:12, Bhupathi, Ramakrishna
> <C-Ramakrishna.Bhupathi at> wrote:
> >
> > Mark,
> > Thanks for the tip.
> >
> > Now I get a failure (as below)  during deploy (I have all-in-one) config. I am on stein.
> > It appears to me that this has been presumably  fixed .  But I still running into this error.
> >
> > I do not see anything explicit in the HAProxy/MariaDB logs.
> Are both the haproxy and mariadb containers up and not restarting?
> Does your api_interface have the VIP activated? Is haproxy listening
> on (check ss)?

We ran into mariadb startup slowness recently in Zuul CI jobs. Basically they generate timezone data tables on first startup which can be very slow depending on your disk IO. Not sure if this is the issue, but the workaround is to set an env var to disable the timezone data table generation (if you don't use those tables).

More information about the openstack-discuss mailing list