Riak CS race condition at start-up (was: Riak-CS issues when Riak endpoint fails-over to new server)

Toby Corkindale toby at dryft.net
Sun Jan 22 22:28:23 EST 2017


Thanks Luke.

We're using the official Basho Ubuntu packages for Ubuntu 14.04 LTS.
These don't seem to perform that step; they just start Riak CS
simultaneously with Riak KV.

I'm surprised if we're the only people who hit this. Don't people hit these
issues in your other production deployments, or are they running on another
platforms that have more finessed start-up scripts?

Toby


On Sat, 21 Jan 2017 at 03:19 Luke Bakken <lbakken at basho.com> wrote:

> Hi Toby,
>
> The process you use to run "riak-cs start" could use the "riak-admin
> wait-for-service riak_kv" command to ensure Riak is ready first:
>
>
> http://docs.basho.com/riak/kv/2.2.0/using/admin/riak-admin/#wait-for-service
>
> --
> Luke Bakken
> Engineer
> lbakken at basho.com
>
>
> On Thu, Jan 19, 2017 at 5:38 PM, Toby Corkindale <toby at dryft.net> wrote:
> > Hi guys,
> > I've switched our configuration around, so that Riak CS now talks to
> > 127.0.0.1:8087 instead of the haproxy version.
> >
> > We have immediately re-encountered the problems that caused us to move to
> > haproxy.
> > On start-up, riak takes slightly longer than riak-cs to get ready, and so
> > riak-cs logs the following then exits.
> > Restarting riak-cs again (so now 15 seconds after Riak started) results
> in a
> > successful start-up, but obviously this is really annoying for our ops
> guys
> > to have to remember to do after restarting riak or rebooting a machine.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.basho.com/pipermail/riak-users_lists.basho.com/attachments/20170123/1c2b81c9/attachment-0002.html>


More information about the riak-users mailing list