Dead connections with load balancer

Vitaly E 13vitamins at gmail.com
Fri Apr 10 10:21:26 EDT 2015


Yes, it looks like the same issue. Thanks Csaba.

On Fri, Apr 10, 2015 at 5:11 PM, Csaba Hoch <csaba.hoch at erlang-solutions.com
> wrote:

> Hi Vitaliy,
>
> This ticket is apparently about the same issue:
> https://github.com/basho/riak_api/issues/88
>
> Regards,
> Csaba
>
>
> On Wed, Apr 8, 2015 at 8:50 PM, Vitaliy E <13vitamins at gmail.com> wrote:
> >
> > Dear group,
> >
> > Our application connects to a Riak cluster via a load balancer (Radware
> Alteon), and we encounter problems whenever a Riak connection stays
> inactive for a while. What we think happens is that the load balancer
> removes its session that corresponds to the connection without closing the
> connection itself. Later on it tries to re-establish the connection with
> the same source port, but the Riak server seems to ignore the request,
> probably because it didn't close the old connection on its side.
> >
> > 1. Does it make sense?
> > 2. Does anyone have experience using Riak with a load balancer?
> > 3. Are there configuration parameters to enable closing idle connections
> on the server side?
> >
> > Thanks,
> > Vitaly
> >
> > _______________________________________________
> > riak-users mailing list
> > riak-users at lists.basho.com
> > http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.basho.com/pipermail/riak-users_lists.basho.com/attachments/20150410/1bbe32a8/attachment-0002.html>


More information about the riak-users mailing list