ignite-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Vladimir Ozerov <voze...@gridgain.com>
Subject Re: igfs-meta behavior when node restarts
Date Thu, 08 Jun 2017 07:58:57 GMT

Is it possible to create isolated reproducer which will show gradual
increase of failed operations over time?

On Wed, Jun 7, 2017 at 6:39 AM, joewang <me@joewang.net> wrote:

> Would it be helpful if I uploaded logs collected from the cluster? I can
> point to the time when the behavior begins to occur.
> My concern with this is that the recurrent transactions (removing entries
> from the TRASH) and rollbacks are being performed over and over again for
> the same set of entries -- over time, this set continues to grow. In my
> observation, over the course of 3-4 days, the initial 4 TX op/s grows to
> ~200 TX op/s across all my nodes. I'm assuming if this grows unbounded, the
> cluster performance and stability will eventually be affected.
> --
> View this message in context: http://apache-ignite-users.
> 70518.x6.nabble.com/igfs-meta-behavior-when-node-restarts-
> tp13155p13446.html
> Sent from the Apache Ignite Users mailing list archive at Nabble.com.

View raw message