You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@ignite.apache.org by Evgenii Zhuravlev <e....@gmail.com> on 2017/06/02 15:06:16 UTC

Re: Cache has been stopped

Hi,

I think you should find the reason why nodes segmented at first.

As you have small heap size(044Gb per node), it's most likely that nodes
segmentation was cause by GC pressure, you can check it by turning on GC
logs. As you see, (CPU [cur=31.33%, avg=1.42%, GC=50.87%]) , your CPU
spends more than 50% of time in GC, it's too much.

Just give more heap size to ignite nodes.

Evgenii

2017-05-30 12:40 GMT+03:00 jpmoore40 <jo...@ca-cib.com>:

> It's happened again today requiring a manual restart of all nodes. I had
> previously changed the VM options as described here:
>
> https://apacheignite.readme.io/docs/jvm-and-system-tuning#
> jvm-tuning-for-clusters-with-on_heap-caches
>
> but it doesn't seem to have made any difference. So my question is where
> can
> I find documentation on how to accurately detect and repair segmentation
> issues? I've searched on this and it looks like I should be setting the
> segmentation policy to NOOP and listening for EVT_NODE_SEGMENTED events,
> but
> I'm unsure on what to do when this happens. This link:
>
> http://apache-ignite-users.70518.x6.nabble.com/Ignite-cluster-td9527.html
>
> seems to indicate that a full restart would be required in this situation
> anyway. To clarify I simply have a publisher that is posting data to a
> queue
> and several workers that are pulling the data from the queue and processing
> it.
>
> One further point - we have been using peer class loading up until now
> though this will be switched off on our next release due this week - could
> that be having any effect on this?
>
> Thanks
> Jon
>
>
>
>
>
> --
> View this message in context: http://apache-ignite-users.
> 70518.x6.nabble.com/Cache-has-been-stopped-tp13056p13217.html
> Sent from the Apache Ignite Users mailing list archive at Nabble.com.
>

Re: Cache has been stopped

Posted by jpmoore40 <jo...@ca-cib.com>.
Thanks - I will increase the memory and see how it holds out



--
View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Cache-has-been-stopped-tp13056p13380.html
Sent from the Apache Ignite Users mailing list archive at Nabble.com.