You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@ignite.apache.org by Denis Magda <dm...@gridgain.com> on 2016/02/17 12:18:20 UTC

Please review

Igniters,

Please review the critical minor fixes for the following tickets

IGNITE-2647:Cache is undeployed even when BinaryMarshaller is used 
<https://issues.apache.org/jira/browse/IGNITE-2647>

IGNITE-2666: Node is not stopped when RingMessageWorker is interrupted 
<https://issues.apache.org/jira/browse/IGNITE-2666>

In my understanding the review should be done either by Alex G. or Yakov 
or Sam.

--
Denis

Re: Please review

Posted by Alexey Goncharuk <al...@gmail.com>.

Denis,

Changes in IGNITE-2647 look good to me.

Re: Please review

Posted by Denis Magda <dm...@gridgain.com>.

Hi Sam,

I saw some users reporting the following error

[tcp-disco-msg-worker-#2%null%] o.a.i.s.d.tcp.TcpDiscoverySpi -  TcpDiscoverSpi'smessage worker thread failed abnormally. Stopping the node in order to prevent cluster wide instability.
	java.lang.InterruptedException: null
		at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
		at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088)
		at java.util.concurrent.LinkedBlockingDeque.pollFirst(LinkedBlockingDeque.java:522)
		at java.util.concurrent.LinkedBlockingDeque.poll(LinkedBlockingDeque.java:684)


This exception is shown only when a node is not being stopped explicitly 
by the user meaning that someone else interrupted this thread.
Probably it's a reasonable interruption. That's why in addition I 
decided that we need to store a stack trace of the thread that 
interrupted  the ring message worker.

--
Denis

On 2/18/2016 1:44 PM, Semyon Boikov wrote:
> Hi Denis,
>
> Yakov and me looked at changes at ignite-2666 and we do not understand 
> how ring message worker can be interrupted except case when node is 
> stopped? Should we care about this case?
>
> On Wed, Feb 17, 2016 at 2:18 PM, Denis Magda <dmagda@gridgain.com 
> <ma...@gridgain.com>> wrote:
>
>     Igniters,
>
>     Please review the critical minor fixes for the following tickets
>
>     IGNITE-2647:Cache is undeployed even when BinaryMarshaller is used
>     <https://issues.apache.org/jira/browse/IGNITE-2647>
>
>     IGNITE-2666: Node is not stopped when RingMessageWorker is
>     interrupted <https://issues.apache.org/jira/browse/IGNITE-2666>
>
>     In my understanding the review should be done either by Alex G. or
>     Yakov or Sam.
>
>     --
>     Denis
>
>

Re: Please review

Posted by Semyon Boikov <sb...@gridgain.com>.

Hi Denis,

Yakov and me looked at changes at ignite-2666 and we do not understand how
ring message worker can be interrupted except case when node is stopped?
Should we care about this case?

On Wed, Feb 17, 2016 at 2:18 PM, Denis Magda <dm...@gridgain.com> wrote:

> Igniters,
>
> Please review the critical minor fixes for the following tickets
>
> IGNITE-2647:Cache is undeployed even when BinaryMarshaller is used <
> https://issues.apache.org/jira/browse/IGNITE-2647>
>
> IGNITE-2666: Node is not stopped when RingMessageWorker is interrupted <
> https://issues.apache.org/jira/browse/IGNITE-2666>
>
> In my understanding the review should be done either by Alex G. or Yakov
> or Sam.
>
> --
> Denis
>