[
https://issues.apache.org/jira/browse/IGNITE-6313?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16210936#comment-16210936
]
Ilya Lantukh commented on IGNITE-6313:
--------------------------------------
The test fails because Commiter threads die due to uncatched exception - either
ClusterTopologyException or CacheException with ClusterTopologyException as
cause. As I understand, this is expected behavior. If you add catch {} block
for these exceptions, test will pass with 100% probability.
> transaction hangs while node left cluster
> -----------------------------------------
>
> Key: IGNITE-6313
> URL: https://issues.apache.org/jira/browse/IGNITE-6313
> Project: Ignite
> Issue Type: Bug
> Affects Versions: 2.1
> Reporter: Stanilovsky Evgeny
> Assignee: Ilya Lantukh
> Attachments: TxDeadLockOnNodeRestart.java, run.py, tpre.tar.gz,
> trun.tar.gz
>
>
> in attached reproducer there are 2 projects,
> first - tpre simple fills cache
> second - trun concurrently update cache values.
> expected behavior 100 or sometimes 99 numbers in output, but after several
> grid autorestart python script we found grid have no active transactions ...
> try to find time for pure ignite reproducer, but for now it need be run like :
> sequentially start:
> *java -jar tpre-1.0-SNAPSHOT-jar-with-dependencies.jar 1000
> java -jar trun-1.0-SNAPSHOT-jar-with-dependencies.jar 1000 20 1
> run.py*
> console with *trun* process will soon output something like : BBB that
> signals no transactions found.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
