Hi, issue mentioned early is related with both pessimistic and optimistic transactions. So it is still possible that it is your case.
Anyway, I don't have any other ideas about the problem. It would be great to have reproducer in order to debug issue. But from my point of view migaration to newer version is the best way. On Fri, May 19, 2017 at 12:23 PM, yfernando <[email protected]> wrote: > The issue you mention doesn't seem to be related to the issue we are having > as, > - The entire grid does not lock up as in the situation described in the bug > - There are no threads blocked or locked when our key lock occurs > - The bug seems to occur on Optimistic locking whereas our scenario occurs > on Pessimistic > > > agura wrote >> There was a problem with incorrect transaction timeout handling [1] >> that was fixed in Ignite 1.8. It is possible that it is your case. >> >> [1] https://issues.apache.org/jira/browse/IGNITE-2797 > > Another observation we had was that the key that was locked occurred on a > cache that gets loaded at startup and never changes content. The cache does > however take part in transactions in terms of cache reads (using > IgniteCache.get() ) > > > > -- > View this message in context: > http://apache-ignite-users.70518.x6.nabble.com/Pessimistic-TXN-did-not-release-lock-on-a-key-all-subsequent-txns-failed-tp10536p13021.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com.
