It's obviously bug. NPE occures if node doesn't have cache but got respone message during deadlock detection about thiscache on toher node. It's possible for example when you have node filter and caches start on some subset nodes of cluster. Is it your case?
I'll try to reproduce this problem, create JIRA ticket and answer here. On Mon, Apr 3, 2017 at 8:23 PM, bintisepaha <binti.sep...@tudor.com> wrote: > Sorry for the late response. We do not close/destroy caches dynamically. > could you please explain this NPE? > > > > -- > 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-tp10536p11676.html > Sent from the Apache Ignite Users mailing list archive at Nabble.com.