Andrey, thanks for getting back. The long pause is between 2 different threads, so isn't that normal?
Also the 9990 ms and 10 ms used earlier for some previous step in the tn, is this how ignite breaks out the time? we have always seen the timeouts for the value in ms that we set in our code. Also the stack trace does not come from our code which it usually does on genuine timeouts that do not leave the key locked. We are trying 1.8 in UAT environment and will release to production soon. Unfortunately this issue does not happen in UAT and we have no way of reproducing it. Is there any way we can force release a locked key without restarting the whole cluster? Thanks, Binti -- 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-tp10536p10910.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.