Hi Ilya, Regarding exchange worker, the jira saying that 2.8.1 and 2.9 but we are using 2.7.6 version. So it could be in 2.7.6 as well. But still I am not understanding here is, there is no node join/left or cluster activation or cache creation near to the error logging time. Don't know why PME triggered.
Regarding communication worker first four thread dump, if we see it's blocked 110secs it's huge time. It's happening in production, it's should not blocked this much time even we have given failure detection time out and networktimeout 60 secs. I don't think there is network issue in prod, there is something which blocked this one. We have around 300 thick clients connected to the prod cluster. Both are very important for us as it's in prod, any more pointers which could have cause these issues. Regards, Gangaiah ----- Thanks and Regards, Gangaiah -- Sent from: http://apache-ignite-users.70518.x6.nabble.com/
