Vladimir, right now I have no idea what is happening. Maybe Community will help. The stack trace goes deep into Big-Endian code that I cannot even troubleshoot since I have no Big-Endian platform available. Gathering logs might be helpful: I would stop all the nodes, clean $IGNITE_HOME/work/logs/*, start running the nodes one-by-one until the problem is reproduced and then share logs from all the nodes with the community.
On Thu, Nov 2, 2017 at 4:10 PM, Vladimir <vlads...@yandex.ru> wrote: > No single and simple configuration is available. But notging special: > default > or slightly changed memory policy, replicated and partitioned caches > (mostly > atomic), write-through persistence, tcp-discovery, continous queries, > services > > > > -- > Sent from: http://apache-ignite-users.70518.x6.nabble.com/ > -- Best regards, Alexey