I saw a lot of discussion on this topic in https://issues.apache.org/jira/browse/GEODE-332. Does that help?
Anthony > On Nov 10, 2015, at 4:50 AM, Hovhannes Antonyan <[email protected]> wrote: > > Who are the AsyncSocketCloser threads? > I had an issue once when in my multi node deployment one of the nodes was > heavily performing garbage collection and eventually crashed OOM. > This was due to bug in business logic, however looking at the crash heapdump > I saw large amount of threads named "AsyncSocketCloser for > <IP>(name)<v12>:PORT". Just curious who do these threads are. > > An interestingly when this happened (only one of the nodes was under GC), the > gemfire client's weren't able to establish connection with the system. > So there were around 500 AsyncSocketCloser threads and also around 2400 P2P > handshake reader threads. > I am trying to understand what was the reason the gemfire clients weren't > connecting to other nodes. > Could it because of reaching some limit on threads numbers or probably is was > due to other reasons? > Is there any limit on amount of P2P handshake reader threads? Is there any > limit on ServerConnection threads?
signature.asc
Description: Message signed with OpenPGP using GPGMail
