java - Eclipse WorkerPool does not start new Workers for Jobs -
in rcp eclipse application have number of sessions each run controller run job. each controller makes request/response-communication should synchrounous per controller run in job (for display of progress , parallelism of long running requests in different sessions).
in 99% of cases works correctly. there no new worker objects produced in workerpool , spare ones in pool gone. have multiple heap-dumps our customers proof.
the calling code follows:
job job = new job("name") { // request/response work, try/catch errors -> alsways return status.ok } job.schedule(5l); job.join(jobdeadlocktimeout,monitor); // wait each request/response job before next can scheduled in sync per session
usually in heap-dump can see workers controller-jobs , spare (sleeping) worker-objects used request/response jobs.
in case application hangs can see controller threads in workerpool , no sleeping or unused workers. have been removed no new workers produced. heap-dumps state eclipse jobmanager active , not suspended. threads waiting on job.join none of them can completed because don't have worker-threads complete work.
unfortunately no 1 can trigger problem emerges after hours or days, , @ few users @ different locations/customers.
the system-specifications follows:
- windows 10, 8 , 7
- java 1.8.0 - 131 32bit
- eclipse-rcp-platform based on eclipse 4.6 neon 3
has experience sort of behaviour or hint elese check ?
Comments
Post a Comment