Zombie graphs: Runtime env. release has been already called

Hi, we are having problems with graphs that fail/abort for any reason but the graph is still running, like a zombi process. The CloverETL list them are RUNNING, and that’s been the case for at least 700+ hours. Even by killing them thtrough the Monitoring Window, the log says this:


2014-05-19 14:10:10,448 INFO  30118172 [jdkTimerFactory] Runtime env. release has been already called! runId=30118172
2014-05-19 14:10:30,449 INFO  30118172 [jdkTimerFactory] Runtime env. release has been already called! runId=30118172
2014-05-19 14:10:50,451 INFO  30118172 [jdkTimerFactory] Runtime env. release has been already called! runId=30118172
2014-05-19 14:11:10,452 INFO  30118172 [jdkTimerFactory] Runtime env. release has been already called! runId=30118172
2014-05-19 14:11:30,455 INFO  30118172 [jdkTimerFactory] Runtime env. release has been already called! runId=30118172
2014-05-19 14:11:34,395 INFO  30118172 [1374003118@qtp-1699608420-575] Runtime env. release has been already called! runId=30118172
2014-05-19 14:11:50,457 INFO  30118172 [jdkTimerFactory] Runtime env. release has been already called! runId=30118172

It’s been doing this for more than a month now. That second to last line was me trying to kill it from the Monitoring window. We have a whole bunch of these guys, doing the same, what’s going on there? Is there an effective way to kill them and recover the resources? I there something we can do to fix this behavior?

I really appreciate your help!

We are using version 3.2.1
build.number 63
build.time 2012-01-04 13:40:23
build.user hudson
container jetty6
obfuscation obfuscated
version 3.2.1

Does anybody have a fix for this ?

Hi jfuentesve,

Since 3.2 releases we were working a lot on CloverETL Server stability. We do not have issue in our bug tracker reporting same problem as you have. But it may be it was fixed as side effect of another changes. So I would either try to upgrade to 3.5 or reboot CloverETL Server periodically.