From our experience, on a windows server, with some jobs I'd suggest that you look in Task Manager and see if it is one of the java processes taking all the CPU. If it is then increase the java memory for the specific job. We found that some 'bounced the limit' and then started to hog CPU but giving them a bit more java memory enabled them to run very quickly and not hog CPU. We do have a revisit to the jobs on the 'to do' list to see if we can reduce the amount of memory they need but the 'instant' fix is more java memory for the 'problem' jobs.
I meant sometimes the job while running as standalone job hangs in between and doesnt process further. Due to which the CPU usage goes 100% and also one thing is we have used parallel execution in the job.Let me know if you still confused with my question
Ah right, when the Java memory is too small we think, that Talend starts paging the memory and seems to tie its self up in doing that while doing very little of the job and using 100% cpu. As I say, for us, in these cases just using the 'JVM Parameters' tab for the job and adding '-Xmx####M', where #### is the amount of memory in mega bites to add, fixes the issue. As far as we can tell this happens for us when the job retrieves a large amount of data and then has to do some other processing with the data. Our DB is on another machine.
