RQM floating licences swithed to free ones
Hello!
Today happened something strange with our Jazz CLM Server v4.0.7! Several hundredth of "Rational Quality Manager - Professional-Floating" licenses suddenly switched to "Rational Quality Manager - Contributor" ones.
Accordingly, the users weren't able login to server when they came to work.
Strange because switching happened early morning when nobody couldn't work in the system and switch the licenses. Next report displays that the time when switching happened is about 07:00AM in accordance to the chart:
I analyzed logs but can not find any problematic behaviour. If somebody could give an advice in clarrifying the situation, I would be thankful very much.
Thank you in advance!
Today happened something strange with our Jazz CLM Server v4.0.7! Several hundredth of "Rational Quality Manager - Professional-Floating" licenses suddenly switched to "Rational Quality Manager - Contributor" ones.
Accordingly, the users weren't able login to server when they came to work.
Strange because switching happened early morning when nobody couldn't work in the system and switch the licenses. Next report displays that the time when switching happened is about 07:00AM in accordance to the chart:
I analyzed logs but can not find any problematic behaviour. If somebody could give an advice in clarrifying the situation, I would be thankful very much.
Thank you in advance!
One answer
Dmirty, if you're still concerned, please consider opening a PMR.
Comments
Paul, we were having the problem two times. Those days, we were deploying hard BIRT report which consumed a lot of server resources. As well ,that fact was confirmed by some external software which monitors for operability of all resources in the Customer's network (it was fixed that server resources consumption reached 100%).
Afterwards, we optimized the report and the problem is disappeared. We believe that it was resulting of Overflow Memory that was giving such a strange behavior.
Comments
Dmitry A. Lesin
Oct 26 '16, 9:20 a.m.Although, I am wrong. The system was failed yesterday at 07:00PM in accordance to picture above. In relation to that time, I found a problem in native_stderr.log (WAS v8.0.0.3):
Log file started at: [10/21/16 2:30:06:842 MSK]
**** End Display Current Environment ****
Exception in thread "Bundle File Closer" JVMDUMP039I Processing dump event "systhrow", detail "java/lang/OutOfMemoryError" at 2016/10/21 19:18:24 - please wait.
JVMDUMP039I Processing dump event "systhrow", detail "java/lang/OutOfMemoryError" at 2016/10/21 19:18:24 - please wait.
JVMDUMP039I Processing dump event "systhrow", detail "java/lang/OutOfMemoryError" at 2016/10/21 19:18:24 - please wait.
...
May be this is a reason of the problem. But why licences were switched is a very big question!
Donald Nong
Nov 01 '16, 1:35 a.m.In your screen shot, the license type is "connector", not "contributor". It's always there, so I don't see any "switch". It could be that due to the OOM error, JTS was thrashed and could no longer serve any licenses.
Dmitry A. Lesin
Nov 01 '16, 1:56 a.m.Hello Donald,
Sorry that I did not explain the problem correctly. Really happened next - all users which were working under commercially purchased floating license unexpectedly were switched to a non-commercial one. The problem was quickly solved by manual reassigning the licences back. But yesterday morning, the situation was repeated again!
It must be mentioned that for last several years this happened only two times but during a week only.
I can not find too many records in the logs regarding this case. I have a plan to increase the logs detalization and try to catch the problems in the log.
Donald, you wrote that JTS is trashed. You mean that we need try full re-indexation, did I understand you correctly?
Thank you in advance!
Donald Nong
Nov 01 '16, 2:41 a.m.No. When OOM occurs, you just cannot predict what can go wrong. I just suspect that OOM (and how it affects JTS) could contribute to the strange symptom that you have observed.
Dmitry A. Lesin
Nov 01 '16, 3:18 a.m.Thank you!
I'm thinking now that may be initially the problem is related to our activities in last days. We are developing a new BIRT report for RQM, and several days ago we deployed it from our development environment in the production one. Currently, the main problem is to enhance its performance and optimize the memory used. May be the lack of resources was followed for these operations when we were running the report for debugging.