Calls for using jchem operations hangs

Hi.





I was setting up users synonyms, when suddenly that stopped working.


Oracle works, but any call to jchem specific synonyms just hangs.


I've tried to restart the tomcat server but that does not help.





Any suggestions on what to do ?





Regards


Mikael

Mikael,





It's most probably some of the objects in the cartridge owner's schema (which the synonyms are created for) which got stuck locked for some reason. I'd bet one of your CREATE SYNONYM ... statement is also hanging. (If not on the client side, the Oracle server might see so.) Have you checked pending locks in the instance?





Thanks


Peter

After about 10 minutes the lock went away.


However, It looks like the lock will occure when I try to setup synonyms for one user, and another user is using any of the synonyms.





Do you know if that is true ?





/Mikael

I vaguely remember having similar problems, but I don't remember the exact context.





Are you able to identify (using the data dictionary views) the objects, sessions and SQL statements involved?





Logically, unless you're trying to drop/replace/alter the objects themselves, Oracle should use some kind of "read" locks (not mutually exclusive locks) to alter privileges or synonyms on the objects. Still, I am not sure if the notion of multilevel locks exists in this context. (I also don't see any immediate reason why it shouldn't exist, but who knows...)





I've just performed a minitest: I executed a query using jc_compare and created a synonym for jc_compare for a third user. No hang observed. It is possible, however, that doing the same with more complex composite objects results into a different behaviour.





Whatever the exact reason for your problem, it would be helpful if you could provide details on who locks what by doing what (which ultimately results in all cartridge operations waiting on that lock).





Thanks


Peter

Hi.


Ok, I'll try and do that when things are a bit calmer.





/Mikael

Today I got a outofmemory error, I then increased the memeory from 360 to 720 and restarted the apache server.


After that I retried the structure search operation, it seemed to run, but was taking a long time, it turned out that all calls to jchem_core_pkg results in a hang .








Those hangs does not show up on locks in the database, so it seems they are already parsed to the cartridge.





Even if I restart the apache server, a simple jchem_core_pkg.getenvironment() hangs.





What mechanism in the cartridge can produce a hang like this ?,


the only way I can get rid of it (as far as I know now) is to restart the database.





/Mikael

I found an old process with the old memory configuration running that was still running dispite the shutdown of the apache , and since that was listening on the standard port, the new start of apache would fail.





I'll monitor that to see if I nailed down the problem.





Regards,


Mikael