J Z
asked on
Oracle taking 50% CPU constantly
Question for the oracle experts:
We have a VM running oracle 11g (I'm not sure which version) that has 1 vCPU with 4 cores. It has 2 DBs and 2 oracle.exe processes which a few minutes after restart each start start taking 50% of one of the cores. The other 2 other cores are left unused. That makes a total usage of 25% (2 x 50% of 2 cores on a total of 4 cores).
We don't know what is generating the CPU usage and why it's a constant. In any case it is a flat line we see all the time once it starts. But we do know that also when the machine is completely network disconnected we see the same thing happening. It only stops after you stop the service and restart it. But after a few minutes it starts again.
Question is why is it doing this what is the easiest way to figure out what is causing this constant 50% CPU load on 2 cores.
It has to be something internally on the Oracle since it happens even if the machine is completely isolated.
Thanks
We have a VM running oracle 11g (I'm not sure which version) that has 1 vCPU with 4 cores. It has 2 DBs and 2 oracle.exe processes which a few minutes after restart each start start taking 50% of one of the cores. The other 2 other cores are left unused. That makes a total usage of 25% (2 x 50% of 2 cores on a total of 4 cores).
We don't know what is generating the CPU usage and why it's a constant. In any case it is a flat line we see all the time once it starts. But we do know that also when the machine is completely network disconnected we see the same thing happening. It only stops after you stop the service and restart it. But after a few minutes it starts again.
Question is why is it doing this what is the easiest way to figure out what is causing this constant 50% CPU load on 2 cores.
It has to be something internally on the Oracle since it happens even if the machine is completely isolated.
Thanks
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
I'm positive that it is version 11g.
Anyone any more proposals?
An Oracle expert to take a look is also welcome...
Anyone any more proposals?
An Oracle expert to take a look is also welcome...
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
OK, It is a Windows installation but in task manager the processes that are taking the constant CPU load of 50% each on one of the cores is just oracle.exe. Other processing are doing nothing. Does you explanation then still apply?
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
I appreciate all the help. But we are looking in the wrong direction I think.
The reason that makes me believe that is that the server had exactly the same config before they upgraded their code and made change to the database. Only then it started.
The strange thing I see is that on 2 database we see unusual high portions of CPU for certain queries that it should be doing when querying the TOP SQL calls by CPU. Seeing that looks like t a loop in the code to me. The problem is how do I proof that to the developers? :-)
The reason that makes me believe that is that the server had exactly the same config before they upgraded their code and made change to the database. Only then it started.
The strange thing I see is that on 2 database we see unusual high portions of CPU for certain queries that it should be doing when querying the TOP SQL calls by CPU. Seeing that looks like t a loop in the code to me. The problem is how do I proof that to the developers? :-)
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
OK, I took all the screenshots and queries in a PDF. But it seems I forgot to upload them. So here's a second attempt...
I hope these queries and screenshots help a bit. Mind that it's 2 databases which both are showing a loop or bug in the code according to me.
Top-SQL-queries-for-Experts-Exchange.pdf
I hope these queries and screenshots help a bit. Mind that it's 2 databases which both are showing a loop or bug in the code according to me.
Top-SQL-queries-for-Experts-Exchange.pdf
ASKER
additional problem is that it's not in-house developed software. It's a package we buy from fuji. So I don't have a view on the code. But the problem started ever since they made changes in November.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
I chose my own solution since it's not really solved and that is the solution people will see as the final answer when they search for this problem. Thanks again to everyone
ASKER
I explained it in my last answer. The problem was not solved.
ASKER
select job_name,state from dba_scheduler_jobs;
Open in new window
select job, this_date from dba_jobs;
Open in new window
Any other ideas?
I want to start the ADDM. But I can't find how to open it. It says the Database Homepage. But that doesn't seem to run.