cancel
Showing results for 
Search instead for 
Did you mean: 

Problem with Oracle Processes - ora_cjq0_<SID> and ora_qmnc_<SID>

Former Member
0 Kudos
1,466

Hello,

We are having problem with few Oracle Processes which are failing by themselves and also cosuming high CPU load. Few of the Processes are :

ora_cjq0_<SID>

ora_qmnc_<SID>

I did not find any relevant SAP Note about the above two Processes.

Could you please throw some light on the above two processes? Will they stop being acitve once their work is done?

Thanks,

Subbu

Accepted Solutions (0)

Answers (2)

Answers (2)

Former Member
0 Kudos

Hi,

Could you please show us some error logs from trace files or alert.log?

Also, would request you to check if the latest SBP has been implemented or not.

And of course, the Oracle parameters as well.

830576 Parameter recommendations for Oracle 10g

Br,

Venky.

Former Member
0 Kudos

Hi Subbu,

ora_cjq0_<SID>: It means "Job Queue Process". This process is used for job scheduler

ora_qmnc_<SID>: It means "Queue Monitor Process". This process is used for Oracle Streams Advanced Queuing.

Check the database jobs on the database by executing the statement, below;

select * from dba_jobs;

And check currently running jobs;

SELECT * FROM user_jobs_running;

Best regards,

Orkun Gedik

Former Member
0 Kudos

Hello,

Thanks for the response. We are on Oracle 10.2.0.5.

select * from dba_jobs; - returned no rows

SELECT * FROM user_jobs_running; - Gave me an error that Table doesnt exist.

i have a basic idea of these processes which has been mentioned below. But I need info like - will these processes be killed by Oracle automatically when not in use for some period of time?

How to monitor these proceeses CPU usage and optimize their performance as CPU utilization is high by the above two.

We have a tool which will raise a ticket when any process is killed. Daily we are receiving a ticket that above processes are killed...

Thanks,

Subbu

Former Member
0 Kudos

Hi,

You can have further information, if the trace files exist, belong to the these processes under "/oracle/<SID>/saptrace/background".

Best regards,

Orkun Gedik

Former Member
0 Kudos

Hello All,

Thanks for the response.

Parameter AQ_TM_PROCESSES is already 0 as suggsted. As already informed my problem is we are receiving tickets daily from the automatic monitoring tool saying "Oracle processes are killed.." My doubt remians, will Oracle kill these processes when it is not doing any work??

My Os team is complaining that the above two processes are consuming more CPU memory. I found only about the start and stop info. of the above processes in alert log file, i couldnt find any Trace file related to these..

Could you please help how to start my analysis.

Thanks,

Subbu

Former Member
0 Kudos

Hi Subbu,

CJQ0 and QMNC processes waits at the ACTIVE state. There's no problem with it.

The main problem is identifying why these processes are getting killed. So, let's go through step by step, in order to identify the problem.

As a first question, to be sure, I would like to ask that what is your Oracle release? And did you applied latest Oracle patches into the system.

At th second step, you can start the analyze by measuring wait times, on these processes. To check this, go to ST04 -> Performance -> Wait Event Analysis -> System Event -> Busy Waits Summary. Then, check the "Busy Wait %" and "Busy wait time (ms)" columns, in order to find out how much these processes are loading on the database.

At the third step, check are these processes killed at the same time period, everyday? Or randomly?

Additionally, in my previous message, I noted that did you check whether exist the trace files belong to QMNC and CJQ0 processes under "/oracle/<SID>/saptrace/background"? These files are named as "<sid>_qmnc_<process_id>.trc" and "<sid>_cjq0_<process_id>.trc". If they are exist, could you provide their content, here?

Best regards,

Orkun Gedik

Former Member
0 Kudos

Hello Orkun Gedik,

Thank you for the help. We are on 10.2.0.5. We have applied all the patches too for this version.

I have been to St04 and found below:

Session Type UserName ProcessName Sessions Busy WaitTime(ms) Total WaitTime(ms) Busy W%

USER SAPSR3DB CJQ0 1 3.437 182.871 1,88

The reason why i didnt find a log in /Oracle/<SID>/saptrace/background is the session type of process CJQ0 is USER type.

But as you know, logs in UserTrace will not be differentiated by Process and its hard to find a trace for this process.

Can you help me by reading above values?

Thanks,

Subbu

Former Member
0 Kudos

Hello,

One more info. to add apart my previous post is I am seeing a WARNING in all the User trace files as below:

WARNING:Could not lower the asynch I/O limit to 144 for SQL direct I/O. It is set to -1

Is this warnign to be considered? I googled about the above error but couldnt get any hint..

Thanks,

Subbu

Former Member
0 Kudos

Hello,

>

> One more info. to add apart my previous post is I am seeing a WARNING in all the User trace files as below:

>

> WARNING:Could not lower the asynch I/O limit to 144 for SQL direct I/O. It is set to -1

>

> Is this warnign to be considered? I googled about the above error but couldnt get any hint..

>

> Thanks,

> Subbu

Hi Subbu,

You hit the bug 8412426 "WARNING:Could not lower the asynch I/O limit" messages. Check the document, below;

https://support.oracle.com/CSP/main/article?cmd=show&type=NOT&doctype=PATCH&id=8412426.8&addClickInf... search_text="%22Could%20not%20lower%20the%20asynch%20I/O%20limit%22" search_result_size="16" search_result_count="16" powerview_id="" on_off="off" item_position_in_list="1"/>

As far as I understand that this bug has no side effect, except filling your trace file, consequently the storage. Oracle suggest upgrade to 11g.

Best regards,

Orkun Gedik

Former Member
0 Kudos

Hi,

>> Session Type UserName ProcessName Sessions Busy WaitTime(ms) Total WaitTime(ms) Busy W%

>> USER SAPSR3DB CJQ0 1 3.437 182.871 1,88

>> The reason why i didnt find a log in /Oracle/<SID>/saptrace/background is the session type of process CJQ0 is USER type.

>> But as you know, logs in UserTrace will not be differentiated by Process and its hard to find a trace for this process.

Under the normal conditions, CJQ process should be a background session, but is a user session, in your case. With this parameter, I understand that is is forked by a client application. In order to find out root cause, I recommend you check remote connections to the database and user traces, as well. From this point on, it is really hard to guess what's running at the backstage, without connecting to your system.

Best regards,

Orkun Gedik