JobexWatchdog killed due to Out of memory

[Former Member]

Dear fellow practitioners.

Jobex Watchdog thread of HANA DB(fa/hana1sp12) in my Production CRM System got killed, citing out of memory issue.

[7546884.294078] Out of memory: Kill process 1615 (JobexWatchdog) score 962 or sacrifice child

[7546884.294084] Killed process 1615 (JobexWatchdog) total-vm:515474356kB, anon-rss:413832784kB, file-rss:95323348kB

# ps -ef | grep -i JobexWatchdog

root 64847 64575 0 02:46 pts/0 00:00:00 grep -i JobexWatchdog

I could only find note#

2114710 - FAQ: SAP HANA Threads and Thread Samples

having jobexwatchdog information.

"Watchdog for job execution, collection of statistics (e.g. for controlling OLAP / OLTP share)"

Jobex Watchdog is not running anymore. Would it affect the functioning of DB?

How can I get it started without bouncing the database? Moreover, I could not find the root cause. Which parameter do I need to tune to have it corrected from occurring in future

Regards.... Khan

[lbreddemann · Accepted solution]

There is no way for you to restart this HANA internal thread. The 'ps' command on OS level won't ever show you this thread, so, it's not an indication for whether or not the instance currently runs a thread of type "JobexWatchdog".

What you can do instead is to check the M_SERVICE_THREADS view:

select thread_id, thread_type 
 from "M_SERVICE_THREADS" 
where thread_type = 'JobexWatchdog' ;

Even though the thread got killed, it is a core component and HANA will restart it automatically.

If you, in fact, find that it is not running, then bouncing the instance is the only option to get it running again.

[Former Member · Accepted solution]

Thanks Lars,

It really helped I can see threads of type "JobexWatchdog".

Sorry for responding late.