It's all about the answers!

Ask a question

High CPU Usage of DB2 Process after Upgrade to RTC 4.0.4

Daniel Reinhardt (3849) | asked Nov 21 '13, 4:00 a.m.
edited Nov 21 '13, 4:01 a.m.
Were running RTC 4.0.4 with DB2 10.1 Express-C on RHEL 6.
The server is configured with 4 cores and 8GB RAM.

After the upgrade from 4.0.3 to 4.0.4 we see the db2sysc processs running at nearly 200%.

Linux top-command says:
 2234 db2inst1  20   0 7540m 890m 581m S 199.9 11.3 127028:57 db2sysc

Prior to 4.0.4 it used only very little cpu.
However, the machine and RTC runs very smooth, no user complaints, we only discovered this issue through our monitoring tool.

Any suggestions on how to solve this?


Kevin Ramer commented Nov 21 '13, 2:23 p.m.

There are a couple db2 utilities that might shed light on this in addition to what you might see in 'top' or iostat.   I like db2top.  Run as the instance owner

db2top -d DBNAME 

where DBNAME is one of the databases.

This can show session activity, table/tablespace activity, locks, etc.  I recommend these pages

  • sessions (lowercase L) displays agents.  Typing i will suppress inactive
  • Tables ( T )  table activity
  • Locks ( U )  there should be none or few of these

 You may or may not find the 'smoking gun'.

db2diag  lets you query the sqllib/db2dump/db2diag.log in many ways. 

nmon  this one is like top or aix topas and you can run it in the background and collect lots of stats.   There's an nmon analyzer Excel worksheet that can import nmon files and draw nice graphs.  pGraph (a java tool) can also read the nmon files in case you're unable to use Excel.

Hope this helps

One answer

permanent link
Daniel Reinhardt (3849) | answered Nov 22 '13, 3:00 a.m.
with db2top I found only low activity, but there are some locks on the CCM-Database:

          Locks held.......:          7 [0,00%]
          Agents waiting...:          0
          Appls Connected..:          5

                    Agent        Application          Application
                    Id(State)    Name                 Status
                    ------------ -------------------- -----------------------------------
                        25479(*) DSSNAP_logMessage    UOW Executing
                        25479(*) DSSNAP_logMessage    UOW Executing
                        25470(c) db2evmt_JAZZ_DEADLOC Connected
                        25470(c) db2evmt_JAZZ_DEADLOC Connected
                        25470(c) db2evmt_JAZZ_DEADLOC Connected
                        25470(c) db2evmt_JAZZ_DEADLOC Connected
                        25470(c) db2evmt_JAZZ_DEADLOC Connected

Is this normal behavior?

Kevin Ramer commented Nov 22 '13, 8:02 a.m.


I've seen the JAZZ_DEADLOC(K).   The DSSNAP I've never seen; but it's probably 'normal'.  The thing to watch for there is "Agents waiting" > 0 for long periods of time.

Googled and found:
See the comments by "Marcus A"

The chase:  you may be getting misled by what you see.

Your answer

Register or to post your answer.

Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.