Results 1 to 13 of 13
  1. #1
    Join Date
    Apr 2011
    Posts
    5

    Unanswered: db2fm - fault monitor service does not startup after issue db2stop

    hi all,
    I am encountering a weird issue here, i have configured db2fm as below:

    root@chialeep:/opt/ibm/db2/V9.7/bin# grep db2 /etc/inittab
    fmc:2345:respawn:/opt/ibm/db2/V9.7/bin/db2fmcd #DB2 Fault Monitor Coordinator

    started db2fm with below command:
    ./db2fm -i db2inst1 -m /home/db2inst1/sqllib/lib/libdb2gcf.so.1 -U -u

    however, when i issue db2stop via user: db2inst1 (instance owner), the fault monitor service will not able to startup by itself:

    root@chialeep:/opt/ibm/db2/V9.7/bin# grep db2 /etc/inittab
    fmc:2345:respawn:/opt/ibm/db2/V9.7/bin/db2fmcd #DB2 Fault Monitor Coordinator

    Any idea?

  2. #2
    Join Date
    Apr 2011
    Posts
    5
    Quote Originally Posted by vpong View Post
    hi all,
    I am encountering a weird issue here, i have configured db2fm as below:

    root@chialeep:/opt/ibm/db2/V9.7/bin# grep db2 /etc/inittab
    fmc:2345:respawn:/opt/ibm/db2/V9.7/bin/db2fmcd #DB2 Fault Monitor Coordinator

    started db2fm with below command:
    ./db2fm -i db2inst1 -m /home/db2inst1/sqllib/lib/libdb2gcf.so.1 -U -u

    however, when i issue db2stop via user: db2inst1 (instance owner), the fault monitor service will not able to startup by itself:

    root@chialeep:/opt/ibm/db2/V9.7/bin# grep db2 /etc/inittab
    fmc:2345:respawn:/opt/ibm/db2/V9.7/bin/db2fmcd #DB2 Fault Monitor Coordinator

    Any idea?
    I tried this, and it always shows down:
    root@chialeep:/opt/ibm/db2/V9.7/bin# ./db2fmcu
    FMC: down

  3. #3
    Join Date
    Aug 2008
    Location
    Toronto, Canada
    Posts
    2,367
    Quote Originally Posted by vpong View Post
    however, when i issue db2stop via user: db2inst1 (instance owner), the fault monitor service will not able to startup by itself
    "With Fault Monitor facility, if the DB2® Universal Database™ (DB2 UDB) instance exits prematurely, the fault monitor will restart it."

    db2stop != prematurely

    Check this technote for info on how to configure and test it:
    https://www-304.ibm.com/support/docv...id=swg21209001

  4. #4
    Join Date
    Apr 2011
    Posts
    5

    db2fm - fault monitor service does not startup after system restart

    hi BELLO4KA,
    Thanks for your replied. I have follow the link given by you to setup the fault monitor.

    root@chialeep:/opt/ibm/db2/V9.7/bin# cat /etc/inittab | grep db2fmcd
    fmc:2345:respawn:/opt/ibm/db2/V9.7/bin/db2fmcd #DB2 Fault Monitor Coordinator
    root@chialeep:/opt/ibm/db2/V9.7/bin# ps -ef | grep db2fm
    root 1618 1462 0 09:00 pts/0 00:00:00 grep --color=auto db2fm
    root@chialeep:/opt/ibm/db2/V9.7/bin# ./db2fm -i db2inst1 -U
    root@chialeep:/opt/ibm/db2/V9.7/bin# ./db2fm -i db2inst1 -u
    oot@chialeep:/opt/ibm/db2/V9.7/bin# ./db2fm -i db2inst1 -f on
    root@chialeep:/opt/ibm/db2/V9.7/bin# ./db2fm -i db2inst1 -a on
    root@chialeep:/opt/ibm/db2/V9.7/bin# ./db2fm -i db2inst1 -f on
    root@chialeep:/opt/ibm/db2/V9.7/bin# ./db2fm -i db2inst1 -a on


    But db2fmcu still showing:
    root@chialeep:/opt/ibm/db2/V9.7/bin# ./db2fmcu
    FMC: down

    Any idea why?

  5. #5
    Join Date
    Aug 2008
    Location
    Toronto, Canada
    Posts
    2,367
    Did you reboot the server? What's the output from step #11?

  6. #6
    Join Date
    Apr 2011
    Posts
    5

    db2fm - fault monitor service does not startup

    restarted the server and try to tail log:

    root@chialeep:/home/db2inst1# cd /home/db2inst1 && tail -f ./sqllib/db2dump/db2diag.log
    FUNCTION: DB2 Common, Fault Monitor Facility, db2fm, probe:70
    MESSAGE : ECF=0x900003A4=-1879047260=ECF_FM_POSSIBLE_TIME_HOLE_BEFORE_SET_RE AD_LOCK
    Possible timing hole between releasing write lock and getting read lock

    2011-04-23-09.28.33.592193+480 I82022G347 LEVEL: Warning
    PID : 2752 TID : 1629152 PROC : db2fm
    INSTANCE: db2inst1 NODE : 000
    FUNCTION: DB2 Common, Fault Monitor Facility, fmNotifyDaemon, probe:40
    MESSAGE : ECF=0x90000369=-1879047319=ECF_FM_NO_DAEMON_RUNNING
    There is no db2fmd daemon running


    Here is my fault monitor registry file content:
    root@chialeep:/home/db2inst1# cd /home/db2inst1 && tail -f ./sqllib/db2dump/db2diag.log
    FUNCTION: DB2 Common, Fault Monitor Facility, db2fm, probe:70
    MESSAGE : ECF=0x900003A4=-1879047260=ECF_FM_POSSIBLE_TIME_HOLE_BEFORE_SET_RE AD_LOCK
    Possible timing hole between releasing write lock and getting read lock

    2011-04-23-09.28.33.592193+480 I82022G347 LEVEL: Warning
    PID : 2752 TID : 1629152 PROC : db2fm
    INSTANCE: db2inst1 NODE : 000
    FUNCTION: DB2 Common, Fault Monitor Facility, fmNotifyDaemon, probe:40
    MESSAGE : ECF=0x90000369=-1879047319=ECF_FM_NO_DAEMON_RUNNING
    There is no db2fmd daemon running

  7. #7
    Join Date
    Aug 2008
    Location
    Toronto, Canada
    Posts
    2,367
    What's the output from step #11?

  8. #8
    Join Date
    Apr 2011
    Posts
    5

    db2fm - fault monitor service does not startup

    step11 output:

    chialeep@chialeep:~$ sudo su - db2inst1
    [sudo] password for chialeep:
    $ db2fm -s -S
    Gcf module 'fault monitor' is INSTALLED PROPERLY but NOT ALIVE
    Gcf module '/opt/ibm/db2/V9.7/lib32/libdb2gcf.so' is INSTALLED PROPERLY but NOT ALIVE
    Last edited by vpong; 04-22-11 at 22:50. Reason: add in result

  9. #9
    Join Date
    Aug 2008
    Location
    Toronto, Canada
    Posts
    2,367
    Something is not right. I'd suggest to redo it from the beginning, including db2fmcu -d, and save the output from each step. If this still doesn't work, consider opening a PMR.

  10. #10
    Join Date
    Sep 2013
    Posts
    2

    fault monitor in DB2

    Quote Originally Posted by vpong View Post
    hi all,
    I am encountering a weird issue here, i have configured db2fm as below:

    root@chialeep:/opt/ibm/db2/V9.7/bin# grep db2 /etc/inittab
    fmc:2345:respawn:/opt/ibm/db2/V9.7/bin/db2fmcd #DB2 Fault Monitor Coordinator

    started db2fm with below command:
    ./db2fm -i db2inst1 -m /home/db2inst1/sqllib/lib/libdb2gcf.so.1 -U -u

    however, when i issue db2stop via user: db2inst1 (instance owner), the fault monitor service will not able to startup by itself:

    root@chialeep:/opt/ibm/db2/V9.7/bin# grep db2 /etc/inittab
    fmc:2345:respawn:/opt/ibm/db2/V9.7/bin/db2fmcd #DB2 Fault Monitor Coordinator

    Any idea?
    ================================================== ================================================== ===========================================

    DB2 fault monitor comes into play only when there is abnormal termination from any user other than instance owner himself..

    for ex:

    If root user kills instance process then the fault monitor will activate the db2 instance

    try this

    [root@localhost home]# ps -ef|grep -i sysc
    db2inst1 18126 18124 0 06:48 ? 00:00:00 db2sysc 0
    root 18912 17626 0 06:57 pts/1 00:00:00 grep -i sysc

    above you can see db2sysc is running.. note the process id "18126"

    NOw Im trying to kill that process by root which is abnormal

    [root@localhost home]# kill -9 18126
    [root@localhost home]# ps -ef|grep -i sysc
    root 19368 17626 0 06:59 pts/1 00:00:00 grep -i sysc
    [root@localhost home]# ps -ef|grep -i sysc
    db2inst1 19382 19380 6 06:59 ? 00:00:00 db2sysc 0
    root 19398 17626 0 06:59 pts/1 00:00:00 grep -i sysc

    NOw if you see.. db2sysc got killed.. and agiain if i tried ps -ef, db2sysc started
    so If there is only abnormal termination fault monitor comes alive..


    If you stop the db2 instance by db2stop ... faultmonitor will become inactive..

    hope you got this..

  11. #11
    Join Date
    Sep 2013
    Posts
    1

    agreed

    Mr.Ksr i agree with ur answer..

  12. #12
    Join Date
    Oct 2012
    Posts
    21
    You could also run:

    db2nkill 0 as the instance owner. If the FMC is alive it should bring the service up. This will also cleanup all memory process running under the db2 engine.

  13. #13
    Join Date
    Sep 2013
    Posts
    2
    Quote Originally Posted by manawa View Post
    You could also run:

    db2nkill 0 as the instance owner. If the FMC is alive it should bring the service up. This will also cleanup all memory process running under the db2 engine.
    ================================================== ================================================== ====================

    We can also try db2_kill to kill the instance process..

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •