DB2v9.7 FP1 installfixpack problem

[Former Member]

Hello all,

we've got troubles with installation of fixpack on DB2 9.7 on SuSE 11

- installfixpack done successfully;

- when I call db2start I receive

06/21/2010 15:05:06 0 0 SQL1042C An unexpected system error occurred.

SQL1032N No start database manager command was issued. SQLSTATE=57019

so, we can't start db6_update_db.sh (application is SAP)

files in /db2/db2sid/sqllib/adm are not updated and are differ from /db2/db2sid/db2_software/adm

Regards,

Konstantin

[Former Member]

Yes, of course

1. ps -eaf |grep db2

root 7039 3401 0 10:21 pts/0 00:00:00 grep db2

Regards,

Konstantin

[Former Member]

Hello Konstantin,

any errors withing the db2diag.log.

DBI1122E states that there are Db2 processes running. Whaht about db2as (Administration Server)?

Deprecated with DB2 V9.7

Best regards,

Joachim

[paul_power]

Hi Konstantin,

Lets focus on seeing if something related to the db is remaining active since the database appears to be cataloged fine.

Apologies if you have done some of these already but lets be thorough to make sure we find whats blocking the instance update:

can you check if any DB2-Process is running ?

ps -ef grep db2

Maybe, there is some DB2-related servive started as system startup. Can

you check your /etc/inittab ?

If this is the problem, it should help to comment out the related line

in the /etc/inittab and reboot the box.

To stop all DB2 processes, perform the following steps:

Log on as root.

Determine which instances are associated with the DB2 copy. Issue the

command:

DB2DIR/instance/db2ilist

where DB2DIR represents the location where the DB2 copy is installed.

The default installation path for UNIX® is /opt/IBM/db2/V9.5.

Run the following commands for each instance in the DB2 copy:

su - db2

. $HOME/sqllib/db2profile

db2 force applications all

db2 terminate

db2stop

db2licd -end # run at each physical partition

exit

If you are an HACMP# user, you must use the ha_db2stop command to stop

DB2 instead of the db2stop command. If you use the db2stop command

instead of the ha_db2stop command, you will trigger a failure event.

If the DB2 Administration Server (DAS) belongs to the DB2 copy that you

are updating, stop the DAS:

su - aname

. $HOME/das/dasprofile

db2admin stop

exit

where aname represents the DAS owner name.

Note: Since there can only be one DAS on the system, this step affects

all other DB2 copies on the system.

Ensure all DB2 interprocess communications are cleaned for the instance

to be updated. As the instance owner, run the following command at each

physical partition:

$HOME/sqllib/bin/ipclean

Check for ipcs processes by running ipcs grep db2

Run slibclean to unload unused shared libraries from memory before

installation:

/usr/sbin/slibclean

Disable the fault monitor processes. To stop the Fault Monitor Daemon,

issue the command:

DB2DIR/bin/db2fm -i iname -D

where DB2DIR is the location where the DB2 copy is installed and iname

represents the instance owner name. The command must be performed once

for each instance in the DB2 copy.

If the Fault Monitor Coordinator (FMC) is enabled, prevent your

instances from auto-starting:

To determine whether the FMC is enabled, issue the command:

DB2DIR/bin/db2fmcu

where DB2DIR is the location where the DB2 copy is installed. If the FMC

is enabled, you will see output similar to the following:FMC: up: PID =

3415 . If the FMC is disabled, the output from the db2fmcu command will

be: FMC: down.

If the FMC is enabled, determine whether any instances are configured to

auto-start after each system restart. Issue the command:

DB2DIR/instance/db2iset -i iname -all

where DB2DIR is the location where the DB2 copy is installed and iname

represents the instance owner name. The command must be performed once

for each instance in the DB2 copy. If the output from the db2set command

includes the following, it means that the instance is configured to

auto-start:DB2AUTOSTART=YES

Prevent the instances from auto-starting. Issue the command:

DB2DIR/instance/db2iauto -off iname

where DB2DIR is the location where the DB2 copy is installed and iname

represents the instance owner name. If desired, re-enable instance

auto-start after you have completed the fix pack installation:

DB2DIR/instance/db2iauto -on iname

If there are any db2 related ipcs resources these need to be removed

manualy with ipcrm command.

If all the above checks out then we can be sure that the fail isnt due to something related to the instance still being active.

If you still encounter problems after running the commands above you

might want to install the image into a new path, then use db2iupdt to

update the instance to the new code level. Another option might be to

reboot the AIX server and restart the fixpak install.

Regards,

Paul

[Frank-Martin]

Hello Konstantin,

DB2 may not have been stopped cleanly. Maybe there are some leftover IPC segments. Try "ipcs | grep db2" . If those exist you can try to remove them using "db2_kill" or manuel "ipcrm" . An alternative would be to restart your DB server.

After this you can try "db2start" again.

Regards

Frank