I had this kind of problem a couple of times before. Sometimes the instance does not start up and the messages in the db2diag.log are related to one of the follwing:

- apparently symbolic links are broken;

- shared libraries are not loaded by operating system;

- permissions problems on DB2 executable or configuration files.

Errors examples:

1 - Command line error

SQL1042C An unexpected system error occured
SQL1032N No start of database manager command was issued. SQLSTATE=57019

2 - db2diag.log error

2008-05-02-01.54.24.134147-300 E19379A322 LEVEL: Error (OS)
PID : 512304 TID : 1 PROC : db2sysc 0
INSTANCE: hrprd88 NODE : 000
FUNCTION: DB2 UDB, oper system services, sqloRunInstance, probe:275
CALLED : OS, -, unspecified_system_function
OSERR : EPERM (1) "Not owner"

Based on these messages I´d say that DB2 installation is messed up. So, what the hell to do now?

Before re-installing DB2 or call DB2 support, I suggest to run (or re-run) db2iupdt command. This command is supposed to be executed when you are upgrading DB2 fix packs, but has solved my problems in the past.

These are the steps to run this command:

1) Stop all DB2 related processes. Make sure

ps -ef | grep db2

does not return any line;

2) as root, run the following command:

$DB2INSTALLATION_DIR/instance/db2iupdt <instname>

or

$DB2INSTALLATION_DIR/instance/db2iupdt -e

(to apply this command to all instances on the server)

Seems that this command, among other things, recreate all symbolic links and reset the permissions to all files related to DB2. The point here is that you must to stop all db2 processes before running this command.