DB2 AIX WSE V9.5 FP2a

Database was not brought down and our AIX team member has rebooted the server. The server had two instances, after the server came up. We started instances. One of them started and another wouldn't start.

After issuing db2start, no error is seen in diag.log for 10 mins and after 10 mins, this error comes out

2010-07-01-16.15.02.343257-240 I300354A347 LEVEL: Warning
PID : 360612 TID : 1 PROC : db2bp
INSTANCE: teiemsnz NODE : 000
EDUID : 1
FUNCTION: DB2 UDB, Connection Manager, sqleUCappImpConnect, probe:150
MESSAGE : ZRC=0x81360012=-2127167470=SQLZ_RC_CMERR, SQLT_SQLJC
"External Comm error"



Any help would be appreciated.

I tried doing ps/kill, ipcs/ipcrm, ipclean, db2nkill{as it is wse}, db2ftok [create new file] and even then instance wouldn't start. When I do ipcs -a | grep -i [instname] . I see these leftover messages which wouldn't go with ipcrm. Are those the one creating problem



m 1048576 0xffffffff D-rw-rw-rw- teiemsnz db2admqt teiemsnz db2admqt 1 140848008 577566 577566 14:41:11 14:50:46 14:41:35
m 3 0xffffffff D-rw------- teiemsnz db2admqt teiemsnz db2admqt 1 72548352 577566 577566 14:38:22 14:50:46 14:41:35
m 4 0xffffffff D-rw------- db2fenc1 db2fadm1 teiemsnz db2admqt 1 268435456 553026 553026 14:38:22 14:41:35 14:41:35

Finally it started.

I did db2stop on the other instance of that server and killed all the processes of that instance and cleared all the db2 processes running on the box. After doing that, the above three messages disappeared and db2 started sucessfully. Not sure what caused it not to start.

Those three shared memory segments were defunct (D in the first position under mode column). From "man ipcs":

D
If the associated shared memory segment has been removed. It disappears when the last process attached to the segment
detaches it.