hi

in the diag file we can each entry has PID

incase the all the application is forced off and new connection comes in does the PID changes not sure not tried yet.

Regarding
SQL1224N error, one of the feature of this error is when ur shared memory is reached the limit db2 internally force off the connection (in which the user gets SQL1224N error) and it does not put any error in diag file and continues working as normal and taking new connection, it that possible.

regds
Paul

I dont remember to have observed. My response is theoretical

assuming unix,
In pre-9.5, it depends on which db2agent from the agent pool is used.


in 9.5 and above, PID will change only if the instance is restarted. tid will change depending on which agent is used.

use the commands

db2pd -edus
db2pd -agents

to understand.


have you tried diaglevel 4?

Logically..if a connection is completely forced off then a new connection should come with a new Pid.

Thanks
Naval K

Naval, If there is an avaialble agent in the pool, why will you get a new pid (and hence a new agent being created)? Cant the same agent be used? Can you pl clarify

Thanks

When a connection is forced completely means, related connection pool agent is also released. And if agent is also released means a new pid would be allocated to a new connection. Let me know if my understand is not correct.

Thanks
Naval K

In 9.7, after forcing the applications the db2agent (edu id 20) goes back to the idle pool as above. The second column is the thread-id

I would expect similar behavious in pre-9.5 also.

thks for the reply satya & naval
no i have not tried with diag 4,

1) Is is possible coz of shmmax hit db2 can internally force all connection and do a db2start by its own and not putting any info in diag file ?

2) few days back i was getting error continously about crash recovery there was no db2stop or start, again and again it was going into crash recovery this could be reason behind the internal restart of db2 by its own coz of memory issue, i do accept the fact the db is having a memory issue. pls suggest

regds
paul


regds
paul

Lets get down to the basics. What is the problem you are facing?

Anyway, let me answer you questions :
1) unlikely.
2) some application is trying to reconnect to the db repeatedly. Your crash recovery is failing every time, stopping and starting again on the next connection try.