|
View previous topic :: View next topic
|
| Author |
Message |
Pandora-Box
Global Moderator
.jpg)
Joined: 07 Sep 2006
Posts: 1592
Location: Andromeda Galaxy
|
|
|
|
Hi,
I have a transaction id TST1 which calls program A and Program A is main module which connects between Java and Mainframe.
Program A in turn links 30 different CICS modules
Now I have a situation where DB2MSTR shows up something like
CORRELATION-ID=ENTRTST10005 deadlocked with
CORRELATION-ID=ENTRTST10025
Does the last 4 digit number in Correlation id has any significant meaning ?
How can we figure out which 2 programs were contending for resource?
Note : I felt this query is closely related to CICS so posted here |
|
| Back to top |
|
 |
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hello,
Your situation is caused by a design problem using DB2. While this is run in CICS, it it a DB2 situation that needs corrected (i believe).
As with any deadlock, multiple processes are trying to "lock" the same data but not in the same order.
When i looked in the db2 doc just now, CORRELATION-ID is involved with MQ. |
|
| Back to top |
|
|
Pandora-Box
Global Moderator
Joined: 07 Sep 2006
Posts: 1592
Location: Andromeda Galaxy
|
|
|
|
Hi Dick,
Thanks for the response but here this program is not involved with MQ
So will increasing the LOCKMAX number of Table space help to reduce the deadlocks ? |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hi PB,
No, i don't think so. It only takes 2 tasks to generate a deadlock situation. Upping the number of possible locks would most likely not help this deadlock.
About MQ - i too was surprised to see those entries in the DB2 manual . . .  |
|
| Back to top |
|
|
Jose Mateo
Active User
Joined: 29 Oct 2010
Posts: 121
Location: Puerto Rico
|
|
|
|
Good afternoon, Pandora!
Have you scanned the MVS log for more information on this problem? |
|
| Back to top |
|
|
agkshirsagar
Active Member
Joined: 27 Feb 2007
Posts: 691
Location: Earth
|
|
|
|
Please check this link to find more information on this topic.
As far as your first question goes, the number is to uniqely identify the threads. Looks like the correlations are two instances of transaction TST1. |
|
| Back to top |
|
|
colin777
New User
Joined: 06 Jun 2013
Posts: 19
Location: Singapore
|
|
|
|
Hi,
These Correlation Ids :
CORRELATION-ID=ENTRTST10005 deadlocked with
CORRELATION-ID=ENTRTST10025
are first 4 chars is the type of Thread, ENTR is Entry Thread. Then next 4 chars is the Transid, then followed by task number of the task.
But please note, this is only valid for knowing the Transid/Task number, if this transaction did the signon to DB2, meaning if the Thread was reused it would be the Transaction Id, that used the thread first, that did the signon. Only the Attachment Facility can change the Correlation Id, by another Transaction issuing signon.
HTH
Colin777 |
|
| Back to top |
|
|
Pandora-Box
Global Moderator
Joined: 07 Sep 2006
Posts: 1592
Location: Andromeda Galaxy
|
|
|
|
Thanks all for your replies
As of now we have increased the LOCKMAX and this has almost eliminated the deadlocks for past few days |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hello,
Good to hear this helped, but i believe the process(es) need to be revised to eliminate the deadlocks. . . |
|
| Back to top |
|
|
|
|
