Project currently locked

[ArndW]

The Search can be used to find the many threads in DSXChange that discuss this. I think that the documentation is very weak, the only mention is in the BASIC guide,

Deadlocks Deadlocks occur when two users who acquire locks incrementally try to acquire a lock that the other user owns, and the existing lock is incompatible with the requested lock. The following situations can lead to deadlocks: v Lock promotion from a shared record or shared file lock to a stronger lock v Lock escalation to file locks when two users try to escalate at the same timeYou can configure WebSphere DataStage to automatically identify and resolve deadlocks as they occur through the Deadlock Daemon Administration menu, or you can manually fix a deadlock by selecting and aborting one of the deadlocked user processes. You use the deadlock daemon dsdlockd to identify and resolve deadlocks. For more information, see Administering WebSphere DataStage.

(the referenced manual doesn't exist at V8 any longer)

[Brian Zhang]

The Search can be used to find the many threads in DSXChange that discuss this. I think that the documentation is very weak, the only mention is in the BASIC guide,
(the referenced manual doesn't exist at V8 any longer)

Thank you very much ArndW.

I'm new to this forum, and hope I could gain more knowledge here.

Thanks again!

:D

[ArndW]

I just noticed Ray mentioned how to start the daemon, you will also need to edit the $DHSOME/dsdlockd.config text file and enable the daemon.

[Brian Zhang]

I just noticed Ray mentioned how to start the daemon, you will also need to edit the $DHSOME/dsdlockd.config text file and enable the daemon. ...

ArndW,

I open this configuration file, and see the below default parameter setting,

start=0
timer=900
res=0
log=

Could you elaborate more details on these parameters?
Thanks in advance!

Brian

[Brian Zhang]

I just noticed Ray mentioned how to start the daemon, you will also need to edit the $DHSOME/dsdlockd.config text file and enable the daemon. ...

ArndW,

I open this configuration file, and see the below default parameter setting,

start=0
timer=900
res=0
log=

Could you elaborate more details on these parameters?
Thanks in advance!

Brian

[ArndW]

Set "start=1". It will start automagically next time you bounce the DataStage server, or use Ray's method at the beginning of the thread to start it immediately.

[Brian Zhang]

Set "start=1". It will start automagically next time you bounce the DataStage server, or use Ray's method at the beginning of the thread to start it immediately. ...

Thank you.

And the "log" parameter will indicate the path and log file name which will be used to store log message from Deadlock Daemon process, right?

And how about "res" parameter?

[Brian Zhang]

I noticed that there is a situation depicted in the manual,

Lock escalation to file locks when two users try to escalate at the same time

So, this could be happened on the # of access to certain file.
If so, it might have the relationship with the parameter MFILE that I mentioned above, right?
Because we tried to enlarge the number of the parameter MFILE to see what will happen, and after that, the issue somehow was solved.

Any comments on this?

Thanks.
Brian

[Brian Zhang]

I got the list of option from dsdlockd command, and think it could be referenced for configuration to dsdlockd.config.

dsdlockd [options]
------------------
-t timer interval timer to have daemon check for deadlocks
-r res resolution to use(0=Random;1=Youngest;2=Fewest Locks)
-l log location of the log file
-query requests a report of current status
-victim sig identifies the signature of the victim to be aborted
-p performs a check on dead process cleanup
-stop shutdown the daemon
NOTE: The -query, -victim, and -stop options cannot be used in conjunction with the other options