DSXchange

In my DataStage environment under /tmp is px.connect.fail#### files

I see them on few of my remote servers where datastage scratch disk is mounted (nfs) .

I am assuming some connection fail while accessing scratch disk and this is generating these logs..

Did any one see such files in your servers ?
if yes please shed some light .


Thanks

Don't have access to check so need to ask whether you have any content in those files or they are just 0 byte files?

these logs have info about
hostname , ports ,dsd.run and connections established and wait state

So basically it has lot of information about the server.


After lot of research I figured when a datastage job is initiated and failed abruptly then these logs are generated.


If you have more info pls share.

I think you have your answer. And even in case of abrupt failure, the job related information is saved in &PH& folder in project directory. These files only generated when there is a connector stage in your job which failed abruptly, IMO.

I have tried with a job having a connector stage and by killing the job in the middle of execution but could not generate these files .

I think the failure has to be from the database end.

I am seeing these files on compute nodes (server2)also ..
where my scratch disk is located . ..

And my engine is located on conductor node server 1.

Yes, I believe they're written there when the connector detects a failure in what it's connected to.

is this connector failure have to do with kernel limits ?

Almost certainly not. Problems with kernel size will usually be picked up at installation time.