DSXchange

Hi,

We have just started receiving warnings messages of the format below in jobs that have been running fine for a number of months. The jobs complete OK but issue these warnings:

Unable to read Temp_<jobname> from OMD.XML.DIR
Unable to write Temp_<jobname> from OMD.XML.DIR

I have tried searching for this file / directory in the /opt/ and other locations but to no avail.

Has anyone else encountered this warning and if so, what are the causes / resolutions?

Many Thanks,

Daren

I haven't seen messages of this type, could you perhaps post the full log message? Is this for all jobs or certain jobs?

That is the full message from the director log, replacing jobname with the job I am running. There is no other detail.

Daren

do you have a basic transformer in your job?

I have a Parallel Transformer, which is probably the same, yes.

I have a Parallel Transformer, which is probably the same, yes.

No, Parallel and BASIC transform stages are very different in how they get executed. I wonder if this isn't a message from the repository. (quick login to our XMETA) There is no table with that name, but several starting with OMDMODEL... so it is likely a repository error after all.

I have seen this warning only when there was a basic transformer in the job (mostly calling some server routines). So It will be better if you check it once again to confirm.

I don't remember the exact name of the environment variable but I think its DSIPC_TIMEOUT or something like that. Try increasing that from 30 to 600. But I would prefer to replace the basic transformer with Parallel one.

Sorry about the delay in replying.

The transformer is a parallel transformer. We do not use basic transformers in our PX jobs.

I have amended DSIPC_OPEN_TIMEOUT as suggested above to see if that resolves the error in the meantime.

Thanks for the responses so far

Daren

The errors in the PX jobs have not re-appeared. However, we created some new reporting jobs which use a Server shared container and basic transformer and this message is now appearing for that job:
If it is a known issue with a basic transformer, what is the resolution?

Many Thanks,

Daren

If DSIPC_OPEN_TIMEOUT setting change made the job to run without problem then there seems to be a basic transformer in that job.

This is because Inter Process Communication is now way related to parallel engine unless it comes to Basic transformer. Because it uses IPC to communicate between processes.

I won't suggest to increase it further because sometime jobs may take exeptionally high time to finish. If you can play with the settings I would suggest to change(increase) all the settings related to IPC (don't set it too high else you may find other problems)

Basic transformer creates some intermittent problems hence I don't prefer to use it (especially encountered with 8.0). I think the official support provider is your best bet at the moment.

If its not creating any problem with data then I think you can also go for message handler (IMO)

Thanks for the replies guys. We have logged a call with our support.

Daren