DSXchange

Hi All,

Can anyone help me with this issue?
I am trying to FTP a file from DataStage Unix server to Mainframe using FTP Plugin stage in 7.5.1 DataStage.

I am getting the following error.

Error while trying to open connection for sending data to the FTP server, socket call 'connect' failed

More over this error pops up only some times.
Sometimes the FTP works fine.
FTP to any other Unix server is working fine though.
Can any one help me please?

Thanks
Ranjani

Intermittent failures... blah.

Analyze your environment and understand it. Go over what COULD change between executions.

LDAP authentication is often load balanced across a few boxes. Are you getting intermittent authentication problems?

Did you remember to turn on verbose error messages?

Try writing a script on the command line to trap messages if the stage is not handy enough for you.

Is your datastage environment a grid, where your side is load balanced across different servers?

The Mainframe itself is a load balanced environment also. The LPAR you are farmed off onto once the connection is attempted is dynamic I believe. Is your ID properly set up on all of those LPARs?

I'm not a strong mainframer so that's guesswork on my part.

Hi

Thankyou very much for the response.

We have tried to change the 'FTP Connection mode' in FTP plug in stage to 'Active' from 'Passive' mode. It is working fine as of now.Testing the same.

But not sure what that means.

Not to hijack the thread topic, but I'd recommend staying away from the FTP stage in DataStage. Due to its transfer methodology it is very slow compared to an external ftp command.

You'll see a significant performance improvement if you use an external ftp command via a script. The script can be called in either a Before Job or After Job ExecSh. It can also be run from an Execute Command object in a Job Sequence.

The change I have mentioned didn't work.
The job failed again..

I would still recommend using a separate FTP script instead of trying to use the stage. However, if you are getting occasional FTP connect failures it is probably because the mainframe has a limit on the number of active FTP sessions and it isn't allowing the connection at that time.

If that is the case, then you'd need to run the job in a sequence loop that traps the error, sleeps for a few minutes, and tries again.

Hi Paul/Andy...

Is there any way that we can see the error code of the FTP process in UNIX where the Datastage resides..

As the 'Socket call failure' is not clear enogh to find out the reason for failure

The error msg in director is not clear.

Thanks
Ranjani

Hi Andy,

The FTP plugin stage has the retry interval of 15 seconds now and i have changed it to 60 seconds and the retry attempts from 2 to 5.

lets see if that works.

Thanks,
Ranjani

Mr G, if you're environment is grid or at least a cluster, don't do the FTP from the head node please. Use an external source stage, drop it down to sequential mode. That farms off your script to your compute node. Limit it to your first defined node.

Try to minic your FTP outside datastage. Determine where the problem resides.

You'll have to rope in your Mainframe brothers to help out on that side.