DSXchange

Lately our jobs with Aggregator stage started failing with Below errors:
"sendWriteSignal() failed on node <Server name > ds = 1 conspart = 0 Broken pipe"
"Fatal Error: Unable to allocate communication resources"
"Premature EOF on node <Server> Socket operation on non-socket"
Only jobs with aggregator are failing. I checked Scratch disk, tmp space. everything looks good. Looks like a space issue, But not sure which space is full. Please advise.

Thanks a lot in advance..

Did you monitor the disk space while the job is running?
Not much point checking afterwards.

Also try running it with a subset of the data. Does it run successfully?

Why do you think "unable to allocation communication resources" is a disk space issue?

Does your configuration file mention more than one fastname value?
What stage types are upstream and downstream of the Aggregator stage? What operators are upstream and downstream of the group operator in the score?
Are you using Sort method or Hash method in the Aggregator? How many distinct groups are there/will there be?