Error writing to named pipe

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
cyclogenisis
Premium Member
Premium Member
Posts: 48
Joined: Wed Jan 07, 2015 3:30 pm

Error writing to named pipe

Post by cyclogenisis »

I am getting this error message at ~ 4M rows on the target Netezza connectors stage:

Code: Select all

TargetNetezaStage: Failed to write to named pipe /tmp/server02_49736_nzw_0_1_20160310041042104013. Error code: 32 (CC_NZPipe::write, file CC_NZUnixPipe.cpp, line 321)
The setup is

Code: Select all

Netezza (source) --> Netezza (Target)
Anyone have advice or solution? It is stopping our jobs from loading target table.
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

Typically that would mean that /tmp filled up.
-craig

"You can never have too many knives" -- Logan Nine Fingers
cyclogenisis
Premium Member
Premium Member
Posts: 48
Joined: Wed Jan 07, 2015 3:30 pm

Post by cyclogenisis »

I've tried to set it to other directories in thr named pipe location option which should have space but same error. Any other suggestions?

By the way what is the named pipe used for?
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

Should have enough space != does have enough space. Necessarily. Monitor the space during the load to be certain. Then we can move on to other possible causes.

The named pipe is used to transport the data through the ETL server.
-craig

"You can never have too many knives" -- Logan Nine Fingers
cyclogenisis
Premium Member
Premium Member
Posts: 48
Joined: Wed Jan 07, 2015 3:30 pm

Post by cyclogenisis »

What's weird is that I didn't see this file increasing in size while the job was running, so I'm wondering if data is transferred into named pipe in one large batch...
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

There's a number of factors since many UNIX processes can leverage the /tmp space. If you've moved it elsewhere and still have the issue you could be seeing a problem with the connector itself. Unfortunately all the "32" means is "broken pipe" which doesn't really tell us anything and several issues could cause the pipe to break. Probably best to involve support on this one.
-craig

"You can never have too many knives" -- Logan Nine Fingers
Post Reply