DSXchange

After a server reboot, our environment script wasn't run correctly, so we needed to take down DataStage, run the env script, then bring it back up.

When we tried to shut down DataStage, we had the following errors:

$ uv -admin -stop
Waiting for segment 18 to be destroyed...
Unable to destroy segment 18.
1 error(s) encountered during shutdown procedure.
DataStage Engine may be in an inconsistent state.

Now we can't bring DataStage back up. This is our production environment. I saw some similar threads dealing with 6.x and 7.x, but nothing with 5.x unirpc. Thanks!

Make sure you have root or super-user privileges, then rerun the shutdown script; i.e. uv -admin -stop. This should bring all the engine components down.

The message you got earlier, complaining about a segment that cannot be removed, is due to a piece of code in the shutdown script that does not evaluate engine segments correctly and inadvertantly may contain a segment number NOT associated with the engine. Because it is not that particular segments owner, it cannot be killed, hence the message. Note that this error has been hunted down, identified and corrected since. Upgrading would alleviate this issue.

Hope this helps.

Oddly enough, we are upgrading to 7.1r1 tomorrow

I ran the uv -admin -stop about 5 more times with no success. I finally got a hold of someone from Ascential support, but I'm nervous because they keep telling me to run "dsrpcd" in debug mode, but 5.x has "unirpcd".

Anyway, I'm going to continue to do what they tell me, but I'll post the outcome on here in case it will benefit anyone.

Thanks so much!

Perhaps some dsrpc or unirpc sessions are still open.

Stop DataStage, and wait for the to time out, then restart DataStage:

bin/uv -admin -stop

netstat -a | grep dsrpc {-or - netstat -a | grep unirpc}

If the netstat is clear, then you can restart DataStage

bin/uv -admin -start

Chances are, that you will see dsrpc session in the FINWAIT state.

And do a "ps -ef | grep phantom". If you still have some when DataStage has been shutdown, kill them. Or you may wait a very long time on the "nestat -a" to clear.

Ogmios

It turns out a recent AIX patch we applied messed up the ipcrm utility, so that's what was causing the problem.

Now on a different box, we're encountering a similiar issue. The uv -admin -stop works fine, but then when we do a uv -admin -start it says that ../DataStage/unirpcshared/unirpc/unirpcd cannot be found. This is a new instance, so I'm wondering if the installation was messed up or something? Any thoughts?