Engine status property of uv -admin -info

Ultramundane · Post by **Ultramundane** » Tue Oct 24, 2006 1:45 pm

$ uv -admin -info | egrep -i 'Engine status'
Engine status : Running

How is the Engine status derived? Reason I ask is because part of failover testing we performed crash of the server and then datastage would not start. Running a uv -admin -start said that the instance started! However, when I checked for the dsrpcd process and for the listener neither was started. Yet the command send it started successfully.

I ran the above command and it said that the Engine was running. It was not. I ran uv -admin -stop and this cleared this status. I was able to perform a successful startup after that.

So, I have two problems.
1. Seems uv -admin -info does not accurately know whenever the engine is running or not.
2. The startup indicating a successful startup when nothing was started.

I can code around these issues, but I think that IBM should fix this.

Back to my question.

How is the Engine status derived as it does not seem to be related to the state of the dsrpcd process or the listener?

ray.wurlod · Post by **ray.wurlod** » Tue Oct 24, 2006 1:52 pm

Engine status is obtained from the existence of the DataStage disk shared memory segment. This has a shared memory key of 0xadecnnnn where nnnn is the DataStage version number, for example 0xadec7511 where I'm currently working. The "ade" may be different on an -itag installation. You can use the ipcs command to determine what shared memory segments are in use; on most UNIX systems this can be restricted to shared memory segments by the -m option. On Windows servers use the shrdump command that ships with DataStage.

Ultramundane · Post by **Ultramundane** » Tue Oct 24, 2006 8:30 pm

Thank you for this information!