DSXchange

I am trying to revise our Datastage file system backup strategy. To ensure a clean backup we are following the steps below. Unforuntely when a client connection exists when stopping dstage process they are put into a "CLOSE_WAIT" state. When restarting dstage after backup the dsrpc process does not fully come up and jobs cannot run. The "CLOSE_WAIT connections should timeout after 15 minutes but some do not. We also would like to add a test into the process to see if any jobs are running before the process is started. Please provide feedback!!

How can you tell a client connection versus a job running?

Current backup process:
1) stop dsrpc
2) backup file system
3) wait 15 minutes for closed_wait connections to timeout
4) start dsrpc

Proposed backup process:
1) check if any jobs are running
- if so then go back to sleep for 15 minutes. will try 3 times
- if not proceed to step 3
2) kill all client connections to dstage
3) stop dsrpc
4) backup file system
5) start dsrpc

The running job should be a phantom.

ps -ef | grep phantom

Clients should be a dsapi_slave from memory.

ps -ef | grep slave

Don't wait "15 minutes", wait for the ports that are in use to clear. You can use 'netstat -a' and grep for 'dsrpc'. Once that returns zero results then DataStage can be restarted.