To Start & Stop a TWS FTA use the below commands: 1. Start a TWS FTA:
./<TWS_home>/StartUp
./<TWS_home>/bin/conman "start & link @;noask"
2. Stop a TWS FTA:
./<TWS_home>/bin/conman "shutdown;wait"
And a small script to start / stop TWS FTA:
#!/bin/sh
#
# Start/Stop the Tivoli Workload Scheduler FTA
#
case "$1" in
start)
/<TWS_home>/StartUp
/<TWS_home>/bin/conman "start & link @;noask"
;;
stop)
/<TWS_home>/bin/conman "shutdown;wait"
;;
*)
echo "Unknown flag passed to $0."
echo "Usage: $0 {start, stop}"
esac
exit 0
Note: The above script is the simplest possible, I like simple things, even we have a error as it will try to start / stop all the agents but it does not have rights, which is normal.
The easiest way to migrate jobs / job streams between the agents without any impact and complete invisible for end users is to use composer commands as below:
1. Create a dump containing all the jobs & the job streams from the source agent (S_FTA):
2. Rename the source (S_FTA) agent with destination agent (D_FTA) and save the files using a different name (D_FTA_all_jobstreams.txt & D_FTA_all_jobs.txt) so you can restore it in case is needed (you can use any text converter you want).
3. Remove the dependencies using temporary files (if the dependencies exist when importing the job streams we will get an error message and the job stream is not imported, and we need to import them again and again until all the dependencies are met, or it will not work at all if we have a close cycle):
7. Delete the job and after, the job streams from source agent.
The jobs will start to run on the destination agent (D_FTA) starting with the new plan.
Note: If the jobs are scheduled using time zones it may take 48 hours (2 default plans) for a complete move.
To restore the Symphony files when it gets corrupted on any of FTA agent perform the following using the TWS_user:
Method a. (manual restore):
On the FTA with corrupted Symphony: 1. Change the limit to 0, to prevent any jobs from beeing launched once the new Symphony file is created:
./<TWS_home>/bin/conman "lc;0"
2. Stop all TWS process on the FTA with corrupted Symphony:
./<TWS_home>/bin/conman "shutdown;wait"
3. Delete / Rename the following files:
Symphony, Sinfonia, Jobtable, <TWS_home>/*.msg and <TWS_home>/pobox/*.msg
On the TWS MDM: 4. Unlink all the agents:
./<TWS_home>/bin/conman "unlink @;noask"
5. Stop the TWS MDM agent:
./<TWS_home>/bin/conman "shutdown;wait"
6. Rename Sinfonia file and copy the Symphony file over the Sinfonia.
7. Start the TWS MDM agent:
./<TWS_home>/StartUp
8. Link all the agents:
./<TWS_home>/bin/conman "link @;noask"
On the FTA with corrupted Symphony:
9. Start the agent:
./<TWS_home>/StartUp
On the TWS MDM: 10. Start & Link the FTA with corrupted Symphony:
./<TWS_home>/bin/conman "start FTA_name & link FTA_name;noask"
At this stage the new Symphony file should be deployed to the FTA with corrupted Symphony, it will be exactly how it was before the corruption occurred, also as the *msg files were remove / delete some jobs / job stream information may be lost, verify the jobs / job streams status to make sure none of them will be ran 2 times.
11. Change the limit to the original value:
./<TWS_home>/bin/conman "lc;100"
Method b. (semi-automated restore): There a script utility developed by IBM for the corrupted Symphony file that can be found HERE . Hint: If there are multiple agents with corrupted Symphony, repeat the above process for each one.