Problem with TSM Scheduler on AIX clients 1

[Guest_imported]

Hi every body !!!
I have problem with schedulers on my rs/6000 clients....They are are declared as node on the server TSM but the backup planed are missed if ever i don't kill the dsmc process on the clients. I need to kill the process to generate a new one; cause i have add an entry in the inittab file
tsm:2:respawn:/usr/tivoli/tsm/client/ba/bin/dsmc sched >/dev/null 2>&1 #Start the TSM Scheduler

Is there a way to fix the problem instead of killing the process (ps -ef|grep dsmc and the kill the appropriate ps)

 

[gwichman]

You should never set the TSM client to respawn in inittab. It should be "once". If the dsmc sched process dies you don't want it automatically respawning. You want to troubleshoot why it's dying by examining the appropriate error logs because they don't typically die for no reason. i.e. it means you have a problem!

Other then that i don't know what you mean by "They are are declared as node on the server TSM but the backup planed are missed if ever i don't kill the dsmc process on the clients."

The backups scheduled miss if you don't kill the dsmc process? That doesn't make any sense. You'll have to elaborate.

 

[abdoul]

the backup failed if the dsmc sched process is not regenerated. It seems that if I start the scheduler it hangs after a certain time. So The planed backup failed ...But I I refresh the scheduler process by killing the existing (another will be generated)..it works ....
I kill the sched process by ps -ef|grep sched , and kill -9 PID

 

[gwichman]

Then you have a problem that needs troubleshooting. Killing and spawning a new dsmc sched process isn't the solution. Check your dsmsched.log and dsmerror.log files on the client for indications as to what exactly it's hanging on or what errors are coming up. What do you mean by "it hangs after a certain time"? You should easily be able to see where it's hanging by looking at dsmsched.log when it's hung. dsmc logs everything it does in dsmsched.log with timestamps. Also check dsmerror.log around the same timestamp to see if theres anything of note there. Also check the TSM server activity log around the same time to see if anything gets logged regarding the problem. If you want to send me any of the logs feel free and i'll take a look at them.

What OS and version is your TSM server and the client in question (e.g. TSM server is TSM 4.1.1 on AIX 4.3.3, client is TSM 4.1.1 on AIX 4.3.2)?

 

[tskimiz]

Hi~! tsmfax..

If you'r schedule missing, you must check association.

issue command in tsm console or terminal "q association"

And check association and Associated Nodes.

If there is blank or null, You must define association with node name like this,
"define association DOMAIN-NAME SCHEDULE-NAME NODENAME"

Good luck, From Korea, Republic of.

 

[adamsd]

Hello,

I am having a some what similar problem. The shedular process does not startup at system reboot. I have looged this with Tivoli support, but they are still trying to find the problem. I can start the process from the prompt no problem.

Cheers
Dexter

 

[gwichman]

Dexter..

you'll have to provide more info.. what platform and what mechanism are you using to automatically start the scheduler (/etc/inittab, rc script, windows service, ?)?

Whatever mechanism you're using it would be helpful to see the relevent entry/file where you've got it to start up.

 

[adamsd]

I have the entry in the /etc/inittab file. We are running AIX5.1 on a p670 (LPAR). These configuraton files was copied accross from an AIX4.3 TSM server by someone else.
Below is an extract from my inittab file:

autosrvr:2nce:/usr/tivoli/tsm/server/bin/rc.adsmserv /dev/console 2>&1 #Start the Tivoli Storage Manager server
tsm:2nce:/usr/bin/dsmc sched > /dev/null 2&>1 #TSM schedular
dsmcad:2nce:/usr/tivoli/tsm/client/ba/bin/dsmcad /dev/null 2&>1 #ADSM web client

Cheers
Dexter

 

[gwichman]

can you start the process by simply cutting/pasting what you put above at the cmdline? i.e.:

/usr/bin/dsmc sched

seems correct at first glance.. provided dsmc does indeed reside in /usr/bin (or a link to it). The next question would be whether or not you have password access set to generate in your dsm.sys file? i.e. if running the above command without the >/dev/null 2>&1, it should get you to the tsm> prompt without asking for a username or password. If it asks for a username or password, thats why it isn't starting automatically. If indeed that works properly, the next thing i'd check is making sure it gets called at all from /etc/inittab.. check dsmsched.log and see if there are any entries in it at the time of reboot. It should log *something* or coredump (any core files on the server after reboot?). check errpt for and errors around the time of reboot (a coredump would show up there). i.e. there should be SOME indication it's at least being run.

 

[itr0754]

gwichman, I am having the exact problem you mention. When I run:

dsmc sched

I get a prompt for a userid and password. Thus, my scheduler isn't starting automatically. So, where do I declare the password ?? I am new to Tivoli so please excuse the newbie status.

 

[amphibient]

the userid and password were set up when you registered the node on the server, correct?

 

[itr0754]

Yes, I got confused by the fact that I was running AIX but, actually, it works the same way as W2K. I registered the node on the TSM server and then put in the password when I ran the configuration wizard on the client. Problem solved !!