Questions[tivoli](server)

We are using the Tivoli TSM Backup/Archive client to back up user files on a largish multi-user system. Users come from different countries and thus can create files according to multiple locales.

However, the Tivoli dmsc client insists that all file names are valid in the locale it was started in (e.g., root's locale) and does not backup files that are not valid in this locale:

Sending of object '.../x<B2>+y<B2>=z<B3>' failed
ANS4042E Object name '.../x<B2>+y<B2>=z<B3>' contains one or more unrecognised characters and is not valid.

I'm looking after either a config option that turns off file name validation, or a locale that makes any byte string valid. (E.g., ISO-8859-15 strings cannot contain bytes in the numerical range 128-159 which gives the error above.)

Is there a way to avoid this error and back up all files, regardless of the validity of their names in a locale?

TWS is Tivoli Workload Scheduler.

I have scheduled a job, and made it repeatable every 10 minutes. I put it in a job stream with a run cycle of daily.

Now, the job runs once every ten minutes on the first day, twice on the second day and so on.

What is the correct way to schedule a job in TWS to run many times every day?

Is there anything wrong with this job definition:

Workstation       Job Stream Name   Valid From  Updated On  Locked By
----------------  ----------------  ----------  ----------  ----------------
VMWSOAAPP01-TST   PDSSITRNS_INTG_H  -           01/17/2012  -

SCHEDULE MACHINE01-TST#JOBSTREAM
DESCRIPTION "SomeTransfer.sh  param1"
ON RUNCYCLE RC1 DESCRIPTION "Daily every hour" "FREQ=DAILY;INTERVAL=1"
KEYSCHED
:
MACHINE01-TST#JOBNAME
 SCRIPTNAME "/shared/app/app2/fds/bin/TWS/SomeTransfer.sh  param1"

 STREAMLOGON auser
 TASKTYPE UNIX
 RECOVERY STOP
 AT 0645 UNTIL 0545 +1 DAYS
 EVERY 0010
 KEYJOB
END

We have some data collection machines running WinXP. After a new file is written, we would like to immediately copy the new file to network storage (a NetApp CIFS share) automagically.

We need realtime or near realtime copies generated (copy upon filehandle close would be fine -- these are not long-running system logs). Two commercial applications I've found so far are MirrorFile and IBM's Tivoli CDP. Are there any reliable open source programs or simple ways to get Shadow Copy to do something similar?

Bonus points if it runs as a service.

I'm looking for a free solution that allows operators to keep an eye on systems and applications. It should digest log files from different sources and allow for alerts to be triggered in certain conditions.

I guess this is a trivial question for most of you, but I am less familiar with these types of applications.

Thank you! J.

We currently use IBM Tivoli to back up our Linux servers and we are looking for suggestions on the best way to restore to bare metal. I've read IBM's doc on this issue. Is that still relevant or is there a better way? Also, how do you handle testing a restore to a vm or another piece of hardware on the same network that the original server lives?