Ramesh & Fred...

That depends...UNIXes will buffer a process' writes to a file, typically with a 
4 kB buffer, unless the process calls a flush.  In ADSM's case, you may find 
that it won't write enough to the
schedlog during the time that it is backing it up to cause the OS to write the 
buffer, in which case the file does not change during backup and the backup is 
successful.

I bet that the system it fails on has a fair number of files being backed up 
before and after it reaches the schedlog.  Those systems will see the log 
change as the OS writes the buffer and,
assuming you are not using dynamic it will fail.

As with any text log, it is certainly valid to backup whatever you can get, so 
if you value the schedlog you may want to bind it to a 'textlog' mgmtclass that 
is dynamic.

Why would you want the schedlog?  It may be useful to know what was backed up 
last night when the disk (containing the schedlog and some other important 
stuff) dies this morning.

Cheers...
Bruce

Ramesh Razdan wrote:

> These two files will always fail as they are error log and schedule log which 
> adsm reports and updates when it doing a scheduled process. U can  ask 
> dsm.sys to specifically to skip the two files
> thx
> razdan
>
> >>> Fred Johanson <fred AT MIDWAY.UCHICAGO DOT EDU> 05/12/99 02:30PM >>>
> A user backs up about a dozen Sun boxes a night.  Every one reports in the
> SCHEDLOG that it failed to backup the ERRORLOG and the SCHEDLOG.  The
> schedule for one of the machines is regularly marked as FAILED, all the
> rest are marked as COMPLETED.  Our suspicions are either OS level (it's
> old) or length of time on port.  It's just an annoyance, but if anyone has
> any thoughts...
> Fred Johanson
> System Administrator, ADSM
> S.E.A.
> University of Chicago
> 773-702-8464

--
Bruce Elrick, Ph.D.
Bruce Elrick, Ph.D.
mailto:belrick AT home DOT com
http://members.home.net/belrick/