Geoffrey,

The "switch" is this...

The Unix clients include processing option "ARCHSYMLinkasfile Yes/No".  Our
Unix CAD server is the hub of our engineering infrastructure.  It has links
to the various CAD workstations that themselves have links back to the CAD
server.  Before the introduction of that option in 3.1.0.7, what would
happen is the client would ping-pong between links and generate archives
to:
/dir1/myfile
/dir2 (on workstation)/dir1/myfile
/dir1/dir2/dir1/myfile
/dir2/dir1/dir2/dir1/myfile
...
just looping away until either the max path length was reached or the
recovery log hit 100% and the server fell over dead.  Not good.

By setting that to "No" *SM treats the link file as just another file and
eliminates that ridiculous behavior.  Then all you have to do is clean up
your prior archives that are so crapped up that expiration takes forever if
you don't include the "Skipdirs=Yes" option.

If each archive client is different then you're right, there is no
opportunity to group them.  Then it becomes a question of where you want to
focus your archive job management: lots of individual archive schedules
with embedded properties, or lots of individual archive batch files with
embedded properties.  Since it's six of one, half-dozen of the other it's
pretty much whatever you're confortable with.

Tab






"Gill, Geoffrey L." <GEOFFREY.L.GILL AT SAIC DOT COM>@VM.MARIST.EDU> on 
08/21/2001
09:59:08 AM

Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>

Sent by:  "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>


To:   ADSM-L AT VM.MARIST DOT EDU
cc:
Subject:  Re: Terminated sessions


Tab,

See questions/comments below please;

Geoff Gill
TSM Administrator
NT Systems Support Engineer
SAIC
E-Mail:   gillg AT saic DOT com
Phone:  (858) 826-4062
Pager:   (888) 997-9614


>Our main CAD server runs an archive job at 1:00 am on Sunday
>morning that
>sends 145,000 files totalling 40 GB to our TSM server. Takes
>5-6 hours but
>runs successfully - now that the Archive Link switch is set properly.

Can you be more specific here, I'm not sure what an Archive Link switch is.

>3. Does anyone have a better idea on how to run archives?
>
>Well, approximately 1/4 of all the data in my system is
>archives. That's
>about 2 TB out of 8 TB.
>We've seen where the archive job will fail if one path in a
>list of paths
>is missing or if it denies access.  For example, we had some
>NT boxes that
>were supposed to archive "C:\my documents" and "D:\CAD files" in that
>order.  Every PC had "D:\CAD files" but a few didn't have "C:\my
>documents".  On those PCs, the entire archive job failed even
>though the
>second path was valid.

I also tried this and had the same problem. If there were computers without
specified directories they would report as failed even though they did
sucessfully archive the data they did have. Unfortunately I doubt I have 2
machines alike when it comes to this so that's why I chose the .cmd file on
each server. This way I could be specific to each.