ADSM-L

Re: Mac client and special characters question

2005-11-25 10:27:03
Subject: Re: Mac client and special characters question
From: Richard Sims <rbs AT BU DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Fri, 25 Nov 2005 10:26:53 -0500
On Nov 25, 2005, at 9:40 AM, Farren Minns wrote:

Thanks for that.

I have had a look at the filespaces and as far as the TSM server is
concerned, they are indeed set to "Is Filespace Unicode = NO". So,
from
what I have read I have to put an entry in the TSM System
Preferences on
the client that sets autofsrename to yes. I also understand that
this will
then rename the old fs to something_OLD and then create a new fs.

Refer to the Technote or the client manual: you can opt to either
have TSM automatically perform the filespace rename, or do it
manually. I would manually rename the filespace on the server, as I
would not want things happening without my control, particularly as
it affects later access to the data.


So does this mean that the entire fs will be backed up again?

Yes - it's starting fresh.


and

When it comes to future restores, will I then see both the new
unicode fs
and also the *_OLD one? And how does this effect restores, will I
have to
select the files I may want from the *_OLD fs and restore them to a
different location (i.e. the new fs)?

Yes, just like restoring across file systems in the same client. The
client owner has to remember about the old filespace.


I have one more question. On this particular client there are a
couple of
filespaces that no longer exist but that I want kept in backup.
Will this
process effect those also or just the file spaces being backed up
again?

Affected is just the one filespace that receives this special
attention. File systems which no longer exist on the client do not in
any way participate in backup processing. Orphan filespaces - the
bane of TSM server administrators - need tracking, periodic
communication with client administrators (email), and manual deletion
at some time in the future. The orphans continue to participate in
time-based objects expiration, but not versions-based expiration, as
there is no new incoming data.

  Richard Sims

<Prev in Thread] Current Thread [Next in Thread>