Davina Treiber <Davina.Treiber AT PeeVRo.co DOT uk> wrote on 06/12/2008
04:25:13
AM:
> MIchael Leone wrote:
>
> > My NetWorker server is down, at the moment, while a tape library is
being
> > repaired. But I have jobs that save to disk (disk space on my SAN),
and
> > from there clone onto tape. This is all done in the job
configurations,
> > and it all Just Happens. There is no need for me to manually clone
data
> > from saved disk to tape; NetWorker does it all.
>
> In simple environments, the automatic cloning as part of a group may be
> sufficient. However when an environment becomes more complex, a scripted
> solution to handle cloning is usually more reliable, more controllable,
> and more flexible. Below are some examples where a scripted solution is
> superior to automatic cloning.
Of course; I never meant to imply otherwise. Merely that scripting does
not *have* to be a solution. Sometimes (as in my case), you can let
NetWorker handle all (much?) of the work involved.
> If you have multiple savegroups which all require cloning, you may not
> want to run the cloning as soon as the backups have completed, since you
> may not want to tie up tape drives that are still required for other
> backups.
Granted.
> If automatic cloning fails for any reason, you may not easily know which
> save sets did not get cloned. Also a failure of cloning does not cause
> the group to fail so you may not be aware of the failure.
I mail savegroup completion results to myself. I only have a dozen or so
clone jobs, so it's not too hard to scan the first few lines for the word
failed.
If automatic cloning failed (for whatever reason), I have small scripts
that query for the previous night's saveset ID, and that have copies < 3
(meaning cloning failed, and these savesets must be manually cloned). And
then it clones them.
But kicking off these scripts only happens in case of failure; when jobs
all complete successfully (as they do the majority of the time), as I
said, it Just Happens, and NetWorker takes care of it all. I only involve
myself in case of failure. I suppose I could script the whole checking
process, but I prefer to do it by hand. If you have more than the 11 or 12
clone jobs that I run, I can see where that situation would require a
different procedure.
> You may choose a solution where save sets are cloned to different pools
> depending on level, automatic cloning could not do this.
Granted.
> If you have a lot of tape drives available, you may want to run multiple
> cloning sessions to optimise drive usage. This leads to some scripting
> challenges but can be done.
As someone pointed out here earlier, clone jobs are one-to-one (i.e., only
1 clone job may be writing to a device at a time). So (unless I am
completely off-base), I can run as many cloning jobs at one time as I have
available tape drives.
>
> There are probably other good examples.
>
> You say that NetWorker "does it all", well I wouldn't quite agree, but
> what it does do that some rivals don't is to provide you with the hooks
> and tools to fill in the gaps in the way that you want.
I did not say it "does it all"; I said it "Just Happens". :-) The
difference being that it "Just Happens" when it all works; I don't need to
run any scripts or take any action, etc, because NetWorker handles it all.
When it fails, I involve myself. Sure, it would be nice for the error
recovery to be more robust. But the thrust of my statement was about the
effort involved on my part when it does what I want (i.e., clone). I don't
need to write or execute a script or set of actions for cloning to take
place, unless something has gone wrong in the night.
> I suppose that hundreds of NetWorker users have written cloning scripts
> (I have probably written a dozen or so variants myself over the years),
> and some might say this is something that the product should do better
> natively. However, cloning requirements vary widely for different users
> and it would be hard to meet all those needs with a single product.
Exactly. In my case, most of the time, I don't have to do anything for
cloning to properly happen. When it doesn't properly happen, it's not that
hard (*for me*) to make it happen, via small scripts.
To sign off this list, send email to listserv AT listserv.temple DOT edu and
type "signoff networker" in the body of the email. Please write to
networker-request AT listserv.temple DOT edu if you have any problems with this
list. You can access the archives at
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
|