Hi
I'm considering submitting a Request for Enhancement (RFE) to Legato for
NetWorker's cloning function. Several years ago, I ended up writing a
rather complex script to handle our group's cloning needs. At the time (and
now), I thought that this could (maybe should) be built into the standard
NetWorker product.
Our group did backups of ~160 servers into a single backup pool (the Default
pool, in fact). We used a single pool for the obvious reasons of
simplifying our NetWorker setup, ensuring the best utilization of the tape
media (if you have multiple pools, you have a greater chance of one pool
running out of media, while another pool had plenty, etc).
Our site had a data retention strategy that dictated 5 different retention
periods (1 year, 2 year, 5 year, 7 year, and 10 year, I think), based on the
criticality of the data, legal, governmental, and corporate requirements,
etc. So, we could not use NetWorker's automatic cloning mechanism, because
it places all savesets in the same backup clone pool. In almost all cases,
the retention period for all data backed up from any one system was the
same.
Our cloning strategy was to clone all full backups, and to clone these full
backup savesets to clone pool media that matched the retention period.
Because of the volume of data saved and the price of media, we decided it
wasn't cost-effective to clone all data and send it away for 10 years (the
maximum retention period), since most of the data would expire after 1 to 2
years.
So, if system FOO's data needed to be saved for 2 years, then we would clone
FOO's full savesets to the '2_Year' backup clone pool.
Also, we didn't want to have any more machinery than necessary (tables
outside of NetWorker, self-modifying scripts, etc.).
So, here's what the script that I wrote did:
* created backup clone pools called '1_Year', '2_Year', '5_Year',
'7_Year' and '10_Year', and labeled media into them;
* created savegroups with the same names as the backup clone pools
above. These savegroups were never executed; rather, they were just
''markers' we used to assign a backup clone pool name to a client.
* assigned one of these savegroups to every backup client [because
this is the only way I could think of to equate a backup client with a
particular clone pool; more on that latter];
* wrote a single ksh script that:
* ran every morning at the conclusion of nightly backups;
* used mminfo to determine which full backups had not yet been cloned,
and which backup clients had created these savesets;
* used nsradmin to determine which savegroup with 'Year' in the name
was assigned to each of these backup clients
* loop:
* for every retention period, clone those savesets belonging to the
clients with the savegroup equal to the 'n_Year' currently being processed
to the corresponding backup clone pool
This ran very well for many months, until I left that job. But creating
this machinery and writing the script took a long time, and it was very
difficult for anyone other than me to understand.
So, here's the question: is this a useful idea, so useful that NetWorker
should be enhanced to do it? Do you need to clone to different clone pools?
Does your data have different retention periods?
Comments, please!
Thanks
tl
Terry Lemons
> CLARiiON Applications Integration Engineering
EMC²
where information lives
4400 Computer Drive, MS D239
Westboro MA 01580
Phone: 508 898 7312
Email: Lemons_Terry AT emc DOT com
Terry Lemons
> CLARiiON Applications Integration Engineering
EMC²
where information lives
4400 Computer Drive, MS D239
Westboro MA 01580
Phone: 508 898 7312
Email: Lemons_Terry AT emc DOT com
--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
|
