Skip Hanson wrote:
All,
Hello again. As you may know we added scheduled cloning to the 7.6.1
release. With your help we can continue to evolve this functionality.
Therefore, I have a few questions for you regarding how you script nsrclone.
1. In what scenarios are you doing this?
We use cloning a lot. Certain special groups have automatic cloning
enabled because they write their data to special tape pools (each of
these groups uses a separate backup and clone pool). This is done on a
daily basis, but having cloning done at the end of each group completion
is not a major issue because there are never that many of these groups
running in parallel so little to no drive contention.
However, most groups use the *same* backup tape pool, and cloning is not
enabled and is manually done only once per month, following full
backups. I have to run cloning manually for these because 1. there would
be too much drive contention if cloning was enabled for those groups
since there are many running in parallel, all using the same pool, and
2. it would force all backup levels to get cloned if I had it enabled;
unless, of course, I set up different groups and schedules for the fulls
versus the incrementals for each group - kinda of a pain. So I basically
have to wait until all the fulls are done, and then I run the clones
manually.
It would be nice if you could specify that cloning be carried out
automatically, following a backup for the group, BUT only if it's a
certain backup level, without having to create a separate group/schedule
for each possibility. But that would only help for point 2. above, but
not point 1.
2. Do you make use of the -J option?
No. we're only using one storage node now.
3. Do you make use of the -d option?
What does that option do? I don't see if for our release (7.5.1)?
4. How do you schedule the clone operations?
As noted above, for the fulls, it's manual. For the special groups,
cloning is enabled for the groups. We do also have two Oracle servers,
and the Oracle backups for these (RMAN with NMO) are broken down into
multiple groups, each group having cloning enabled. HOWEVER, it's very
important for us that the groups be run in a special order. There is no
reliable way to do this in the GUI, so the groups are disabled, and we
run the backups from a script, using the 'savegrp' command for each
group. This way we can force the order, with cloning automatically
occurring after each group finishes but before the next one starts. That
way, not only do they run in the right order, but there's no contention.
Each of the Oracle servers runs its backups at different times. There's
little overlap, but never a guarantee that there won't be some overlap.
Regardless, both servers use separate backup and clone pools. Also, the
server writes its index and bootstrap to a file type device, and cloning
is auto-enabled for the group, with the clones going to tape, We also
use staging to move the data on the file type device to a tape pool.
5. Would the ability to chain save groups or clone operations be useful
(having a group or clone operation's execution tied to the ending of a
previous operation)?
YES!!!!!!!!!! This is something I've wanted for a long, long time. As I
mentioned above, we have to do this manually now from scripts. Of
course, you could guess what a reasonable time window would be for job
completion and just have the groups start far enough apart to allow for
that time, but it's a guessing game at best and would be inefficient
best case. Having group C start only after group B and group B only
after group A would be nice.
6. Would the ability to have groups or clone operations just run once be
useful?
Not sure I understand what you mean here?
Feel free to respond back to the list or to our usability alias at
networker_usability AT emc DOT com. Thank you again for all the help and
feedback.
On a somewhat related note, I notice that the group backup schedules
allow only by week or month. Suppose I want to have a full run on the
last Saturday of every month? If I select week, it will be every
Saturday. If I select month, it will be on that same date (e.g. 30th) of
every month which will not always be a Saturday. Sure, I can override
the backup schedule manually, but that requires having to go in every
month and dork around with it or set it all up months in advance - a
pain when you're dealing with a lot of schedules and groups. So I solve
this by running the full backups from a script that determines when that
day is and then runs the full backups. The script runs from cron every
day, so if it's not yet the date, it does nothing.
George
Cheers,
Skip Hanson
NetWorker Usability
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
--
George Sinclair
Voice: (301) 713-3284 x210
- The preceding message is personal and does not reflect any official or
unofficial position of the United States Department of Commerce -
- Any opinions expressed in this message are NOT those of the US Govt. -
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
|
