Hello,
On 1/19/2006 10:13 PM, Ralf Gross wrote:
Arno Lehmann schrieb:
On 1/19/2006 3:11 PM, Ralf Gross wrote:
[snip]
At the moment, tapes are the only way to archive data and put them into a
safe here. 200GB+ disk are still more expensive and need more space than
tapes.
No, I wouldn't recommen to use disks as backup media *shudder*...
rather, I suppose you might find that an autoloader with more slots
might better suit your needs.
You are right, the changer is 6 six years old!
Nothing wrong with that - I've got some old stuff running here as well.
My advantage is that tape changing doesn't need long ways.
Unfortunately the
budget situation is a bit tight at the moment. But that's an other
story.
Apart from that, I don't exactly understand why you make a distinction
between archive and regular jobs. The simplest solution is to keep the
full backups as long as you need them, and have the differential and
incremental jobs go to pools that cycle more quickly.
I want to keep the monthly archive tapes for 2 years. But the regular full
backup should be kept only 14 days. I want to change the tapes only one
time a month. Thus the regular backup must be overwritten more frequently
and I think I have to use 2 different jobs/pools to achieve that.
Difficult.
You could do this differently, too, but that would require clever
scripting (for the retention times etc.) and a rather complicated setup
of your schedules.
I'd like to keep this as simple as possible, because the setup is not
for a typical IT department. I'll be the only person that (hopefully)
will have detailed knowledge about bacula.
This might be a philosophical difference, but personally, I think you're
doing two things wrong: First, having only one person to handle backups
and restores is not enough. You *need* at least a second person,
preferraby on the location, who can (with as much "emergency manual" as
necessary) change tapes, trigger backups, handle restores and forward
everything critical he doesn't understand to you. Second, if your setup
has some limitations and working around them is complicated anyway it's
best to do that behind the scenes, i.e. better make heavy use of
(tested!) scripts than require much manual intervention from people not
qualified for that sort of task.
But don't be bothered by my concerns too much - it's your setup.
Still I wonder why you need that distinction - I mean, you're doing the
full backups anyway, and why not simply remove the volumes for one of
the full backups once a month, set their status to Archived, and thus
keep things simpler (and needing less autochanger slots)?
If remove the tapes and put them into a safe - which might be placed
at a differnt location - and I need them for a restore I'll lose some
time. That was one reason for the 2 differnt jobs/pools.
I think I've to take a deeper look at the documentation to get the
whole picture, what I can accomplish with bacula.
In that case, you could leave the diff/incr volumes in the changer and
only remove the full backup volumes.
Hm, you would use 2 separate pools for the full and inrc. volumes, take
out the monthly tape and put it in the safe? How can I be sure the monthly
tape is not overwritten by a regular job if I forget to take it out? (I'm
sure this is covered by the manual...).
True. Retention times is the keyword...
I'm quite sure that I'll have some questions to the retention times in
the future (no threat! ;)
The focus should be - as mentioned
above - to change the archive tapes only once a month.
Should be possible... an actual estimate with real numbers and all that
stuff would usually be something I get paid for :-) but I suppose you
might find that you don't have enough slots:
Agreed!
Only talking about server 2. Three tapes per full backup. You keep them
for two weeks and need space for the next backup: 9 tapes already.
Ok, with 12 slots it seems to be the best method to have just 2 jobs
for the 2 server and handel the archive tapes as you mentioned above.
[snip]
Maybe I'm a just bit to confused at the moment ;)
Maybe... start with a simple experimental setup, with a small fileset,
disk based volumes, and run your backups multiple times a day. This
would allow you to observer the volume management live.
That's a good idea. Is it possible to simulate the changer behavior with
slots in any way (I guess not)?
That's not possible, but you will not need that, too. You can emulate
your setup by manually creating the volumes you'd have in the
autochanger. These volumes can belong to different pools, and you can
limit the size each volume can have.
Whenever you get an operator intervention request you know that you'd
have to take some tapes and go to the autochanger in "real life".
I'll try that after building 1.38.5 from source.
Then I suppose you are already at work, right ;-)
Arno
Thanks, Ralf
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
--
IT-Service Lehmann [EMAIL PROTECTED]
Arno Lehmann http://www.its-lehmann.de
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
