On Mon, Jul 11, 2005 at 12:13:21PM +0200, Christopher Rasch-Olsen Raa wrote:
> Hi,
>
> I have been using bacula to backup one host with a relatively small
> amount of data, and it is all working very well. :)
>
> I am now about to start doing backups of a host wich has ALOT more data
> (+-30GB). I backup to file, and have got a storagedevice with 38 GB free
> space.
>
> The questions I have yet to find a answer to are these:
> 1. How much space for spooling is required? Have I understood it
> correctly if I say that the amount required is equal to the biggest file
> to be backed up?
If you're backing up to file, there's little point in spooling, and in
fact in such a configuration it will only increase your disk space
requirement with no corresponding increase in storage. The purpose of
spooling is to allow you to dump backup data quickly on a fast device (a
disk) and get the backup over with, then transfer the data to a slower
device such as a tape drive at your leisure.
To my knowledge, Bacula can "pause" in the middle of a file if the spool
fills up, but as noted above, if you're backing up to a disk file then
there's no point spooling anyway.
> 2. I plan on doing a full backup _once_ and then only doing incremental
> backups every night. Are there any problems with this? Will this present
> any problems?
The immediately apparent problem with this is that in order to ever
perform a full restore, you will need to keep all of your incremental
backups forever. If you have to restore a machine after six months of
this, you'll have to restore over 180 jobs.
It also means that since your Incremental jobs need to be kept forever,
all the catalog data on them needs to be kept forever as well. This
means that no records can ever be pruned from your catalog, and your
catalog is going to grow HUGE... and you should be backing up the
catalog as well, remember.
A better scheme is to perform a Full backup (possibly several times a
year), then perform Differential backups (which will record all
differences between the current state and the Full backup) monthly or
bi-weekly (I do so weekly), then nightly Incremental jobs that record
only changes since the last backup of any level. This way, after each
Differential job, you can in reasonable safety discard all the
intervening incrementals and differentials since the Full backup, as all
you need at any time to perform a complete restore is the Full backup,
the *most recent* Differential, and the Incremental jobs *since* that
Differential. In my backup cycle, where differentials are performed
weekly, I can do a full restore of any machine by restoring at most
eight jobs -- the Full, one Differential, and six Incrementals. This
also means that as older records are pruned, your Catalog should reach a
given size and then more or less stop growing.
--
Phil Stracchino [EMAIL PROTECTED]
Renaissance Man, Unix generalist, Perl hacker
Mobile: 603-216-7037 Landline: 603-886-3518
-------------------------------------------------------
This SF.Net email is sponsored by the 'Do More With Dual!' webinar happening
July 14 at 8am PDT/11am EDT. We invite you to explore the latest in dual
core and dual graphics technology at this free one hour event hosted by HP,
AMD, and NVIDIA. To register visit http://www.hp.com/go/dualwebinar
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
