Josh Kuo wrote:
> Hi all:
>
> I am new to bacula, I have read through the documentation and a
> portion of the mail archive, but perhaps I am searching using the
> wrong keywords, I could not find what I am looking for, and I was
> hoping that someone could point me in the right direction. Apologies
> if this has been brought up and answered before, I could not find the
> answers myself.
>
> Here is the problem I need to solve:
>
> - Client site has low bandwidth (upload) to the Internet (less than 0.5Mbps).
> - Client site has several PCs, with a combined data of 30GB that needs
> to be backed up. Assume new data rate at 1GB per week.
> - I am bringing a server on-site to serve as on-site backup, with
> bacula installed. I install bacula on all the client boxes, and they
> back up at least once a day to the on-site bacula box. Let's call this
> the LAN backup tier.
> - I also need to replicate some of the data off-site, for emergency
> recovery in case of catastrophic event (say fire burned down the
> office). I need to do this at least once a week (maybe even more
> frequently, nightly would be preferred). Let's call this the WAN
> backup tier.
>
> My plan is to use bacula for LAN backup at least once a day (may be
> more frequent depending on our needs), and use some sort of 'rsync'
> mechanism for the WAN backup, so each remote site's data is replicated
> in our data center.
>
> Question:
> Can I use bacula for my WAN backup? I did not think it will work too
> well because the LAN backup would leave me with some binary
> "tarball-ish" files every night, in my example, if I do a full backup
> (not incremental), then I need to re-transmit 30GB of data across the
> WAN link, which would take literally days to complete. I only want to
> send changes, and only do the 30GB massive dump the very first time.
> In fact, my plan is to send a guy on-site to manually carry the 30GB
> (maybe more) data back to our data center (where possible).
>
> My thoughts currently is to do my LAN backup (this part is easy with
> bacula), and then immediately after each successful backup, do a
> restore locally somewhere, so it is decompressed back out to
> directories and files; then I can use rsync (or bacula client) to do
> my WAN level backup, which should only transmit changes over the said
> 0.5Mbps link to the data center. Hopefully, since I am only
> transmitting changes, it will only take a few hours every night,
> instead of several days if I cam transmitting the whole thing.
>
> So another question:
> How do I setup Bacula to automatically restore somewhere? The docs I
> am reading so far talks about using the command-line console to
> restore files, and I don't want to write a expect script to do the
> restore...
>
> I am hoping that someone out there has already solved this problem
> before, and came up with a much more elegant solution than me. If so,
> I would love to hear what others have done in this situation. If I am
> way off on my solution, please tell me, as I am still fairly new to
> the whole large scale backup thing.
>
> Thanks in advance for any help, and again, my apologies if this has
> been answered in the past. I will keep reading the docs and mailing
> list archive in search of my answers.
>
> -Josh
>
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Bacula-users mailing list
> [EMAIL PROTECTED]
> https://lists.sourceforge.net/lists/listinfo/bacula-users
For the WAN link I would suggest the following model modified of what I
developed for roaming clients with slow links:
Schedule a restore job (or many if you want to use bootstraps) to the SD
machine. Then in that job, after dumping the files, run the rsync
script. The script should be with something like this:
-azv --delete-after --bwlimit=40
The bwlimit is good to ensure that the site is not crippled by the job.
Also, running it every night is a good idea split up the work.
The WAN can then run jobs on the imported rsync files if you want more
than just a repository off-site. Before the Full job, run a second
restore job with otherwise the same rsync job but with '-c'. This
verifies every file's contents to be sure it has not been corrupted.
It's way faster than sending every file, but takes a lot of resources
while running.
To schedule a restore job, and to run scripts, see my example:
Job {
Name = JOBNAME
Type = Restore
Client = LAN-SD
Fileset = FILESET
Storage = LAN-SD
Schedule = JOBSCHEDULE
Priority = 10
Pool = Default
Messages = Standard
Where = /tmp/PUSHJOB
Bootstrap = "/var/lib/bacula/CLIENT.bsr"
ClientRunAfterJob = "/etc/bacula/SCRIPT"
}
Schedule {
Name = JOBSCHEDULE
Run = daily at 21:05
}
If you do it like that, you need bootstraps created during Backup. If
you want everything at once, you need something else. The fileset is not
important with bootstraps if I recall correctly.
Keep in mind that there are quite a few steps in the process, which is
generally not a good idea - but a workaround for limited resources. You
could say that you cannot afford having all that data because the link
is too slow.
Also consider if you should run the rsync job on each client instead.
Cheers,
Johan
--
Johan Ehnberg
Email: [EMAIL PROTECTED]
GSM: +358503209688
WWW: http://www.ehnberg.net/johan/
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Bacula-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/bacula-users
