Re: [BackupPC-users] NFS woes
2011-04-19 13:03:43
comfi <backuppc-forum AT backupcentral DOT com> wrote
on 04/19/2011 12:49:00 PM:
> I'll perform the tests you guys recommend, although I've done
> hundreds of NFS tests of the years and I know what to expect there.
> This NFS implementation feels fairly typical, speed-wise.
> Personally, I agree with some of you who say something sounds
> broken. 4.5 minutes to pull up a host summary list of 3 servers,
> whose cummulative backup size is less than 400MB? That's long enough
> to copy the entire contents of the backup pool several times over.
I
> guess what I'm looking for here is some insight into what BackupPC
> is doing at that point, so I can figure out what it is that may be
> broken. Is it crawling the entire backup pool for something? Or is
> it just looking at the config files for all available hosts?
No: the BackupPC server process is merely looking
at log files. It's not (e.g.) scanning the pool to see the size:
that is contained in the "backups" file in each host. As
you've seen with the pool on a local drive, the access should be relatively
instantaneous--no matter *how* big the pool is.
> Or, on a completely attack vector, can all of the BackupPC
> operations be performed from a shell?
Yes. Check on the commands in the <BackupPC
dir>/bin. The GUI merely uses those shell commands on your behalf.
How long are your backups taking? Can you give
us some host size/backup time examples?
Are the backups running at full speed? Could
the problem be with the Apache or BackupPC CGI process and *not* an NFS
issue directly? How are you connecting the two: CGI or modperl?
> As for using a different device, a previous backup admin spent a ton
> of dough on this Data Domain device, which is supposed to be
> designed specifically for storing backups. I'd much rather throw
> together something different, but the purse strings are tied fairly
> tightly for this project.
The idea of a dedicated NAS/SAN device for backup
is not terrible; it's the choice of NFS for BackupPC that is questionable!
:) If you could do iSCSI you likely wouldn't have any of these
issues...
I would *really* like to know the results of the dd
test I mentioned in my VMware thread (it's right in the first message!),
as well as any iozone testing you'd like to do. If we can eliminate
base NFS performance we might be able to give you better answers.
I (or anyone else) can't see your system; all
we know is what we've experienced in similar situations, and our gut tells
us that you have an NFS problem that is 100% separate from BackupPC. It
would be nice to be proven wrong, so that we can quit tripping over that
in our thought process.
Timothy J. Massey
------------------------------------------------------------------------------
Benefiting from Server Virtualization: Beyond Initial Workload
Consolidation -- Increasing the use of server virtualization is a top
priority.Virtualization can reduce costs, simplify management, and improve
application availability and disaster protection. Learn more about boosting
the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev _______________________________________________
BackupPC-users mailing list
BackupPC-users AT lists.sourceforge DOT net
List: https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki: http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/
|
|
|