|
Re: [BackupPC-users] NFS woes
2011-04-19 10:47:42
comfi <backuppc-forum AT backupcentral DOT com> wrote
on 04/18/2011 06:57:45 PM:
> I recently fired up BackupPC as a replacement for our convoluted and
> outdated Amanda setup to backup an environment of about 200 servers.
> So far, I have BackupPC version 3.1.0 installed on an Ubuntu 10.04
> system. I'm using this to back up a grand total of three systems:
2
> other Ubuntu machines and the localhost. Everything was running fine
> and dandy when I had all backup data going to local disk. However,
> I'm having massive performance issues after switching to an NFS
> mount for my backup data.
>
> Per the instructions, I mounted /var/lib/backuppc to my NFS share.
> I'm using the following options:
>
> nfs.server.ourstuff.com:/backup/backuppc /var/lib/backuppc
> nfs nfsvers=3,tcp,hard,intr,rsize=32768,wsize=32768,bg
>
> These options were recommended to me by Data Domain, the
> manufacturer of the storage device containing my NFS mount. This
> device only supports CIFS and NFS, sadly. I have also tried with
> larger and smaller rsize/wsize, and noatime.
>
> Everything appears to be working correctly. I can backup and restore
> with no errors. However, the performance is worse than atrocious and
> I can't believe everything is working
I agree strongly with Adam. Most likely, your
problem is not a BackupPC issue, but an NFS issue. To determine this,
do some performance testing with dd. If you see the same issue there
(and I think you will), you can then start to debug your NFS issues.
I also agree strongly with Adam: NFS and performance
do *not* mix. I have put much work into making NFS work with high
performance for use as a VMware datastore. All of my research led
me to believe that it's just not possible without using very high-end hardware
with large battery-backed cache. You might find my research helpful--at
the very least, it will show you how to perform the testing using dd.
http://communities.vmware.com/thread/263165?start=0&tstart=0
Having said all of the above, the results you are
seeing (*minutes* for a webpage to display) do not seem like expected low-end
performance, they sound like something is *broken*. However, it is
*very* unlikly that it's BackupPC that is broken, but rather something
else at a lower layer. Hopefully, a simple test with dd will help
you to reproduce it simply and easily, and let you test things more quickly.
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/
|
|
|
