BackupPC-users

Re: [BackupPC-users] Excessively High Traffic on Full Backups (Debian Squeeze on ARM Processor)

2014-08-26 12:20:08
Subject: Re: [BackupPC-users] Excessively High Traffic on Full Backups (Debian Squeeze on ARM Processor)
From: <backuppc AT kosowsky DOT org>
To: "General list for user discussion, questions and support" <backuppc-users AT lists.sourceforge DOT net>
Date: Tue, 26 Aug 2014 12:18:10 -0400
Les Mikesell wrote at about 10:22:51 -0500 on Tuesday, August 26, 2014:
 > On Tue, Aug 26, 2014 at 9:48 AM,  <nhoeller AT sinet DOT ca> wrote:
 > > I recently upgraded my ARM plug computer from an ancient version of Ubuntu
 > > to Debian Squeeze running backuppc 3.1.0.  Everything seemed to be going
 > > well until yesterday when a full backup of a relatively quiet server
 > > downloaded 1.5 GB of data.  The original full backup seven days ago was 
 > > only
 > > 2.3 GB and incremental backups have been around 152 MB.  When I checked the
 > > logs, there were 47 'create' entries adding up to about 300 MB (I excluded
 > > 'create d' entries) while the rest were 'pool' and 'same'.  The 300 MB
 > > includes some double-counting since the backup failed (probably 
 > > connectivity
 > > problems) and restarted.
 > >
 > > This reminded me of a problem with File::RsyncP on ARM processes that Craig
 > > fixed back in 2010.  I was not able to verify the File::RsyncP version 
 > > level
 > > (probably 0.68) but used CPAN to upgrade File::RsyncP to version 0.70.  The
 > > incremental backup this morning finished in half the time of previous
 > > incremental backups which is encouraging.
 > >
 > > I am in the process of upgrading another ARM plug computer from Debian
 > > Squeeze to Debian Wheezy.  It looks like Wheezy comes with File::RsyncP 
 > > 0.70
 > > and backuppc 3.2.1.
 > 
 > I don't know the details of this bug, but if you are using the
 > --checksum-seed option you might want to check the behavior again on
 > the 3rd full run.   The rsync block checksums should be cached on the
 > 2nd run so they don't have to be computed on the server side for
 > subsequent runs.
 > 

If I recall correctly, I found two bugs on ARM/debian processors
several years ago. One of which is the one referenced above which is
due to a problem in the type definition in the adler32_checksum in
File-RsyncP-0.68.

I believe the other problem was in the Debian lenny library for md5sum
which resulted in inconsistent checksum naming of pool files relative
to x86 versions.

------------------------------------------------------------------------------
Slashdot TV.  
Video for Nerds.  Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
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/

<Prev in Thread] Current Thread [Next in Thread>