Problem:
========
It isn't unusual that internet connections abort once a day. A lot of
providers are doing that. Therefore it isn't unusual too that a full
backup aborts and will be saved as an partial backup. If the next full
backup aborts too, it will only be saved if it contains more files than
the previous partial.
If a big file will be added on client after a partial backup is stored
than it is possible that the transfer of this new file, within the next
full run, needs a lot of time. If this full backup aborts too than it is
possible that it contains less files than the previous backup. It will be
cancelled and this big file have to be retransmitted within the next full
run.
In average I can transmit 6GB per day, depending on the internet upload
bandwith of the client. So it isn't unlikely that I will never backup this
client, because of the above problem.
Don't advise the $Conf{PartialAgeMax}. This would change a reliable
backup into a game of luck ;-) (I know, that's not really true. But I didn't
believe this would be a reliable solution)
Solution:
========
BackupPC_dump compare the filecount and decide if the old partial backup
must be removed or not. If BackupPC_dump remove the old partial it will
rename the $TopDir/pc/$client/new directory into $TopDir/pc/$client/<nnn>
after this. In the other case it will only remove the
$TopDir/pc/$client/new.
Instead this, BackupPC_dump should "move" the $TopDir/pc/$client/new
always over $TopDir/pc/$client/<nnn>, overwriting existing files and
creating new files in $TopDir/pc/$client/<nnn>. The "NewFileList" contains
all new files anyway. Therefore BackupPC_link should do his job perfectly
too.
Advantage:
========
+ All transmitted files will be saved within the actual backup.
Nevertheless if a full backup will be aborted after a short or long time
or not.
+ The continuation (strictly speaking the 2nd or later continuation) of a
partial backup must not retransmit files which are always transmitted.
Disadvantage:
========
None known.
Such a full backup can have a duration of more than a few days. It is not
sure which version of changed files are in the backup or not. But the
duration of a [full] backup isn't really influenceable. So also without
the above "improvement" we didn't know which version of a file is within
the backup.
e.g.: If we use snapshots and a file will be deleted after the start of a
backup it will be stored within the backup. If we don't use snapshots that
isn't sure. It depends from the time where it was deleted. If we use
different shares the above case isn't sure too.
There is no different between the BackupPC_dump implementation and the above
suggestion, but less retransmits.
What is your opinion about this?
br
Mattthias
--
Don't Panic
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
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/
|