Backup Central

hi,

when will backuppc 2.1 be available? at the moment, we backup the
backup-pool via cronjob, but it would be much more comfortable to do this
via backuppc too.

regards
lorenz



-----Ursprungliche Nachricht-----
Von: Craig Barratt [mailto:cbarratt < at > users.sourceforge.net]
Gesendet: Donnerstag, 18. Marz 2004 09:23
An: Ghate, Niranjan P
Cc: backuppc-users < at > lists.sourceforge.net
Betreff: Re: [BackupPC-users] Backup to Tape/DVDs?


"Ghate, Niranjan P" writes:

I have some systems backing up all the data to a Linux server running
backup-pc. This is great as a backup solution, but does little for
catastrophic events like a Fire. I am thinking of backing up the
backup-pc server to DVD RW disks that I could carry home with me.

I am looking for suggestions on what I should backup and how.

For the "What" part, I am debating if I should backup the entire backup
pool (could get large), or just the last full backup for each pc.

For the "How" part, I was thinking of using scdbackup
[http://scdbackup.sourceforge.net] to burn afio archives on DVD+RW
media.

BackupPC 2.1.x includes an archive feature. From the CGI interface you
can ask it to generate compressed tar archives of the most recent backup
for any or all the hosts. Output files can be optionally split (eg: for
650MB CDs) and optional parity files can be generated. A backend script
can be customized to write the output files to DVD, tape or whatever.

Craig


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
BackupPC-users mailing list
BackupPC-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/backuppc-users
http://backuppc.sourceforge.net/

"Lammersdorf, Lorenz" writes:

when will backuppc 2.1 be available? at the moment, we backup the
backup-pool via cronjob, but it would be much more comfortable to do this
via backuppc too.

I'm hoping to get the first beta version out either this weekend
or next (I'm traveling next week).

Based on the feedback from this release, there will likely be a
couple of additional betas. Based on the experience with 2.0.0,
it took a four months or so for the production release.

Craig


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
BackupPC-users mailing list
BackupPC-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/backuppc-users
http://backuppc.sourceforge.net/

On Thu, 2004-03-18 at 10:42, Craig Barratt wrote:

Based on the feedback from this release, there will likely be a
couple of additional betas. Based on the experience with 2.0.0,
it took a four months or so for the production release.

This may be too weird to consider, but I'd like to see some way
to intelligently 'nest' backuppc copies and this might also satisfy
some people's need to backup the backup server. I have several remote
offices set up with Linux boxes as file servers and internet gateways
and am currently running backuppc using rsync-over-ssh to back them
up from our main office. I'd like to run an instance of backuppc
at each remote site to catch the other machines at that location,
then let that be copied back to the main office too. That would
probably work as-is but to restore the nested copy you would have
to first restore an equivalent of the remote backuppc machine which
would then be able to extract the files. It would be nicer if it
magically understood the backuppc structure on the client side and
merged it into the server as though it had been created there so
you would get pooling of identical files even across the nested
backups and be able to extract any file directly from the web
interface.

My other off-the-wall idea for backuppc can probably already be done
with some external scripting. I'd like to be able to get a diff of
any two files selected by date or backup number, host, and path in
the style of cvsweb. Then with a suitable backup history of /etc on
all my machines I could almost pretend that I had the configurations
under version control even though I've been too lazy to do it.

---
Les Mikesell
les < at > futuresource.com

Les Mikesell writes:

Based on the feedback from this release, there will likely be a
couple of additional betas. Based on the experience with 2.0.0,
it took a four months or so for the production release.

This may be too weird to consider, but I'd like to see some way
to intelligently 'nest' backuppc copies and this might also satisfy
some people's need to backup the backup server. I have several remote
offices set up with Linux boxes as file servers and internet gateways
and am currently running backuppc using rsync-over-ssh to back them
up from our main office. I'd like to run an instance of backuppc
at each remote site to catch the other machines at that location,
then let that be copied back to the main office too. That would
probably work as-is but to restore the nested copy you would have
to first restore an equivalent of the remote backuppc machine which
would then be able to extract the files. It would be nicer if it
magically understood the backuppc structure on the client side and
merged it into the server as though it had been created there so
you would get pooling of identical files even across the nested
backups and be able to extract any file directly from the web
interface.

Yes, this has come up before. I need to think about how this could
be possible. It would be a lot simpler if the master BackupPC server
simply had a mirror copy of each remote data tree (eg: updated via
rsync). Then BackupPC on the master could simply be pointed at
the specific data directory mirror for browsing the remote backups.

This would allow you to download individual files, tar and zip files,
but not initiate direct restores to the remote machines. Would that
be good enough?

My other off-the-wall idea for backuppc can probably already be done
with some external scripting. I'd like to be able to get a diff of
any two files selected by date or backup number, host, and path in
the style of cvsweb. Then with a suitable backup history of /etc on
all my machines I could almost pretend that I had the configurations
under version control even though I've been too lazy to do it.

One part of this is already in BackupPC CVS. You can now display a
backup history of a single directory. It's a table with a row for each
file name and columns for every backup. The cells in the table show
each file version. Since BackupPC knows about identical files, the
cells span multiple columns if the file was the same between backups.
You can click on the cells to download each file version.

This could be extended to allows diffs to be displayed between
different versions. Would you ever want to display diffs between
different files in different directories (the GUI would be a lot
of work), or just the same file in the same direcotry and host between
different backups? The latter would be a lot easier.

Craig


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
BackupPC-users mailing list
BackupPC-users < at > lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/backuppc-users
http://backuppc.sourceforge.net/

On Thu, 2004-03-18 at 23:54, Craig Barratt wrote:
It would be nicer if it
magically understood the backuppc structure on the client side and
merged it into the server as though it had been created there so
you would get pooling of identical files even across the nested
backups and be able to extract any file directly from the web
interface.

Yes, this has come up before. I need to think about how this could
be possible. It would be a lot simpler if the master BackupPC server
simply had a mirror copy of each remote data tree (eg: updated via
rsync). Then BackupPC on the master could simply be pointed at
the specific data directory mirror for browsing the remote backups.

This would allow you to download individual files, tar and zip files,
but not initiate direct restores to the remote machines. Would that
be good enough?

I don't think the direct restore capability matters - you would do that
from the remote server copy anyway unless there was a site disaster. In
many cases the machines backed up by the remote server would not be
directly reachable anyway. However, if you just rsync each tree you
lose the efficiency of linking common files across the remote sites.
Linux has a utility called 'hardlink' that links identical files
that might be sufficient if it wouldn't confuse backuppc. Perhaps
you could do a perl version that would take advantage of the cpool
name convention and would work if the OS didn't include the hardlink
program.

My other off-the-wall idea for backuppc can probably already be done
with some external scripting. I'd like to be able to get a diff of
any two files selected by date or backup number, host, and path in
the style of cvsweb. Then with a suitable backup history of /etc on
all my machines I could almost pretend that I had the configurations
under version control even though I've been too lazy to do it.

One part of this is already in BackupPC CVS. You can now display a
backup history of a single directory. It's a table with a row for each
file name and columns for every backup. The cells in the table show
each file version. Since BackupPC knows about identical files, the
cells span multiple columns if the file was the same between backups.
You can click on the cells to download each file version.

This could be extended to allows diffs to be displayed between
different versions. Would you ever want to display diffs between
different files in different directories (the GUI would be a lot
of work), or just the same file in the same direcotry and host between
different backups? The latter would be a lot easier.

The two common scenarios would be the same file on different dates
and the same file on different machines. The GUI could offer a
choice of 'diff against previous version', 'select for diff', and
'diff against selected'. That way the most likely 'what just changed'
question is answered with one click but you could still pick anything
else by marking one and browsing to the other.

---
Les Mikesell
les < at > futuresource.com