Bruno Rogerio Fernandes wrote at about 19:22:29 -0300 on Thursday, June 2, 2022:
> Hi
> Currently, I have a backup of about 15TB of data, using another backup
> solution.
>
> I'm planning to migrate to backuppc V4, because of some problems I'm having
> with the current solution. This backup is done every day through an
> Internet connection and would be pretty dangerous waiting for many weeks to
> do a first full backup (wait for transfers to complete through Internet -
> about 100Mbps client uplink speed). Also, I don't have enough space to
> accommodate two backups (current and the newly created by backuppc).
>
> So, I'm wondering, is it ok to seed the backuppc pool manually? For
> example, if I just checksum every file that I currently have on my backup
> server and manually put them in the corresponding directory, for example:
> Suppose I have a file /current-backup-solution/FILE01, which checksums
> to b2bc1884a98f60259b1fd57d482855dd
>
> And suppose I just copy or hardlink this file to:
> /var/lib/backuppc/pool/b2/bc/b2bc1884a98f60259b1fd57d482855dd
>
> If I start backuppc and do a backup on a client that has this FILE01 which
> checksums to b2bc1884a98f60259b1fd57d482855dd, I could verify that backuppc
> won't transfer this file over Internet, so solving my issues.
>
> Are there any other things that I'm missing? Is it ok to do that?
>
Yes. You are potentially missing at least 3 things:
1. You would need to disable BackupPC_nightly so that it doesn't erase
pool files that don't have a backup associated with them (initially
this would be all your pool files)
This can be done in the config file
2. You may need to create the poolCnt lists that occur in every:
'/var/lib/backuppc/pool/xx' directory
(though not sure if this is only used by backuppc_nightly and its
ilk or if it is used during backups.
You could write a simple script to populate these refcounts since
they would all start with zero references.
Though, if not needed for the backup itself, you could run a
BackupPC_fsck to create/recreate these pool count files at the end
of your tranfer process.
3. If two files have the same checksum (which of course happens a lot
with common files or across backups), then you should at least in
theory do a full file compare to check for the extremely small
chance of (unintentional) md5sum collisions -- as in approximately
1 in 2^32.
Also, if you want backuppc to work with compression (cpool vs. pool), then
you would need to zlib compress each file first. Again, a simple
script could be written to do this.
There may be other edge cases to consider but this is top of mind.
Either way, suggest you experiement on a small subset of your backups
to start...
Jeff
_______________________________________________
BackupPC-users mailing list
BackupPC-users@lists.sourceforge.net
List: https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki: https://github.com/backuppc/backuppc/wiki
Project: https://backuppc.github.io/backuppc/
