Re: [BackupPC-users] lost pc/*/backup file
Solved, just found /usr/share/backuppc/bin/BackupPC_fixupBackupSummary Robert On 01.09.2010 00:19, Robert Strötgen wrote: > Hi, > > during a crash I lost the file pc//backup (and backup.old). Is > there a way to restore the file (automatically or manually from > pc///backupInfo)? > > Thanks and best regards > Robert -- This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
[BackupPC-users] lost pc/*/backup file
Hi, during a crash I lost the file pc//backup (and backup.old). Is there a way to restore the file (automatically or manually from pc///backupInfo)? Thanks and best regards Robert -- This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] Anyone using r1soft hotcopy as a pre/post backup command?
Ok, I tested the following and it worked. Here is what I used for others future reference> DumpPreUserCmd: $sshPath -q -x -l root $host /usr/sbin/hcp /dev/sda1 (/dev/sda1 is my / partition so I run hcp against that) DumpPostUserCmd: $sshPath -q -x -l root $host /usr/sbin/hcp -r /dev/hcp1 (this stops the snapshot and unmounts it in one fell swoop) RsyncShareName: /var/hotcopy/sda1_hcp1 (or whatever your version of hcp mounts the snapshot to) Obviously I an using rsync to back up this machine. I have tested backups and restores. The restore you just have to tell it to restore to / instead of the /var/hotcopy/. path and you should be good. I hope this helps someone in the future. Donny B. On Tue, Aug 31, 2010 at 11:03 AM, Kameleon wrote: > I am looking at using r1soft's hotcopy to enable a snapshot of the > filesystem before backuppc does it's magic. Similar to how Microsoft Volume > shadow copy works or LVM's snapshots. What I am wondering is this: Does > anyone currently use this type of setup and if so, would you mind sharing > your pre/post commands so I can compare to what I am thinking. Thanks in > advance/ > > Donny B. > -- This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] Backup backuppc
On 8/31/2010 12:09 PM, Josh Malone wrote: > Farmol SPA wrote: >> Hi list. >> >> I would like to ask which is the simplest yet effective way to dump >> backuppc stuff (mainly __TOPDIR__) eg to a removable hard disk that will >> be used in a disaster recovery scenario where the plant were destroyed >> and I need to restore data from this survivor device. Is a "rsync -aH" >> enough? >> >> TIA. >> Alessandro > > If your 'topdir' is its own filesystem, lvm, etc., you can use 'dump' to > back up a snapshot of your pool. Adapting the link earlier: > > #!/bin/bash > > EXTDISK=/dev/sdc > POOLDISK=/dev/mapper/Group0-Vol0 > > lvcreate -l '100%free' -s -n snapshot /dev/volgroup/backuppc > mount $EXTDISK /mnt/tmp > dump -a0f - /dev/volgroup/backuppc | gzip > > /mnt/tmp/pool-backup.0.gz But time a restore before deciding whether the speed will be acceptable or not. I think dump can write the inode tables fairly quickly, but the restore side has to do lookups to rebuild the hard links and might take days to complete. -- Les Mikesell lesmikes...@gmail.com -- This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd ___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] Backup backuppc
Farmol SPA wrote: Hi list. I would like to ask which is the simplest yet effective way to dump backuppc stuff (mainly __TOPDIR__) eg to a removable hard disk that will be used in a disaster recovery scenario where the plant were destroyed and I need to restore data from this survivor device. Is a "rsync -aH" enough? TIA. Alessandro If your 'topdir' is its own filesystem, lvm, etc., you can use 'dump' to back up a snapshot of your pool. Adapting the link earlier: #!/bin/bash EXTDISK=/dev/sdc POOLDISK=/dev/mapper/Group0-Vol0 lvcreate -l '100%free' -s -n snapshot /dev/volgroup/backuppc mount $EXTDISK /mnt/tmp dump -a0f - /dev/volgroup/backuppc | gzip > /mnt/tmp/pool-backup.0.gz -- Joshua Malone Systems Administrator (jmal...@nrao.edu)NRAO Charlottesville 434-296-0263 www.nrao.edu 434-249-5699 (mobile) BOFH excuse #327: The POP server is out of Coke smime.p7s Description: S/MIME Cryptographic Signature -- This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
[BackupPC-users] Anyone using r1soft hotcopy as a pre/post backup command?
I am looking at using r1soft's hotcopy to enable a snapshot of the filesystem before backuppc does it's magic. Similar to how Microsoft Volume shadow copy works or LVM's snapshots. What I am wondering is this: Does anyone currently use this type of setup and if so, would you mind sharing your pre/post commands so I can compare to what I am thinking. Thanks in advance/ Donny B. -- This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
Re: [BackupPC-users] Why does backuppc transfer files already in the pool
Here is how I think an algorithm could work, which requires no
changes on the remote side, so it should work with the plain rsync
protocol, but likely will require some changes to rsyncp.
Let:
B be the BackupPC_dump process
P be the peer (PC to be backed up)
R be the remote rsync process
F be a file on P that backuppc wants to save
Gc, Gc2be two compressed files in the same chain in the server pool
G1, G2 be their uncompressed data
F[x], G1[x] be the xth block of F or G1 respectively
HR(b) be the rsync rdelta-hash of block b
HL(b) be backuppc's hash-for-pooling of block b
0. B has already contacted P and R has been spawned. Maybe some
files have already been transferred;
Now B arrives at F and determines that it need to transfer the
file (i.e. it's new, or the attributes have changed).
1. Since we need the first 8 blocks to determine the pool chain, we
just receive the first 8 blocks. Depending on the protocol, this
possibly requires us to ask for the hashes first, and to pretend
that each hash simply does not match out (fake) local hashes:
newfile = []; // array of blocks
for( x=0 ; x < 8 ; ++x ) {
tmphash = recv( HR( F[x] ));
if tmphash == IMPOSSIBLE_HASH { explode(); }
newfile += recv( F[x] );
}
2. Okay, now we have the first 8 blocks and can determine the pool
chain, if there is one:
localfiles = []
localfiles = get_pool_files( newfile[0], newfile[7] )
3. And now iteratively ask for hashsums from the remote, compare
them to the corresponding blocks in the pool files (which may be
cached!), and assemble the new file:
x = 8
while !F.eof() {
tmphash = recv( HR( F[x] ));
tmpblock = "";
for G in localfiles { // iterate G1, G2, …
if tmphash == HR( G[x] ) {
tmpblock = G[x];
break;
/* possible optimisation: replace localfiles with just
* localfile so that for successive blocks, not all local
* files have to be compared.
*/
}
}
if !tmpblock {
tmpblock = recv( F[x] );
}
newfile += tmpblock;
}
4. At the very end, determine if the new data is identical to
existing data and hardlink. This can either be done with
a full-file hashsum (to be safe), or the above algorithm can
determine that while it's running.
If this logic cannot be implemented inside or atop of rsyncp, then
maybe the following logic would work:
1. Create a buffer in memory and fill it with zeroes;
2. Pass this buffer to rsyncp and ask it to synchronise F into it;
3. After 8×128k blocks have been determined:
a. determine if there is a corresponding pool file;
b. if yes, uncompress it and write it into the buffer;
4. Let rsyncp continue, with a potentially modified buffer.
Thoughts?
--
martin | http://madduck.net/ | http://two.sentenc.es/
"i've got a bike,
you can ride it if you like,
it's got a basket. a bell that rings, and things to make it look good.
i'd give it to you if i could,
but i borrowed it."
-- syd barrett, 1967
spamtraps: madduck.bo...@madduck.net
digital_signature_gpg.asc
Description: Digital signature (see http://martin-krafft.net/gpg/)
--
This SF.net Dev2Dev email is sponsored by:
Show off your parallel programming skills.
Enter the Intel(R) Threading Challenge 2010.
http://p.sf.net/sfu/intel-thread-sfd___
BackupPC-users mailing list
BackupPC-users@lists.sourceforge.net
List:https://lists.sourceforge.net/lists/listinfo/backuppc-users
Wiki:http://backuppc.wiki.sourceforge.net
Project: http://backuppc.sourceforge.net/
[BackupPC-users] Host Summary page never loads
Hi I'm having a problem with the BackupPC web interface host summary page which with a full list of hosts configured (24 hosts) never loads when I click on the link. Reducing the list of hosts in /etc/Backuppc/hosts to a single host, the page loads after a long delay (about 30 seconds). The odd thing is that I have another server in another location which works fine and all hosts are reachable from both servers. The backups seem to be running OK although they do seem rather slow in some cases. The OS is Debian 5 with all updates applied and the BackupPC version is 3.1.0 on both servers. The storage space is on a SAN mounted via NFS with the uid & gid matched on the SAN & BackupPC server (this is the same for both servers). Does anyone have any ideas as to what might be causing this problem? Thanks and regards Mark. -- This SF.net Dev2Dev email is sponsored by: Show off your parallel programming skills. Enter the Intel(R) Threading Challenge 2010. http://p.sf.net/sfu/intel-thread-sfd___ BackupPC-users mailing list BackupPC-users@lists.sourceforge.net List:https://lists.sourceforge.net/lists/listinfo/backuppc-users Wiki:http://backuppc.wiki.sourceforge.net Project: http://backuppc.sourceforge.net/
