Did a server re-install (OS partition only). I have the backups preformed with the previous install (of UrBackup), do I have the ability to re-attach UrBackup to recognize the backups from the previous install?
Thanks,
-Jeff
If you still have the files from
C:\Program Files\UrBackupServer\urbackup (or /var/urbackup on Linux)
Stop the server and copy the old files over the new ones. That should be it if the backup storage path is still the same.
If you installed the OS over the existing files and C:\Program Files\UrBackupServer\urbackup is still there you can just install normally. It keeps the old database.
Excellent. While it was too late for me to grab these files this time, now I know what I should add to the servers backup directories.
Thanks a bunch
-Jeff
Ok. In my continually evolving lab; home network the family depends on :0. I’ve found occasion to do another server install. This time having the folder backed up necessary to bring the UrBackup server back to life.
Did what you recommended, stopped the service, copied the old folder over, restart the service. Worked excellent, all the clients were back online and the server recognized the backup locations. There are some inconsistencies, the server log indicates “Error retrieving last backup.” on each client when attempting to do an incremental file backup. And, the client attempts to do an incremental backup often as a result. There are not new folders being created for incremental backups. A full backup works fine. I thought triggering a full backup might resolve the error for the incremental, no dice.
Haven’t tried an incremental image.
Any thoughts on what I might look at to resolve this?
Hmm normally this error message indicates that it does not find any (not interrupted) file backup in the database. But as you are saying that it still happens after a full backup… I do not have any idea…
If you want to investigate closer the SQL statment that returns nothing is:
SELECT incremental,path FROM backups WHERE clientid=? AND done=1 ORDER BY backuptime DESC LIMIT 1;
You can find out the clientid with SELECT id,name FROM clients;
and use the sqlite shell here to access the database: http://www.sqlite.org/download.html (sqlite3.exe backup_server.db)
I’ll give that a look when I get a chance. Tonight likely. Thanks for the hints.
Tried your query, got a “Error: database disk image is malformed” without output, tweaked the query a bit, similar error
http://dl.dropbox.com/u/114127/UrBackup_QuerySnip_jhughesco.JPG
Is there a repair for SQLite databases, seems to be the issue? Guess I could re-import the .db file, if the backed up copy is not corrupted.
-Jeff
Did some research on the error message. There seems to be some discussion about the integrity of the journal file. This is a likely culprit since…This is not a copy of the …Urbackupserver\Urbackup directory from a machine state after it was shutdown clean. This is a copy of the directory grabbed during an image of the server machine. Not very forward thinking on my part. Of course the condition of the database will be in flux as it is working to make an image, of itself, makes sense.
If you agree, this is likely an "operator headspace " issue and not be repaired, do I have any options at this point? Any way to trigger the server to re-learn the backups? If not, how might I clear everything back to zero (I’ll keep the current backup files until I have replacements) and get the server happy again?
Any thought this may be developed into the server? A server state backup that runs with each backup procedure, maybe drops a clean save of the directory into the root backup location. This might be impractical as it may only be useful to guys like me, trying to run a development/lab/production network. Poor combo, but hey whatta ya do on a home budget.
[quote=“jhughesco”]
If you agree, this is likely an "operator headspace " issue and not be repaired, do I have any options at this point? Any way to trigger the server to re-learn the backups? If not, how might I clear everything back to zero (I’ll keep the current backup files until I have replacements) and get the server happy again?[/quote]
Yeah think so as well. Did you copy the backup_server.db-wal and backup_server.db-shm as well?
What you could try otherwise is dump the database in sqlite3.exe and then reimport it:
sqlite3.exe backup_server.db
.output tmp.sql
.dump
.quit
rename backup_server.db to somethin else
sqlite3.exe backup_server.db
.read tmp.sql
.quit
If you are on Windows you can just install the Client as well and use it to backup this directory. As it uses shadow copying the backup server database state should be consistent.
Should have those on the image backup.
sqlite3.exe backup_server.db
.output tmp.sql
.dump
.quit
rename backup_server.db to somethin else
sqlite3.exe backup_server.db
.read tmp.sql
.quit
Give it a shot and let you know what comes of it.
This is how I have them arranged, client and server on the server box. Is there a difference between an image backup and a file backup as far as the shadow copy is concerned? I captured the directory on an image not a file backup.
[quote=“jhughesco”]
This is how I have them arranged, client and server on the server box. Is there a difference between an image backup and a file backup as far as the shadow copy is concerned? I captured the directory on an image not a file backup.[/quote]
No, should be the same. If you do not know how to extract files from an image backup. It’s in the FAQ: http://www.urbackup.org/FAQ.php?lang=en
Got this one. One of the many beauties of W7 and SVR 2008.
Wha-Wha-Wha
No dice. Got a bunch of output:
Error: near line 18343: database disk image is malformed
Error: near line 18348: database disk image is malformed
Error: near line 18350: database disk image is malformed
Error: near line 18351: database disk image is malformed
Error: near line 18354: database disk image is malformed
Error: near line 18365: database disk image is malformed
Error: near line 18367: database disk image is malformed
Error: near line 18374: database disk image is malformed
etc…
pulled the files from the image again, same.
Is there a way to clear the server brain so it can be happy again?
That is weird. Are you certain you copied the backup_server.db and the backup_server.db-wal ?
Otherwise I do not know what else to do: http://www.sqlite.org/faq.html#q21
I will test this myself if shadow copying the database is enough as well.