Raf wrote:
. . .
Backups run fine for a while and then suddenly, I'm offered the option of rebuilding a new backup. I haven't timed the delay between full backups, but I'd say once every month/two months.
Since about 10.6.4, Time Machine does a quick verification (a few seconds) of backups made over a network at the beginning of every backup. Periodically (perhaps once a month), it does a much more extensive (thus longer) one. Those checks are where, if it finds a problem, it sends the message that it must start over.
2/11/12 9:18:47.710 AM com.apple.backupd: Error: (-43) SrcErr:NO Copying /Users/raphaeljosse/Documents/Microsoft User Data/Office 2011 Identities/Main Identity/Data Records/Message Sources/0T/0B/0M/77K/x27_77091.olk14MsgSource to /Volumes/Time Machine Backups/Backups.backupdb/rjosse-iMac-i7/2012-02-11-072920.inProgress/C6A9C3F2-A 0CB-433B-8467-491356FB0183/Machiavel/Users/raphaeljosse/Documents/Microsoft User Data/Office 2011 Identities/Main Identity/Data Records/Message Sources/0T/0B/0M/77K
That's a file system problem, almost certainly on the backup drive (but theoretically could be a problem with the Office file on the Mac). That backup failed.
The actual problem could be something that happened to the directory earlier, or could have happened on this backup.
2/11/12 10:16:52.583 AM com.apple.backupd: QUICKCHECK ONLY; FILESYSTEM DIRTY
2/11/12 10:16:53.076 AM com.apple.backupd: Runtime corruption detected on /Volumes/TMBackup/rjosse-iMac-i7.sparsebundle (fsck_hfs -q termination status: 3)
The next backup (an hour later) detected the problem during the check.
I'm under the impression that the backup reset has occurred more often as my hard disk occupation increased.
Possibly, but with RAID5 on the backup destination, it's hard to tell. That's just another bit of complexity -- did something go wrong with the data sent to the NAS, or did the NAS's controller drop the ball?
Are you connected to the NAS via Ethernet or WIFI?
I used to do backup on a Time Capsule but switched to the Qnap as the 1Tb capsule bacame slightly too small for my backup needs (it only holds backups).
An alternative is to connect a larger drive to the USB port and back up to it, instead of the TC's internal HD.
And I too had issues with the restart of a backup under Snow Leopard on that capsule, so both Lion and Snow Leopard had this issue.
Do you mean a backup failed, and you had difficulty finding/fixing the problem? There could be lots of reasons for that.
My best guess would be that for big backups, there comes a time where the sparsebundle has so many parts that the likelyhood of an error is directly linked to the increasing number of parts included in the sparsebundle (in short the chaining of these parts).
Not really. Most users don't back up more than 1 TB, but those who do don't seem to have problems of this sort more often. The more you back up, of course, the more opportunities there are for problems to arise, just because you're sending more data (especially if you're backing up wirelessly).
Maybe this will get better if Apple allowed a larger size of individual sparse elements.
I don't know what that means. There isn't a limit on the size any of the contents, other than the maximum size of the disk image, which by default is set to the size of the volume it's on.
It's just a special type of disk image -- it has it's own partition map and format, just like a physical disk.
And while the structure of the backups is complex, it's no more so for a large backup. See How Time Machine works its Magic for an explanation.
If you have a larger HD you can connect to the TC, you might want to try backing-up to it. If the problems stop, that will at least narrow it down a bit.
Keeping secondary backups is also prudent. 🙂
It's a bit safer to use a different app, as well as a different destination. See Time Machine - Frequently Asked Question #27 for some suggestions.
