TimeMachine flush index error: 0

Question

I'm using OS X 10.6.8
Time Machine is configured to use an external drive connected via fw800
It seems to be stuck in a cycle (from console logs):
Starting standard backup
Backing up to: /Volumes/video-DRIVE Q/Backups.backupd
Event store UUIDs don't match for volume: Macintosh HD
Node requires deep traversal:/ reason: must scan subdirs|new event db|
(long time passes, an hour or two)
Error: Flushing index to disk returned an error: 0
Backup canceled.
Before this began to happen, I was getting errors about back up disk being full.  I tried deleting old backups manually to correct this (this backup disk stores the backups for three computers).  I transferred the old backup folders out of Backups.backupd to the trash, but I couldn't successfully empty the trash.  I ended up sudo rm -rf the directories in the trash folder to free up the space.  Time Machine has not successfully completed a backup since (syndrome is always the series of messages above)
I've tried repairing the backup disk and the backed-up (main system) disk.  I've also tried "resetting time machine" by moving the TimeMachine.plist file to another location (out of /Library/Preferences)
A file "2011-07-11-072049.inProgress" remains stubbornly in the backups folder. 
Any ideas/answers would be much appreciated!  Thanks!

Answer 1

tonyfromframingham wrote:
Event store UUIDs don't match for volume: Macintosh HD
Node requires deep traversal:/ reason: must scan subdirs|new event db|
(long time passes, an hour or two)
That's required after a failed backup (see #6 in Time Machine - Troubleshooting), but unless you have an enormous amount of data, shouldn't take that long.
Error: Flushing index to disk returned an error: 0
Backup canceled.
A problem writing the index to your TM drive.
Before this began to happen, I was getting errors about back up disk being full.  I tried deleting old backups manually to correct this (this backup disk stores the backups for three computers).
You connect the drive to each Mac to do backups?   If so, it's a bit late now, but it's usually better to partition the TM disk, so one Mac's backups won't "crowd out" those from a different one.  See #4 in  Time Machine - Frequently Asked Questions.
Or, if you're backing the Mini up directly, and the others via your network, the backups made over the network will be in separate sparse bundles.  It's a good idea to have them in separate partitions, too, in case you have to delete the backups for one -- you can just erase the partition.
I transferred the old backup folders out of Backups.backupd to the trash, but I couldn't successfully empty the trash.
Unfortunately, Apple doesn't do a very good job of warning folks not to do that via the Finder or Terminal.  You can delete individual backups, or all backups of selected items, via Time Machine.  See #12 in the FAQ.
I ended up sudo rm -rf the directories in the trash folder to free up the space.
Did that get all of them?  If so, that, and repairing the backup disk, usually fix it.
A file "2011-07-11-072049.inProgress" remains stubbornly in the backups folder.
That's a partial backup (actually a "package"), left over from one that failed or was cancelled.  Is it in the folder for the Mac in question, or a different one?  Normally, Time Machine will recover that on the next backup (so the files there don't have to be copied again).  Since all else has failed, you might try deleting it via the Finder and emptying the trash (that will take a while, of course).
It may be that the backups are damaged in a way Disk Utility can't detect.  That's rare, but does happen.
It's also possible that there's a problem with the external disk or the connection to it. 
I'd say, try deleting the .inProgress package.  If that doesn't help, tell us how the backups are done (all direct, some via network, etc., in the same or separate partitions), and we'll try to find a fix.