17 days running "Metadata will be read from filesystem instead"

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

17 days running "Metadata will be read from filesystem instead"

John DiMatteo
Hello,

I have an rdiff-backup process that has been running for 17 days with
the following output:

   Warning, could not find mirror_metadata file.
   Metadata will be read from filesystem instead.

Is there any way for me to check how much progress rdiff-backup has
made and whether I should expect this to finish in the next couple
days?  top shows rdiff-backup fluctuating around 8-14% cpu and iotop
shows disk read 10-12 MB/sec (with disk write at 0 when I checked),
and the backup is about 10TB in size.  backup.log just shows what was
written to stdout.  This job normally completes in 1-6 hours.

What is rdiff-backup doing and can I estimate how long it should take
to finish?  Even if it has to read the whole 10TB at 10MB/sec, that
should still have finished in about 12 days.  Here is the exact
command I am running:

rdiff-backup --backup-mode --exclude-other-filesystems --include
/grail/bam --include /grail/projects --include /grail/TONY --exclude
/grail /grail/ /crusader/backup/rdiff-backup/grail

Maybe it has to read both the backup directory and the target
directory so maybe I should estimate it will take ~24 days to finish
this run with my very slow IO?

For context, a couple weeks ago, rdiff-backup failed when the
filesystem ran out of space, and when it ran again it failed
regressing with "Exception 'Bad index order...".  So I deleted the
earlier of the two .data files and moved the later .snapshot.gz file
out of the way based off the suggestion at
http://www.nongnu.org/rdiff-backup/FAQ.html#regress_failure .  Could
my strange exclude everything then include just a couple
sub-directories be contributing to the slow time? (I made this change
after I ran out of space.) I have some more notes here:
https://github.com/BradnerLab/pipeline/issues/41

I am a part time volunteer at an open source cancer research lab and
we have some important genetic and research data to backup.  Any help
would be greatly appreciated!

Thanks,
John

_______________________________________________
rdiff-backup-users mailing list at [hidden email]
https://lists.nongnu.org/mailman/listinfo/rdiff-backup-users
Wiki URL: http://rdiff-backup.solutionsfirst.com.au/index.php/RdiffBackupWiki