[rdiff-backup-users] Exception handling

[Xavier Bertou]

Hi,
I've been using rdiff-backup for a few months but I am facing an
important issue: it looks like I have some hardware issue on one of my
machines, and rdiff-bakcup stops with an exception:
Exception 'CRC check failed' raised of class 'exceptions.IOError'
If I try to start a new backup to another directory, I get:
Exception '[Errno 13] Permission denied:
'/backup/d2/home/tep2/abastod/.galeon/mozilla/galeon/Cache/D1637DE7d01''
raised of class 'exceptions.OSError'
(of course, I run it as root and have all permissions)
Verifying the state of the the mirror I get:
(millions of Warning: Cannot find SHA1 digest for file
asoreyh/offline/External/BoostC++/1.33.1/lib/libboost_date_time-gcc-d.a,
perhaps because these feature was added in v1.1.1)
And then:
Exception '[Errno 5] Input/output error' raised of class 'exceptions.IOError'
Of course, in some sense, it's not rdiff-backup's fault. In /var/log/messages:
Jun 11 17:07:44 localhost kernel: [17270030.372000] attempt to access
beyond end of device
Jun 11 17:07:44 localhost kernel: [17270030.372000] hdd1: rw=0,
want=33774020944, limit=390716802
(which is "funny" BTW)
However, I am really concerned by the fact rdiff-backup just stops
doing the backup when it finds these problems. I would really expect
it to keep on doing the backup and give serious warnings about the bad
files. Does this behaviour make sense for a future release of
rdiff-backup?
What can I do to "repair" my backups?
Thanks for any hint.
Cheers,
--
Xavier