Re: [rdiff-backup-users] questions about error messages meaning

[Ben Escoto]

>>>>> "P" == frache  <address@hidden>
>>>>> wrote the following on Tue, 13 May 2003 18:11:27 +0200

  P> by the way, it could be very usefull if rdiff could build a
  P> repport at the end of the backup process, and tell more about 1/
  P> the result of the backups (succes or not), 2/ wich files failed,
  P> and why, and 3/ is the coherence of the backup ok ? If my
  P> yesterday backup failed, are all oldest backup lost ? I do
  P> netheir understand what this long error messages mean nor if my
  P> backup coherence is still ok.

Yes, you are right.  The newer versions try to make it clearer which
errors happened to which files and what this means.  For instance,
see:

http://rdiff-backup.stanford.edu/error_policy.html

  P> Error '[Errno 5] Input/output error:
  P> 
'/export/sda1/backup/diuflx03_SUNdata/rdiff-backup-data/increments/ftp/soft_eng/WEBSs''
  P> processing ftp/soft_eng/WEBSs Skipping
  P> ftp/soft_eng/WEBSs/.message because of previous error Skipping
  P> ftp/soft_eng/WEBSs/WEBSS30B2.HQX because of previous error
  P> Skipping ftp/soft_eng/WEBSs/WEBSSPAPER.HQX because of previous
  P> error Skipping ftp/soft_eng/WEBSs because of previous error

In this case what happened is that there was an "Input/output" error
processing ftp/soft_eng/WEBSs, so all the files under that directory,
like ftp/soft_eng/WEBSs/.message are skipped.  The idea is that we
shouldn't mess with files underneath a file with an error.  For
instance, if we couldn't create or update a directory correctly we
wouldn't want to write new files into that directory.

I don't know what caused the initial input/output error on the WEBs
file.

Subsequent backups should work, but any changes in that directory
haven't been recorded for that session.


-- 
Ben Escoto

pgpqVLvjUUiDV.pgp
Description: PGP signature