[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [rdiff-backup-users] "Found too many current_mirror incs!"

From: Ron Leach
Subject: Re: [rdiff-backup-users] "Found too many current_mirror incs!"
Date: Tue, 02 May 2017 19:18:18 +0100
User-agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-GB; rv: Gecko/20110414 Lightning/1.0b2 Thunderbird/3.1.10

On 02/05/2017 18:06, Yves Bellefeuille wrote:
  But I must have done something wrong in the new set up.  The first
  attempt to check that mail backup would continue working failed with

  Exception 'Found too many current_mirror incs!' raised of class
'<type 'exceptions.AssertionError'>':

Does this thread help?


Yves, thank you, yes, very helpful. I hadn't thought to search for the error.

In that thread, Ben Escoto had said:

The current_mirror.* files are just 0 length regular files that
rdiff-backup touches so later version know when the last backup was.
I call them "increment files" because their filenames have that form,
and that is how the computer thinks of them.

So you can have as many old increments.<time>.dir files as you want,
but there should only be 1 "current_mirror.<time>.data|snapshot" file,
because the current mirror should only have one time.

I have 3 'current_mirror.*' files which indicates a fault (Ben earlier mentioned that 1 file is normal, 2 files mean the last backup went wrong, and 3 files is just bad).

But renaming one or another - so that it wasn't a current_mirror.* file - resulted in this error:

Server5:/home/ron# rdiff-backup --check-destination-dir --print-statistics -v3 address@hidden::/srv/Data/101vmail
Fatal Error: Destination dir /srv/Data/101vmail does not need checking
Server5:/home/ron# Fatal Error: Lost connection to the remote system

I had to ^C to get a prompt back.

I was very surprised, because I thought that --check-dest... would regress if the check failed, and *not* error if the check passed. To crash is worrying.

I searched for this new error and, in this archive,


Dominic announces what I imagine is an early version of his regress script which overcomes this second error.

But I am unsure whether to run it here because - really - I don't know which of these current_mirror.* files I can, safely, remove. (And I need to remove one to get to the 'doesn't need checking' state.) Because the backups are incremental, I don't want to cause an intermediate diff item to be deleted as a result of my deleting a current_mirror.* file. (This is the same dilemma that the OP in the thread you posted also faced - but the thread doesn't say what he finally decided.)

regards, and thanks - again - for posting that link,


reply via email to

[Prev in Thread] Current Thread [Next in Thread]