Michael Weidel wrote:
Since no one answered till now, I hope I can help out. Looking at the code,
you can see:
If the file "<vault-dir>/dirvish/default.hist" exists, dirvish does the
following:
dirvish looks at the file <vault-dir>/dirvish/default.hist and takes the
latest backup date (let's call it "<image-date>") in this file for which the
directory "<vault>/<image-date>/tree" exists.
Then it checks if it was called with the option "-init" (I think you didn't
call it with this option) or if the directory "<vault>/<image-date>" exists
(which has to exist, because the subdirectory "tree" existed).
So I see the following possibilites leading to this error message:
The file "<vault-dir>/dirvish/default.hist" doesn't exist or no one of the
directories in this file does exist. Could you please have a look at the file
"<vault-dir>/dirvish/default.hist" to see if this is the case?
If this file doesn't exist, dirvish takes "--reference" as reference or if it
isn't called with this option, it takes "--branch". In this case please have
a look at the reference or branch directories - do they exist?
I hope this information helps you and I hope I have listed all possibilities
above.
I looked at these settings and they all seemed correct, however your
deletion suggestion ended up pointing me in the right direction! I
looked at the image summary file and found out much to my surprise that
according to the "Status", dirvish actually considered them all *failed*
backups. Apparently some of the mail spool files had been unable to
transfer because they were deleted by the time rsync got to them. The
actual error was "Status: error (24) -- file vanished on sender".
I did some searching into that error on this list, and found out how to
fix it by changing the status line to "Status: success", and I also
fixed the RSYNC_CODES to not consider code 24 an error, and now my
latest backup is considered successful once again!
So everything seems to be backed up and running smoothly again... it
apparently wasn't that I deleted the latest backup or anything like
that, it was simply that most of my backups were considered "errored",
and I deleted all the ones that were marked successful...
Now that I think about it, maybe the "error" was what was causing the
strange and sudden increase in backup disk usage in the first place!
Since apparently dirvish doesn't look at errored images when it's doing
its next incremental backup, wouldn't it have to create a new copy of
any files added since the last success, every night? That would
certainly explain it!
Now I should probably go and set up my cron job to alert me when the
backup isn't successful, huh... That would've solved this the "right"
way :) Thanks for your help, and your patience.
Even though there are definitely full, complete backup images there...
is it possible to repair my dirvish vault do you think? I'd like to
avoid having to toast everything and re-initialize.
Also, what is the "proper" way to delete images on a one-off basis? My
expiry policy generally works well, but there has been a lot changing on
the drive recently, resulting in abnormally high growth, which is why I
had to go in and clean some stuff out... it seems I did it wrong.
I don't know if this is the "proper way", but you could expire images the
following way:
Change the date of the image you want to expire in the file
"<vault-dir>/<image>/summary" in the line which begins with "Expire". Then
run "dirvish-expire".
That's a clever way of doing it, I haven't tried it yet, but this was
certainly a helpful suggestion nonetheless, as it helped me figure out
what the problem was!
Cheers,
Michael
I really appreciate the help Michael. Thanks!
Bradley.
_______________________________________________
Dirvish mailing list
[email protected]
http://www.dirvish.org/mailman/listinfo/dirvish
