On Wed Feb 14, 2018 at 12:53:16 +0000, Holger Levsen wrote:
> On Wed, Feb 14, 2018 at 01:41:31PM +0100, Andreas Beckmann wrote:
> > Since htdocs/ contains a lot of hardlinks to master/ and df counts the
> > files only on the first occurrence. I'm afraid master will be much
> > larger than 60GB. Probably 200+ GB.
> hmpf. also DSA still seems to be unhappy about too many files in
> master/. hmmm.
If you refer to my IRC message, this was only a general question. I
would like to repeat it here so all the others have the chance to read
it as well. And i would like to understand why you think it is important
to have those files backuped.
master/ contains (log-)files per distribution, sorted by passed, failed,
Maybe i am wrong here, but most of these files (at least for the stable,
oldstable, oldoldstable) distributions should be (sort of)
regeneratable. Also most probably the more interestingly ones should be
the failed, bugged, ... ones. The ones that consomed the most space from
what i saw was in the pass/ directory.
I understand that loosing files in case of a complete crash of this VM
is a pain and not nice. But more importantly i see that we have a backup
of the code to regenerate those files (which we have). Is it really
worth that we need to transfer all of those files with a full backup
which we do every few days?
This is NOT about "please don't have those files on disk". This is about
"do we need to transfer all those files with every single full backup?".
I try to understand why those files need to be in a full backup. If i
understand that, i can better discuss that with you (or maybe we do not
need a discussion any more, becuase you convinced me).
Best and warm regards,
- trying to be constructive -
Martin Zobel-Helas <zo...@debian.org> Debian System Administrator
Debian & GNU/Linux Developer Debian Listmaster
http://about.me/zobel Debian Webmaster
GPG Fingerprint: 6B18 5642 8E41 EC89 3D5D BDBB 53B1 AC6D B11B 627B