[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: [rdiff-backup-users] rdiff-backup + smbfs = :'(
From: |
Marc Provitt |
Subject: |
RE: [rdiff-backup-users] rdiff-backup + smbfs = :'( |
Date: |
Wed, 13 Oct 2004 13:42:37 -0700 |
> -----Original Message-----
> From: Nathaniel Smith [mailto:address@hidden
> Sent: Tuesday, October 12, 2004 7:00 PM
> To: 'address@hidden'
> Subject: Re: [rdiff-backup-users] rdiff-backup + smbfs = :'(
>
>
> On Tue, Oct 12, 2004 at 06:11:00PM -0700, Marc Provitt wrote:
> > my GUESS is this is why it's b0rking on trying to erase
> that directory. What
> > I don't get is; why, when I do the same thing on an NFS
> mount, do the
> > python2.3 processes not hang onto the directory and when I
> use SMBFS it
> > fails this way?
>
> Guess: on NFS it's holding onto the directory as well, it's just that
> UNIX filesystem semantics allow you to delete an open object (and the
> kernel does some clever reference counting things to make this work),
> while Win32 filesystem semantics do not. So what it's doing is legal
> on a UNIX filesystem like NFS, but not on a Win32 filesystem like
> SMBFS.
>
> -- Nathaniel
>
actually no, the NFS share lets go of the link, lsof comes back with no
processes holding onto the directory.