Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 14 Aug 2011 22:18:49 -0300
From:      "Carlos A. M. dos Santos" <unixmania@gmail.com>
To:        "deeptech71@gmail.com" <deeptech71@gmail.com>
Cc:        freebsd-current@freebsd.org
Subject:   Re: ghost files
Message-ID:  <CAJ4jsadBLCKtO63uw6ocmvQHXcJbYHK2FBjASsTa%2B_eqDf57Vg@mail.gmail.com>
In-Reply-To: <CAF6hryRQUT5i9SgThrgzGwVHOf8uzFnSygWfgokvsH-8KSccUg@mail.gmail.com>
References:  <CAF6hryRQUT5i9SgThrgzGwVHOf8uzFnSygWfgokvsH-8KSccUg@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Aug 7, 2011 at 3:08 AM, deeptech71@gmail.com
<deeptech71@gmail.com> wrote:
> as of recent times, some git rebase operations fail unexpectedly with
> an error: "cannot create .git/index.lock: file exists". an
> investigation session was something like the following:
> $ ls -l .git
> the index.lock file is not in the shown list.
> $ ls -l .git/index.lock
> the file is listed! it's a regular file, its size is ~60KiB.
> $ cat .git/index.lock
> some file content is shown.
> $ mv .git/index.lock .git/someplace
> moving fails with: index.lock: file does not exist.
> $ ee .git/index.lock
> some file content is shown, which i edit and save. then the
> .git/index.lock file really disappears (cat, direct ls, ee, mv, etc.
> do not find the file), and the content i put in the .git/index.lock
> file via ee is now in .git/index.
>
> H$X!!!!!111
>
> i still have some git rebase operations (which are notably
> disk-active) fail from time to time with "impossible" reasons (for
> example: something like "cherry picking failed", or something like
> "cannot move git-rebase-new to git-rebase-todo: file does not exist").
>
> i'd note that the hard drive is kind of old (>7 years), and i recently
> had the power cut during port build operations twice, although the
> (UFS) filesystem is fsck-clean.
>
> has anyone experienced anything like this?
> what is the possibility of a filesystem/kernel bug or fsck bug?

Do not start supposing that this is a filesystem/kernel bug so early.
I have seen similar mysterious messages @work, with Git servers
running on Linux, so I'm inclined to believe that it is a software
issue.

--=20
"The flames are all long gone, but the=EF=BB=BF pain lingers on"



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ4jsadBLCKtO63uw6ocmvQHXcJbYHK2FBjASsTa%2B_eqDf57Vg>