From: Jeff King <peff@peff.net>
To: Owen Ofiesh <oofiesh@maxlinear.com>
Cc: Bryan Turner <bturner@atlassian.com>, Git Users <git@vger.kernel.org>
Subject: Re: Git clone fails with fatal: the remote end hung up unexpectedly
Date: Tue, 11 Dec 2018 05:08:19 -0500 [thread overview]
Message-ID: <20181211100819.GD31588@sigill.intra.peff.net> (raw)
In-Reply-To: <CY4PR19MB152668FA1C3E592CA123E3E5ADA60@CY4PR19MB1526.namprd19.prod.outlook.com>
On Tue, Dec 11, 2018 at 02:25:41AM +0000, Owen Ofiesh wrote:
> > On Mon, Dec 10, 2018 at 4:55 PM Owen Ofiesh <mailto:oofiesh@maxlinear.com> wrote:
> > >
> > > We are seeing an issue where git clone in Linux Ubuntu 14.04.5 LTS fails
> > > with the following error using the HTTP protocol.
> > >
> > > The error on the client is:
> > > fatal: the remote end hung up unexpectedly
> > > fatal: early EOF
> > > fatal: index-pack failed
> > >
> > > The client is writing to an NFS volume.
> >
> > A further detail on this (Owen correct me if I'm wrong), but the same
> > clone performed to a local disk, rather than NFS, succeeds.
>
> Correct. The clone to local disk succeeds.
That's weird. The messages imply that the server has started sending the
packfile, at which point we should be pumping data from git-remote-https
to fetch-pack, and from there into an index-pack process. And the
messages imply that the client saw a hangup of the network connection,
at which point index-pack complained of the truncated packfile (and then
we complained that index-pack complained).
But what's weird is that none of that should really be relevant to the
choice of local filesystem. Is it possible that using NFS stresses the
network to the point that the HTTP connection may be killed?
It's also possible there's something racy in Git's handling of the data,
and NFS is slower to write. What version of Git is this?
> One more data point. I tried this with SSH just now, and the behavior
> is somewhat improved in that I am only seeing the failure on the NFS
> volume so far with 1 in 6 clone attempts (5 successful). Whereas with
> HTTP, we see the failure every time.
If you use ssh's verbose mode, what does a failing case look like? If
you have Git v2.10 or later, you can just do:
GIT_SSH_COMMAND='ssh -v' git clone ...
If it's older, the simplest thing is probably to put:
Host yourserver...
LogLevel Verbose
into your ~/.ssh/config.
I know the problem is more common with HTTP, but I suspect ssh's logging
may be better than ours. ;) You can also try:
GIT_CURL_VERBOSE=1 git clone ...
or if you have recent enough (Git v2.10 again, I think), you can use:
GIT_TRACE_CURL=1 git clone ...
which has a few more details.
-Peff
prev parent reply other threads:[~2018-12-11 10:08 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-12-11 0:48 Git clone fails with fatal: the remote end hung up unexpectedly Owen Ofiesh
2018-12-11 1:15 ` Bryan Turner
2018-12-11 2:25 ` Owen Ofiesh
2018-12-11 10:08 ` Jeff King [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20181211100819.GD31588@sigill.intra.peff.net \
--to=peff@peff.net \
--cc=bturner@atlassian.com \
--cc=git@vger.kernel.org \
--cc=oofiesh@maxlinear.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).