-
Notifications
You must be signed in to change notification settings - Fork 5.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Gitfs - No Content-Type header in response #38630
Comments
@terminalmage have you seen this error before with gitfs + pygit2? Thanks, |
I don't remember seeing this error before. I have same problem, with same configuration Versions Report
|
Also seeing this exact same issue with https remotes
|
I've never seen this before. Those of you seeing this issue, are your masters behind a web proxy? One common thread between all of you is that you are all on pygit2 0.21.4. The pygit2 project is known for making non-compatible changes in point releases (they even have a warning about this in their docs). However, 0.21.4 is not exactly a new release, so I have my doubts as to this being a sudden change in their API. I would have expected to have seen something like this by now if that were the case. One other common thread in all three of you is that you're all using CentOS 7.3 with python-cffi 1.6.0. I know for a fact that there is currently a bug report filed against pygit2 in the RHEL/CentOS repositories, seemingly caused by the newer python-cffi. It does seem like there is an incompatibility with pygit2 and newer python-cffi, however whether that is fixed in newer pygit2 I am not sure. At the very least, there seems to be an incompatibility between pygit2 0.21.4 and python-cffi 1.6.0. One thing you can try is to downgrade python-cffi to 0.8.6. Links to the RPMs are still available in the Fedora Buildsystem here. |
If it helps, I resolved this by upgrading libgit to 0.25.1 and then installing the latest pygit2 using pip. |
Yeah, it definitely appears to be an upstream issue. |
I can confirm that I am now seeing this in cent7. Unfortunately there is nothing I can really do until RedHat fixes this issue on their end. If you don't need authenticated repos, maybe try installing GitPython and set |
FWIW, my fix on CentOS7 went something like:
There may have been some libcurl and libssh2 dependencies as well, but these can be satisfied from OS packages. |
Does anyone know if / where this has been submitted upstream? |
@cedwards I referenced the bug report in an earlier comment on this issue: |
I had the same problem with salt 2015.8.12 and RedHat 7.2 |
@Eldrad53 That might be good information to add to that upstream bug report, if it hasn't already been mentioned. |
A build has been submitted to the Fedora Build System which should fix this issue. https://bodhi.fedoraproject.org/updates/FEDORA-EPEL-2017-dc9f5cde6d The RPM for the build can be downloaded via https://koji.fedoraproject.org/koji/buildinfo?buildID=845387 and installed using Those with Fedora accounts can give positive feedback (click the radio buttons under the check mark) to vote it into the stable EPEL repository. |
The build I mentioned in my previous post is now in the stable repositories (pending in some cases a mirror sync):
The warnings there are due to version 0.21.4 of pygit2's usage of cffi, and are A) not fatal, and B) presumably fixed in later pygit2 releases. Can someone experiencing this issue (@cedwards ?) update to |
@terminalmage Unfortunately it doesn't seem to have fixed the problem. Logs after salt-master restart
|
@timwsuqld yeah, I can reproduce that. I did some investigation, and this is still an upstream issue. I opened a new upstream bug report: https://bugzilla.redhat.com/show_bug.cgi?id=1422583 |
Downgrading to http-parser 2.0.5 via the RPMs at https://koji.fedoraproject.org/koji/buildinfo?buildID=703753 should allow you to proceed until such time as libgit2 is rebuilt for EPEL 7. |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. If this issue is closed prematurely, please leave a comment and we will gladly reopen the issue. |
The upstream bug report was confirmed fixed in April of 2017. Closing this issue. |
Description of Issue/Question
Recently upgraded to 2016.11.1. We use gitfs_remotes to pull in a number of formulas. Started getting a lot of errors pulling gitfs remotes from Github. Our local Gitlab ones don't give any errors, but the local ones use git/ssh not https.
Setup
Steps to Reproduce Issue
Lines when running log level debug
Versions Report
The text was updated successfully, but these errors were encountered: