-
Notifications
You must be signed in to change notification settings - Fork 18k
cmd/go: git remote (Gerrit) did not send all necessary objects #31399
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
Comments
I can confirm that this is happening to me as well. It is not 100% constant. My docker builds experience it about 98% of the time.
A Same issues with go 1.11.5, 1.11.8, 1.12.3 in Docker ( Local git:
I also less frequently get this issue:
EDIT: It appears that this patch resolves the issue for my builds: --- a/packages/go/packagename/go.mod
+++ b/packages/go/packagename/go.mod
@@ -7,7 +7,7 @@ require (
redacted/logging v0.0.0
golang.org/x/net v0.0.0-20190320064053-1272bf9dcd53 // indirect
golang.org/x/oauth2 v0.0.0-20190319182350-c85d3e98c914 // indirect
- google.golang.org/api v0.3.1
+ google.golang.org/api v0.0.0-20180603000442-8e296ef26005
google.golang.org/genproto v0.0.0-20190404172233-64821d5d2107
google.golang.org/grpc v1.19.1
) |
Reproduced locally with the same error. Also seen with the same instructions as above:
Perhaps Gerrit is acting up again? |
I've traced this to a sequence of Git commands that reproduces this issue outside of Go. The Gerrit team has been notified and is working to resolve the issue. |
Is it some workaround for the problem? |
The latest we have from the Gerrit team is that this might be a bug specific to the content of the |
Yes. Second running also helps, but it very ugly hack. I added vendor dir to my projects. It looks like better workaround. |
Hello, this is the team that manages googlesource.com. *.googlesource.com is hosted by using JGit, and we believe that there's a bug in upstream JGit that cannot handle unshallow git-fetch. We recognize that this is reproducible easily, but we cannot spot the buggy code in the JGit. If there's an option to do a full clone always instead of partial clone, that can be a workaround. |
We identified the root cause and the fix is uploaded at https://git.eclipse.org/r/c/140463/. Our in-house JGit comitters are already gone home, so this will be reviewed tomorrow at best. If everything goes well and our SREs are happy about a Friday release, we might be able to cherry-pick this fix tomorrow. The worst case ETA would be next week. |
That did the trick. If it happens again, we want to know. |
Waiting on the following to be resolved: golang/go#31399
Everything was working this morning but I've seen a regression in the past 10 mins:
|
Still doesnt work for me. |
I've tried to reproduce this again, and it's now working for me. @chmking @mr-tron Are you able to reproduce this with a clean module cache using the original instructions at the top of the issue? Could you post the output of |
@jayconrod This is happening on our CI machines running the build in a docker without a go mod cache. It worked this morning without fail many times until it started failing at the time of my post. Now the issues is intermittent. I can try to replicate it for you later. |
I met the same problem when using
$ git --version
git version 2.21.0 |
Is there a workaround for |
Please see the first comment for the workaround. |
Reproduced with JGit:
|
I found another bug in JGit code that made it exclude dcbf3e3c23816703a8f5ebfe6583cc3fc2793526. But after I fix that, I found another bug that doesn't send a parent of v0.1.0. I'm debugging JGit more. |
https://git.eclipse.org/r/c/140900/ and https://git.eclipse.org/r/c/140901/ will fix the the three tag unshallow case. Tried all three previously reproduced cases. |
A possible temporary workaround for docker builds might be
|
Still seems to be happening. Pinged internal bug. |
#31399 (comment) is not rolled out yet. |
Oops, I thought those CLs had been deployed already. @draftcode Do you expect this will fix the |
@iamtakingiteasy not working anymore |
@anasanzari true, but it seems original issue of fetching You can remove it from your pipeline and build as normal if that was your use-case. Also it is indeed odd such fetch would not work. Might be somewhat related issue (which those commits might fix). |
All fixes are rolled out now, and this seems to be working. Thanks @draftcode for working on this! I'll close this but everyone please comment if you're still seeing this, and we'll investigate. |
@jayconrod issue is still there. I've added my case in https://github.com/anasanzari/golang_build_check. |
@anasanzari, thanks. I can reproduce the failure. Gerrit/jgit is still broken:
Reopening. |
@draftcode, please investigate? |
I can reproduce this, but I cannot reproduce this with JGit. |
can someone give a quick fix for this? |
The reproduction case above only failed in production-like environments, so it took longer to track down. The next fix forgoes potential data transfer optimizations in the name of correctness. Unshallow operations will be full clones. We will work on reintroducing the optimizations but that is a multi-week project that won't begin immediately. The patch is in our release process, building and verifying. It should start rolling it out in a few hours. |
Sorry to say we hit a snag in the release process. We are working through it and will have a release ready later this afternoon. But releasing on a Friday afternoon is...dicey. We will wait until Monday to update the servers. Good news is that the fix bypasses all the finicky calculations that have caused this problem to recur, and we have verified it in a production-like environment. |
This was fully deployed by yesterday afternoon, PDT. |
Thanks @tparker and @draftcode for working on this. I'll close this issue now. If anyone is still seeing this, please let us know. |
@jayconrod Hi, I still see this issue:
I get error:
|
Does anyone have a full reproduction recipe? I'm unable to reproduce it. |
@anjmao I couldn't reproduce this. Are you seeing this consistently? Does this happen with an empty module cache (i.e., If you can give us a sequence of git commands that reproduces this (i.e., from |
What version of Go are you using (
go version
)?Does this issue reproduce with the latest release?
yes
What operating system and processor architecture are you using (
go env
)?go env
OutputWhat did you do?
What did you expect to see?
go modules downloaded and binary is built.
What did you see instead?
I could fix it by
then
go install
works as expected.https://bugs.chromium.org/p/chromium/issues/detail?id=951546
The text was updated successfully, but these errors were encountered: