Closed ashmind closed 6 years ago
Not sure what's going om here, but see http://blogs.suwatbodin.com/2013/07/deploy-to-windows-azure-using-git-with.html for some general guidance related to using submodules .
Also, I was not able to repro this. I tried this:
git clone https://github.com/ashmind/SharpLab.git
git clone https://github.com/ashmind/SharpLab.git
git submodule sync
. It just says Synchronizing submodule url for 'source/#external/ILSpy'
, with no errors.@davidebbo Thanks for the research! Yes, same code worked fine before, and still works fine on local.
I googled the exit code from the error message (0xC0000142
) and it seems to be "The application was unable to start correctly", which kind of makes sense -- git submodule sync
tries to start something using fork, but it fails to start.
Not sure why it fails to start -- might be something with a specific system?
https://answers.microsoft.com/en-us/windows/forum/windows_10-performance-winpc/error-0xc0000142/3da27a75-178c-4de5-b43c-c538e6005540?auth=1
Can you try in other Web Apps to see if this seems specific to just this one app? That will help isolate. If it comes down to just one app, we can take a look at it and see what might be going on.
@davidebbo Sure -- same happens to main sharplab
(sharplab.io) deployments/SCM.
Also tried sl-b-dotnet-features-dataflow
through SCM (it doesn't have deployments) -- same issue when running git submodule sync
in any folder.
So you mean that running git submodule sync
fails this way in any folder that's not even a git repo? When I tried that in my app, I get:
fatal: Not a git repository (or any of the parent directories): .git
Yep, surprising, but fail on 0xC0000142
in any directory -- probably because it fails before it can do any checks. Sorry, I have to leave for work, so further answers might be a bit delayed.
I moved you to a different VM, which solved it. But this is certainly puzzling. If you see this again, let me know and we'll dig deeper next time.
Thanks! I'll definitely let you know if I see it again.
Ok, closing for now but we can reopen as needed.
Sorry to comment in a closed issue, but we also had this problem in Microsoft Azure since last friday (for no apparent reason), which uses Kudu. We were pushing with git and the server was receiving the commits correctly, but then auto deployment via git never worked and didn't update our website. It's still failing, actually, but we uploaded manually via FTP.
We were getting errors when "git push"ing:
Counting objects: 25, done.
Delta compression using up to 16 threads.
Compressing objects: 100% (22/22), done.
Writing objects: 100% (25/25), 413.92 KiB | 8.12 MiB/s, done.
Total 25 (delta 12), reused 0 (delta 0)
remote: 2 [main] sh 7056 fork: child -1 - forked process 6500 died unexpectedly, retry 0, exit code 0xC0000142, errno 11
remote: hooks/post-receive: fork: retry: Resource temporarily unavailable
remote: 1094002 [main] sh 7056 fork: child -1 - forked process 6472 died unexpectedly, retry 0, exit code 0xC0000142, errno 11
remote: hooks/post-receive: fork: retry: Resource temporarily unavailable
remote: 3180523 [main] sh 7056 fork: child -1 - forked process 7564 died unexpectedly, retry 0, exit code 0xC0000142, errno 11
remote: hooks/post-receive: fork: retry: Resource temporarily unavailable
remote: 7246861 [main] sh 7056 fork: child -1 - forked process 4892 died unexpectedly, retry 0, exit code 0xC0000142, errno 11
remote: hooks/post-receive: fork: retry: Resource temporarily unavailable
remote: 15312972 [main] sh 7056 fork: child -1 - forked process 6948 died unexpectedly, retry 0, exit code 0xC0000142, errno 11
remote: hooks/post-receive: fork: Resource temporarily unavailable
To https://test-server-website.scm.azurewebsites.net:443/test-server-website.git
9097984..498db46 master -> master
I don't know if this relates to Kudu, but I thought I would let you know. This link also has more details of the same issue, and someone seems to have narrowed it to sh.exe failing.
In our case, following these instructions yields the error, without even requiring git commands:
Go to Kudu Console. e.g. https://{yourapp}.scm.azurewebsites.net/DebugConsole
Run: echo ls | sh
@CanisLupus it could be that your VM got into that same unexplained state. If you share your site name, and can take a look.
Or if you need this fix asap, just scale up and back down to move VM (e.g. Small -> Medium -> Small, or vice versa).
Hey, David, thanks for replying. I just created a new App with the name "kudu-test". The real site is the one that begins with "c" and ends in "www".
Unfortunately, we tried the option of scaling up and down before (waiting a few minutes between the two), and it didn't fix the problem, though we might have done something incorrectly. We managed to update our website manually by uploading everything via ftp.
Is it just a matter of going to "Scale up" and moving, for example, from S1 to S2 and then back again?
Thanks for looking into this. Daniel
@CanisLupus sorry, just saw this. Not sure why scale up / down didn't fix it. I initiated a reboot and now it seems fine.
Hey @davidebbo, we're eternally grateful to you. ;) Git deployment is working fine now. Thank you so much!
@CanisLupus Glad that worked. It's a very strange condition with the git tools, and it is not well understood.
Repro steps.
Pushed update into my GitHub repo, Kudu source refresh failed on
git submodule sync
(before any other build steps). Same command fails when run through SCM.Same command works OK on local.
Azure GIT: 2.14.1.windows.1 Local GIT: 2.8.1.windows.1
Project structures.
https://github.com/ashmind/SharpLab/tree/edge
The log/error given by the failure.
Deployment Log: Updating submodules
SCM (Manual):
Debug your Azure website remotely.
Site name: sharplab-edge (https://edge.sharplab.io)