Open yatima1460 opened 3 years ago
We are facing the same problem. We have linux runners.
We are facing the same problem. We have linux runners.
well then we can exclude ESET or the operating system could be an internal logic error or bad tolerance when network conditions are bad
in fact office network is not the greatest, AWS instead is top notch
I noticed a very strange problem related to actions/checkout@v2
Randomly it would get stuck forever at
"C:\Program Files\Git\cmd\git.exe" checkout --detach
when using a self hosted runner on WindowsIn the beginning I thought it would be caused by slow internet connection or a slow PC, so I left it running..
Turns out even after 6 hours it's still stuck and it seems got cancelled automatically by GitHub Actions
It's pretty random and I can't find any other info about this problem
Self hosted runner info:
Microsoft Windows [Version 10.0.19042.1052]
Git Version on self hosted runner:
git version 2.32.0.windows.1
Docker on self hosted version (shouldn't be relevant but still):
Docker version 20.10.7, build f0df350
ESET:
8.0.2028.0
Other notes: The self hosted runner has multiple GitHub runners installed in different directories
My only theory is that something locks some files inside the repo and git waits forever until they are unlocked
It doesn't seem to happen on a AWS EC2 instance, but only on a local office PC
So even if it's the self hosted runner's fault, antivirus installed or something else it's wise to add an internal timeout to actions checkout I think 🤔 with optional skipping the action that waits for locked files?