-
Notifications
You must be signed in to change notification settings - Fork 3.5k
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
[ci] Flaky Hexagon RPC server in tests #13205
Comments
@driazati Looks like it is trying to access this port |
if its all within the same container it should be since its just on the loopback address, is there any kind of random port selection or generation in the setup that's missing a |
@driazati but this function shouldn't allow |
@driazati I think this function has a bug. I'll send a PR to fix it |
#13207 might fix this issue |
Disabling random port numbers is a big productivity killer since I can no longer re-run pytest while the port is still in |
How about we only seed if |
@supersat I think it's not disabled, it will only start from the same port and increases the port number until it finds a valid port. wdyt? |
@mehrdadh I would say it's effectively disabled because it always tries the same sequence of port numbers. The tracker does come up on the next available port if the one it tries is taken, but then the launcher tries to get Seeding the RNG with a fixed value only in CI (with |
@driazati can we close this? |
the immediate issue is fixed so sure |
Seen on main in https://ci.tlcpack.ai/blue/organizations/jenkins/tvm/detail/main/4580/tests/, the RPC server name is also the same for every failing test and the tests all failed on the same shard
cc @Mousius @areusch @gigiblender @mehrdadh
The text was updated successfully, but these errors were encountered: