`dora` often gets stuck in unfixable state

[alberthli]

Describe the bug
Oftentimes, after hours of development, I will suddenly be unable to start the Dora daemon. Seemingly, the only way I can fix this problem is by rebooting my computer - not even reinstalling works.

The behavior is as follows. If I run

dora check

I get the output

Dora Coordinator: ok


[ERROR]
failed to send DaemonConnected message

Caused by:
    failed to fill whole buffer

Location:
    binaries/cli/src/command/check.rs:68:10

dora up similarly raises failed to send DaemonConnected message, while dora destroy outputs

[ERROR]
failed to send destroy message

Caused by:
    failed to fill whole buffer

Location:
    binaries/cli/src/command/up.rs:60:18

dora daemon shows

2025-07-08T19:32:39.729498Z  WARN dora_daemon::coordinator: Could not connect to: 127.0.0.1:53290, with error: failed to connect to dora-coordinator. Retring in 1s..
2025-07-08T19:32:40.730895Z  WARN dora_daemon::coordinator: Could not connect to: 127.0.0.1:53290, with error: failed to connect to dora-coordinator. Retring in 1s..
2025-07-08T19:32:41.732246Z  WARN dora_daemon::coordinator: Could not connect to: 127.0.0.1:53290, with error: failed to connect to dora-coordinator. Retring in 1s..
2025-07-08T19:32:42.734533Z  WARN dora_daemon::coordinator: Could not connect to: 127.0.0.1:53290, with error: failed to connect to dora-coordinator. Retring in 1s..
# repeats forever...

dora coordinator shows

Listening for incoming daemon connection on 53290
2025-07-08T19:33:29.526054Z ERROR dora_coordinator: failed to listen for control messages

Caused by:
    Address already in use (os error 98)

Location:
    /home/runner/work/dora/dora/binaries/coordinator/src/control.rs:46:10

This suggests that there's some stale connection that I need to close, but I can't figure out how to do so. Any pointers?

To Reproduce
I cannot reliably reproduce this bug. It is extremely flaky.

Expected behavior
I should be able to run all of the above commands or easily understand how to fix the state of dora.

Environments (please complete the following information):

• System info:

Linux <my_computer> 6.11.0-29-generic #29~24.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Jun 26 14:16:59 UTC 2 x86_64 x86_64 x86_64 GNU/Linux

• Dora version: dora-cli 0.3.12

[haixuanTao]

Yes indeed!

I actually don't know what is happening. I genuinely believe we need to fix it.

I couldn't work on it too much before of ongoing PR but will try to look into it

[haixuanTao]

I think it's a coordinator issue

[haixuanTao]

Would love your input on it as well @phil-opp if you have any idea what could be it?

[alberthli]

I appreciate your interest in solving this issue - it's been a huge thorn in my side, because many people use the machine I'm on, so rebooting it frequently is not a good solution. Is there no workaround you can see? For example, clearing some variable, cache, manually closing some connection, etc.?