Skip to content

fix: drop a dead shared iroh connection before reusing it - #134

Merged
ado11231 merged 3 commits into
masterfrom
fix/stale-iroh-share
Oct 16, 2026
Merged

ado11231 merged 3 commits into
masterfrom
fix/stale-iroh-share

Conversation

@ado11231

Copy link
Copy Markdown
Owner

Closes #128.

  • Before the first reuse of the shared ssh connection over iroh, a forward to a closed port goes through it. No answer in 3 seconds closes it.
  • After kill -9 on the Agent, the next run took 29.4s before and 3.8s after.
  • Commands back to back over iroh pay about 0.14s. Local network and tailnet runs are unchanged.

After an Agent crash, the shared ssh connection over iroh stayed open until iroh's 30 second idle timeout, and the next command waited with it. Before the first reuse, a forward to a closed port goes through it, which the Agent refuses at once. A connection that does not answer within 3 seconds is closed. Commands back to back over iroh pay about 0.14 seconds. Short network drops still do not end a run.
The roadmap records the crash test and drops the old 20 second limitation.
@ado11231
ado11231 merged commit 1b0d376 into master Oct 16, 2026
2 checks passed
@ado11231
ado11231 deleted the fix/stale-iroh-share branch October 18, 2026 04:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Reconnect over iroh faster after the Agent crashes

1 participant