You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
noting for future reference: i found when running this script locally that the trap commands herein weren't working as expected; cometbft and pd were sometimes left running, but i haven't tracked down what causes the kill call to fail. 🤨
this squashes two related commits from development:
---
ci: ⭐ `SIGINT` will also kill cometbft, pd, summonerd
when running this locally in a development shell, i found that `ctrl+c`
would not terminate the other running processes. this would stop the
shell script, but logs from pd and cometbft would continue being printed
to the screen.
moreover, after stopping that shell, `ps` showed me that these processes
were left running, causing subsequent runs of the summonerd smoke test
to fail due to ports already being bound.
this adds `SIGINT` to the list of signals that will bring down the
cometbft, pd, and summonerd processes. this way, `ctrl+c` will properly
clean things up.
---
ci: ➕ don't clobber smoke test's child pids
fixes#4370.
this makes the list of process id's provided to `kill -9` when a SIGEXIT
or SIGINT signal is received an _additive_ list. prior, we were
clobbering this list when the phase1 and phase2 steps were run.
now, we append a new id to the list at each point we set a signal
trap.
originally posted by @cratelyn in #4358 (comment)
The text was updated successfully, but these errors were encountered: