Repository navigation
[LEGACY] [CI]: TestRestartWithSignal does not seem to (always?) work correctly (EL8 only?) #4068
Description
Description
Seen on #4056
See attached trace (note that PR #4056 has the new timeout logic).
Signal has been sent and received, but the container has not restarted (and was instead terminated by timeout).
@AkihiroSuda the good news here is that the new timeout logic is working as expected, and this test would have previously hung-up on EL8.
Now we just have to figure out what is happening here.
Steps to reproduce the issue
Describe the results you received and expected
=== FAIL: cmd/nerdctl/container TestRestartWithSignal (21.97s)
container_restart_linux_test.go:163:
container_restart_linux_test.go:163: ======================== Pre-test cleanup ========================
container_restart_linux_test.go:163:
container_restart_linux_test.go:163: ======================== Test setup ========================
container_restart_linux_test.go:163:
container_restart_linux_test.go:163: ======================== Test Run ========================
command.go:435: [2025-04-01T16:46:36Z] [command=/usr/local/bin/nerdctl --namespace=nerdctl-test run --name testrestartwithsignal-27bfb0e4 ghcr.io/stargz-containers/alpine:3.13-org sh -c #!/bin/sh
set -eu
sig_msg () {
printf "received\n"
[ "false" != true ] || exit 0
}
trap sig_msg 10
printf "trap ready\n"
while true; do
printf "waiting...\n"
sleep 0.5
done
] command cancelled
command.go:115: expected: 137 - to be equal to: -1
command.go:115: Expected exit code: 137
command.go:115:
=================================
| Command: /usr/local/bin/nerdctl run --name testrestartwithsignal-27bfb0e4 ghcr.io/stargz-containers/alpine:3.13-org sh -c #!/bin/sh
set -eu
sig_msg () {
printf "received\n"
[ "false" != true ] || exit 0
}
trap sig_msg 10
printf "trap ready\n"
while true; do
printf "waiting...\n"
sleep 0.5
done
| Working Dir: /tmp/TestRestartWithSignal2256824605/001
| Timeout: 20s
=================================
SHELL=/bin/bash
LOGNAME=rootless
XDG_SESSION_TYPE=tty
HOME=/home/rootless
LANG=C.UTF-8
SSH_CONNECTION=::1 41538 ::1 22
XDG_SESSION_CLASS=user
IPFS_PATH=/home/rootless/.local/share/ipfs
USER=rootless
SHLVL=1
XDG_SESSION_ID=4
XDG_RUNTIME_DIR=/run/user/1001
SSH_CLIENT=::1 41538 22
DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/1001/bus
OLDPWD=/home/rootless
_=/usr/local/bin/gotestsum
PATH=/usr/local/go/bin:/usr/local/go/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin
***
DOCKER_CONFIG=/tmp/TestRestartWithSignal2256824605/001
NERDCTL_TOML=/tmp/TestRestartWithSignal2256824605/001/nerdctl.toml
=================================
| Stderr:
=================================
=================================
| Stdout:
=================================
trap ready
waiting...
waiting...
waiting...
received
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
waiting...
=================================
| Exit Code: -1
| Signaled: killed
| Err: command timed out
=================================
case.go:180:
case.go:181: ======================== Post-test cleanup ========================
What version of nerdctl are you using?
main
Are you using a variant of nerdctl? (e.g., Rancher Desktop)
None
Host information
I WANT A GREEN CI AND WILL NOT STOP BEING ENRAGED UNTIL WE GET IT 😡🤣
Activity
apostasie commentedon Apr 1, 2025
Maybe EL8 is just really really really slow and we need to extend the timeout from 20 seconds to 30?
AkihiroSuda commentedon Apr 18, 2025
Maybe yes, as EL 8 is executed on a nested virt
apostasie commentedon Apr 18, 2025
Ok, let's test the hypothesis.
We can merge it - it is rather innocuous. Worse case scenario, it will still fail on EL8, just slower...
#4130
[-][CI]: TestRestartWithSignal does not seem to (always?) work correctly (EL8 only?)[/-][+][LEGACY] [CI]: TestRestartWithSignal does not seem to (always?) work correctly (EL8 only?)[/+]