Yu Watanabe
Yu Watanabe
superseded by #31502.
> @yuwata Currently we use `NLM_F_CREATE | NLM_F_EXCL` for any netlink message that creates a new object. Isn't it as simple as switching to `NLM_F_CREATE | NLM_F_REPLACE` instead to replace...
Fix is waiting in #33944.
> > Bind-mounting these sockets looks nice. But I am not sure we should introduce the new knob. Why not to do that gracefully? > > This is already mentioned...
Another failure: https://github.com/systemd/systemd/actions/runs/9227184474/job/25388567140?pr=33010 ``` [ 102.612678] TEST-46-HOMED.sh[470]: + wait_for_state homedsshtest inactive [ 102.616943] TEST-46-HOMED.sh[470]: + for i in {1..10} [ 102.619418] TEST-46-HOMED.sh[470]: + (( i > 1 )) [ 102.621764]...
Huh, still failed: https://github.com/systemd/systemd/actions/runs/9183330348/job/25256917005?pr=32964 ``` [ 13.619340] TEST-60-MOUNT-RATELIMIT.sh[503]: + journalctl --sync [ 13.767039] TEST-60-MOUNT-RATELIMIT.sh[503]: + timeout 2m bash -c 'until journalctl -u init.scope --since=01:26:46 | grep -q '\''(mount-monitor-dispatch) entered rate...
Huh, still failed: https://github.com/systemd/systemd/actions/runs/9189842613/job/25272756816 ``` [ 13.225589] TEST-60-MOUNT-RATELIMIT.sh[508]: + journalctl --sync [ 13.238213] TEST-60-MOUNT-RATELIMIT.sh[662]: + timeout 2m journalctl -u init.scope --since=11:43:56 -n all --follow [ 13.243660] TEST-60-MOUNT-RATELIMIT.sh[663]: + grep -m...
Hmm, this seems not a issue in journald. ``` $ journalctl --file TEST-60-MOUNT-RATELIMIT.journal -u init.scope --no-hostname | grep 'rate limit' May 22 20:43:47 systemd[1]: Event source 0x55b262896670 (mount-monitor-dispatch) left rate...