BUG/MINOR: mworker: fix -D -W -sf/-st modes

When a new master process is launched like below:

	./haproxy -W -D -p ha.pid -sf $(cat ha.pid)...

The old master process and its workers do not stop. Since the master-worker
refactoring, the code, which sends USR1/TERM to old pids from -sf, is called
only for the standalone mode. In master-worker mode we should receive the READY
message from the newly forked worker at first, in order to be able to terminate
the previous master.

So, to fix this, let's terminate the previous master in _send_status(), where
we parse the READY message from the newly forked worker. And let's continue to
use oldpids array, as it was in 3.0, in order to stop the workers, launched
before the reload.

This patch should be backported only in 3.1.
This commit is contained in:
Valentine Krasnobaeva 2024-12-03 21:13:47 +01:00 committed by William Lallemand
parent 1fead6c0ca
commit ee111d2004

View File

@ -2518,14 +2518,6 @@ static int _send_status(char **args, char *payload, struct appctx *appctx, void
}
}
/* stop previous worker process, if it wasn't signaled during max reloads check */
list_for_each_entry(proc, &proc_list, list) {
if ((proc->options & PROC_O_TYPE_WORKER) &&
(proc->options & PROC_O_LEAVING) &&
(proc->reloads >= 1)) {
kill(proc->pid, oldpids_sig);
}
}
/* At this point we are sure, that newly forked worker is started,
* so we can write our PID in a pidfile, if provided. Master doesn't
@ -2534,11 +2526,19 @@ static int _send_status(char **args, char *payload, struct appctx *appctx, void
if (global.pidfile != NULL)
handle_pidfile();
/* either send USR1/TERM to old master, case when we launched as -W -D ... -sf $(cat pidfile),
* or send USR1/TERM to old worker processes.
*/
if (nb_oldpids > 0) {
nb_oldpids = tell_old_pids(oldpids_sig);
}
load_status = 1;
ha_notice("Loading success.\n");
if (global.tune.options & GTUNE_USE_SYSTEMD)
sd_notifyf(0, "READY=1\nMAINPID=%lu\nSTATUS=Ready.\n", (unsigned long)getpid());
return 1;
}