BUG/MEDIUM: master/cli: Pin the master CLI on the first thread of the group 1

There is no reason to start the master CLI on several threads and on several
groups. And in fact, it must not be done otherwise the same FD is inserted
several times in the fdtab, leading to a crash during startup because of a
BUG_ON(). It happens when several groups are configured.

To fix the bug the master CLI is now pinned on the first thread of the first
group.

This patch should fix the issue #2259 and must be backported to 2.8.

diff --git a/src/cli.c b/src/cli.c
index ec2fc91fb829a006044223fb94f09a56b4bb4924..0a78ed7807338cbe90da8872a4a92be51426cecf 100644 (file)
--- a/src/cli.c
+++ b/src/cli.c
@@ -3244,6 +3244,9 @@ int mworker_cli_sockpair_new(struct mworker_proc *mworker_proc, int proc)
        bind_conf->nice = -64;  /* we want to boost priority for local stats */
        bind_conf->options |= BC_O_UNLIMITED | BC_O_NOSTOP;
 
+       /* Pin master CLI on the first thread of the first group only */
+       thread_set_pin_grp1(&bind_conf->thread_set, 1);
+
        list_for_each_entry(l, &bind_conf->listeners, by_bind) {
                HA_ATOMIC_INC(&unstoppable_jobs);
                /* it's a sockpair but we don't want to keep the fd in the master */