mirrors/haproxy
1
0
Fork 0
mirror of https://git.haproxy.org/git/haproxy.git/ synced 2025-08-13 10:37:02 +02:00
Code Issues Projects Releases Wiki Activity

BUG/MEDIUM: listener: don't report an error when resuming unbound listeners

Browse Source 
Pavlos Parissis reported that a sequence of disable/enable on a frontend
performed on the CLI can result in an error if the frontend has several
"bind" lines each bound to different processes. This is because the
resume_listener() function returns a failure for frontends not part of
the current process instead of returning a success to pretend there was
no failure.

This fix should be backported to 1.5.
This commit is contained in:
Willy Tarreau 2015-04-14 12:07:16 +02:00
parent 10146c9c51
commit af2fd584f3
1 changed files with 5 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
src/listener.c
View File

@ -121,9 +121,10 @@ int pause_listener(struct listener *l)
* may replace enable_listener(). The resulting state will either be LI_READY
* or LI_FULL. 0 is returned in case of failure to resume (eg: dead socket).
* Listeners bound to a different process are not woken up unless we're in
* foreground mode. If the listener was only in the assigned state, it's totally
* rebound. This can happen if a pause() has completely stopped it. If the
* resume fails, 0 is returned and an error might be displayed.
* foreground mode, and are ignored. If the listener was only in the assigned
* state, it's totally rebound. This can happen if a pause() has completely
* stopped it. If the resume fails, 0 is returned and an error might be
* displayed.
*/
int resume_listener(struct listener *l)
{
@ -147,7 +148,7 @@ int resume_listener(struct listener *l)
if ((global.mode & (MODE_DAEMON | MODE_SYSTEMD)) &&
l->bind_conf->bind_proc &&
!(l->bind_conf->bind_proc & (1UL << (relative_pid - 1))))
return 0;
return 1;
if (l->proto->sock_prot == IPPROTO_TCP &&
l->state == LI_PAUSED &&