Hi Lukas,
On Sat, Dec 22, 2018 at 11:16:09PM +0100, Lukas Tribus wrote:
> Hello Oliver,
>
>
> redispatch is broken since commit 25b401536 ("BUG/MEDIUM: connection:
> Just make sure we closed the fd on connection failure"). It simply
> fails to connect to the next server.
>
> 1.9 is affected.
>
>
> Repro:
>
> global
> log 10.0.0.4:514 len 65535 local1 debug
> maxconn 1000
>
> defaults
> log global
> mode http
> option httplog
> timeout connect 1s
> timeout client 30s
> timeout server 30s
> retries 3
> option redispatch 1
>
> frontend http-in
> bind :80
> default_backend mybak
>
> backend mybak
> mode http
> balance first
> server primary-fail 10.0.0.199:80 # this server is unreachable
> server backup-ok 10.0.0.254:80
>
>
>
>
> cheers,
Ooops you're right indeed. The attached patch should fix it.
Thanks a lot for reporting !
Regards,
Olivier
>From 2276c53dac820d0079525730e9bd7abfd3ea408c Mon Sep 17 00:00:00 2001
From: Olivier Houchard
Date: Mon, 24 Dec 2018 13:32:13 +0100
Subject: [PATCH] BUG/MEDIUM: servers: Don't try to reuse connection if we
switched server.
In connect_server(), don't attempt to reuse the old connection if it's
targetting a different server than the one we're supposed to access, or
we will never be able to connect to a server if the first one we tried failed.
This should be backported to 1.9.
---
src/backend.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/backend.c b/src/backend.c
index 2407f8a32..bc38c5710 100644
--- a/src/backend.c
+++ b/src/backend.c
@@ -1129,7 +1129,7 @@ int connect_server(struct stream *s)
srv_cs = objt_cs(s->si[1].end);
if (srv_cs) {
old_conn = srv_conn = cs_conn(srv_cs);
- if (old_conn) {
+ if (old_conn && (!old_conn->target || old_conn->target ==
s->target)) {
old_conn->flags &= ~(CO_FL_ERROR | CO_FL_SOCK_RD_SH |
CO_FL_SOCK_WR_SH);
srv_cs->flags &= ~(CS_FL_ERROR | CS_FL_EOS |
CS_FL_REOS);
reuse = 1;
--
2.19.2
