Greetings,
I patched squid2.6 stable 14 with the tcp probe patch.
It patched two files:
cache_cf.c
neighbors.c
However, After about 14 hours of good runtime, my response times,
began to suck, and began to see errors again indicative of the tcp
probe issue:
2007/10/02 01:57:15| Detected REVIVED Parent: 10.10.10.20
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
2007/10/02 01:57:16| Detected DEAD Parent: 10.10.10.20
2007/10/02 01:57:16| TCP connection to 10.10.10.20/80 failed
The origin server is available, however I keep getting
revivied/connectionfailed/dead
It seems that the only way to recover from this is a restart.
I am running solaris 10, and I had to download the gnu patch utility
in order to patch the src.
Here was the patch applied.
Index: src/cache_cf.c
===
RCS file: /cvsroot/squid/squid/src/cache_cf.c,v
retrieving revision 1.470
diff -u -p -r1.470 cache_cf.c
--- src/cache_cf.c 20 Jul 2007 21:08:47 - 1.470
+++ src/cache_cf.c 28 Aug 2007 23:46:47 -
@@ -1621,6 +1621,7 @@ parse_peer(peer ** head)
p-stats.logged_state = PEER_ALIVE;
p-monitor.state = PEER_ALIVE;
p-monitor.interval = 300;
+p-tcp_up = PEER_TCP_MAGIC_COUNT;
if ((token = strtok(NULL, w_space)) == NULL)
self_destruct();
p-host = xstrdup(token);
Index: src/neighbors.c
===
RCS file: /cvsroot/squid/squid/src/neighbors.c,v
retrieving revision 1.318
diff -u -p -r1.318 neighbors.c
--- src/neighbors.c 20 Jul 2007 21:08:47 - 1.318
+++ src/neighbors.c 28 Aug 2007 23:46:47 -
@@ -1010,12 +1010,13 @@ peerDNSConfigure(const ipcache_addrs * i
debug(0, 0) (WARNING: No IP address found for '%s'!\n, p-host);
return;
}
-p-tcp_up = PEER_TCP_MAGIC_COUNT;
for (j = 0; j (int) ia-count j PEER_MAX_ADDRESSES; j++) {
p-addresses[j] = ia-in_addrs[j];
debug(15, 2) (-- IP address #%d: %s\n, j, inet_ntoa(p-addresses[j]))
;
p-n_addresses++;
}
+if (!p-tcp_up)
+ peerProbeConnect((peer *) p);
ap = p-in_addr;
memset(ap, '\0', sizeof(struct sockaddr_in));
ap-sin_family = AF_INET;
Any ideas is much appreciated. Any special debug info you need, please
let me know.
Also, as I side note, I have monitorurl set as well
cache_peer 10.10.10.20 parent 80 0 no-query no-digest originserver
monitorinterval=30 monitorurl=http://10.10.10.20/test.jpg
Thank you!
