Squid DEAD PARENT problem

Dear squiders,

After a long successful working my Squid-1.0.18 on FreeBSD machine confugured
as used behind firewall begun to mark parents as dead when a single external
link is down. Link is down for a short time frequently enough
but it's interesting that usually
[10/Nov/1996:15:48:11 +0300] comm.c:372: connect: 193.233.9.14:3128: (60) Operation timed out.
[10/Nov/1996:17:29:49 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.
[10/Nov/1996:17:29:49 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.
[10/Nov/1996:19:23:49 +0300] comm.c:372: connect: 193.233.9.14:3128: (60) Operation timed out.
[10/Nov/1996:20:19:37 +0300] comm.c:372: connect: 193.233.9.14:3128: (60) Operation timed out.
[11/Nov/1996:11:08:40 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.
[11/Nov/1996:11:08:53 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.
[11/Nov/1996:13:37:57 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.
[11/Nov/1996:13:38:07 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.
[11/Nov/1996:15:09:12 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.
[11/Nov/1996:15:09:47 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.
[11/Nov/1996:15:10:25 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.
[11/Nov/1996:15:10:32 +0300] comm.c:372: connect: 193.233.9.14:3128: (65) No route to host.

but simetimes
[11/Nov/1996:15:11:15 +0300] neighbors.c:563: neighborsUdpPing: Detected DEAD PARENT: hawk.landau.ac.ru
[11/Nov/1996:15:11:15 +0300] neighbors.c:563: neighborsUdpPing: Detected DEAD PARENT: webcache1.free.net

and only reload (by "kill -HUP" or "kill -KILL") forces "revive" a parent
  
Can anybody help me

Serge Krashakov
Received on Mon Nov 11 1996 - 04:37:50 MST