Here are some details:
$ uname -a
OpenBSD fw-1.my.domain 4.2 GENERIC.MP#1 i386
Hoststated.conf:
ext_if="bge0"
web_1="10.1.1.10"
web_1_1="10.1.1.40"
web_2="10.1.1.11"
web_2_1="10.1.1.41"
virtual_web="public.IP.1"
virtual_web_1="public.IP.2"
interval 30
timeout 300
table http_table {
check http "/" code 302
real port 80
host $web_1 retry 2
}
table https_table {
check https "/" code 200
real port 443
host $web_1 retry 2
}
table https_table_1 {
check https "/test-web_1.html" code 200
real port 443
host $web_1_1 retry 2
}
table backup_http_table {
check http "/" code 302
real port 80
host $web_2 retry 2
}
table backup_https_table {
check https "/" code 200
real port 443
host $web_2 retry 2
}
table backup_https_table_1 {
check https "/test-web_1.html" code 200
real port 443
host $web_2_1 retry 2
}
service http_service {
virtual host $virtual_web port 80 interface $ext_if
table http_table
backup table backup_http_table
}
service https_service {
virtual host $virtual_web port 443 interface $ext_if
table https_table
backup table backup_https_table
}
service https_service_1 {
virtual host $virtual_web_1 port 443 interface $ext_if
table https_table_1
backup table backup_https_table_1
}
Again, I have two boxes with exactly the same configuration running
CARP. And, as I stated before, that load balancing configuration runs
like for a week, then start seeing web_1 and web_1_1 down. Sometime, it
starts using web_2 and web_2_1, or sometimes sees them down, as well.
However, I never saw any stability problem with both web_1 and web_1_1.
So, when such a thing happens, I just reboot openBSD boxes without
touching anything on both web_1 and web_1_1. Then, all comes up fine (at
least for a week again).
Thanks for any help and suggestion!
Rami
