Reputation: 329
I'm trying to use keepalived to monitor some service. See my configuration file "keepalived.conf":
! Configuration File for keepalived
vrrp_script check_haproxy {
script "/usr/local/bin/check_haproxy.sh"
interval 2
fail 2
rise 2
}
vrrp_instance Vtest_2 {
interface eno16777736
track_interface {
eno16777736
}
state BACKUP
virtual_router_id 178
priority 100
nopreempt
unicast_src_ip 172.28.7.132
unicast_peer {
172.28.7.133
}
virtual_ipaddress {
195.221.2.14/32 dev eno16777736
}
track_script {
check_haproxy
}
notify /usr/local/bin/keepalived.state.sh
}
In my script i to compare the number of the pid file and the process of the service :
#!/bin/bash
# check if haproxy is running, return 1 if not.
# Used by keepalived to initiate a failover in case haproxy is down
process=""
HAPROXY_STATUS=""
if [ -f "/var/run/haproxy.pid" ]
then
process=$(cat /var/run/haproxy.pid)
HAPROXY_STATUS=$(ps -aux | grep -w "$process.*[h]aproxy")
fi
if [ "$HAPROXY_STATUS" ]
then
exit 0
else
exit 1
fi
My proble is that when i start keepalived when the command :
keepalived -f /etc/keepalived/keepalived.conf --dont-fork --log-console --log-detail
Everythings works goods keepalived failover works good. But when i start the service with systemd:
systemctl start keepalived
I have this lines in my "/var/log/messages" file:
Mar 15 17:05:51 localhost systemd: Starting Session 10 of user root.
Mar 15 17:06:23 localhost systemd: Starting LVS and VRRP High Availability Monitor...
Mar 15 17:06:23 localhost Keepalived[34182]: Starting Keepalived v1.2.13 (11/20,2015)
Mar 15 17:06:23 localhost systemd: Started LVS and VRRP High Availability Monitor.
Mar 15 17:06:23 localhost Keepalived[34183]: Starting Healthcheck child process, pid=34184
Mar 15 17:06:23 localhost Keepalived[34183]: Starting VRRP child process, pid=34185
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: Netlink reflector reports IP 172.28.7.132 added
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: Netlink reflector reports IP fe80::20c:29ff:fe22:a18e added
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: Registering Kernel netlink reflector
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: Registering Kernel netlink command channel
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: Registering gratuitous ARP shared channel
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: Opening file '/etc/keepalived/keepalived.conf'.
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: Configuration is using : 64365 Bytes
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: Using LinkWatch kernel netlink reflector...
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: VRRP_Instance(Vtest_2) Entering BACKUP STATE
Mar 15 17:06:23 localhost Keepalived_vrrp[34185]: VRRP sockpool: [ifindex(2), proto(112), unicast(1), fd(10,11)]
Mar 15 17:06:23 localhost Keepalived_healthcheckers[34184]: Netlink reflector reports IP 172.28.7.132 added
Mar 15 17:06:23 localhost Keepalived_healthcheckers[34184]: Netlink reflector reports IP fe80::20c:29ff:fe22:a18e added
Mar 15 17:06:23 localhost Keepalived_healthcheckers[34184]: Registering Kernel netlink reflector
Mar 15 17:06:23 localhost Keepalived_healthcheckers[34184]: Registering Kernel netlink command channel
Mar 15 17:06:23 localhost Keepalived_healthcheckers[34184]: Opening file '/etc/keepalived/keepalived.conf'.
Mar 15 17:06:23 localhost Keepalived_healthcheckers[34184]: Configuration is using : 5344 Bytes
Mar 15 17:06:23 localhost Keepalived_healthcheckers[34184]: Using LinkWatch kernel netlink reflector...
Mar 15 17:06:26 localhost Keepalived_vrrp[34185]: VRRP_Instance(Vtest_2) Now in FAULT state
Anyone have an idea ?
Upvotes: 2
Views: 10861
Reputation: 16
I had a similar issue with my track_script
not running when starting keepalived as a systemctl service.
I found out that putting my track_script
in /usr/libexec/keepalived
solved the issue. It is probably related to this issue.
! Configuration File for keepalived
global_defs {
}
vrrp_script odr_check {
script "/usr/libexec/keepalived/health_check.sh"
interval 2
weight 50
fall 1
rise 1
}
vrrp_instance VI_1 {
state BACKUP
interface eno16777984
virtual_router_id 42
priority 90
track_script {
odr_check
}
virtual_ipaddress {
192.168.110.84
}
}
Upvotes: 0
Reputation: 155
I'm a using much simple configuration for the same purpose. Utility pidof must be on most OSs Linux by default. Try this for example:
vrrp_script check_haproxy
{
script "pidof haproxy"
interval 2
weight 2
}
It's a working solution buddy.
Upvotes: 1