Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I am configuring muli-master replication manager for our master-master-slave replication system. Versions of the replication system are, MySQL5.5.15 PHP 5.3.8 MySQL- MMM 2.2.1.

The configuration file is: /etc/mysql-mmm/mmm_common.conf

active_master_role      writer
<host default>
    cluster_interface           eth0
    pid_path                    /var/run/mmm_agentd.pid
    bin_path                    /usr/lib/mysql-mmm/
    replication_user        mmm_replication
    replication_password    password
    agent_user              mmm_agent
    agent_password          password
</host>
<host node5>
        ip              49.212.101.135
        mode            master
        peer            node6
</host>
<host node6>
        ip              49.212.101.136
        mode            master
        peer            node5
</host>
<host node7>
        ip              49.212.101.137
        mode            slave
</host>
<role writer>
        hosts           node5,node6
        ips             49.212.101.140
        mode            exclusive
</role>
<role reader>
        hosts           node5,node6,node7
        ips             49.212.101.132,49.212.101.133,49.212.101.139
        mode            balanced
</role>

Following is the status of mmm_control checks all command;

[root@node4 ~]# mmm_control checks  all
node7  ping         [last change: 2011/09/24 20:20:16]  OK
node7  mysql        [last change: 2011/09/24 20:20:16]  OK
node7  rep_threads  [last change: 2011/09/24 20:20:16]  OK
node7  rep_backlog  [last change: 2011/09/24 20:20:16]  OK: Backlog is null
node5  ping         [last change: 2011/09/24 20:20:16]  OK
node5  mysql        [last change: 2011/09/24 20:20:16]  OK
node5  rep_threads  [last change: 2011/09/24 20:20:16]  OK
node5  rep_backlog  [last change: 2011/09/24 20:20:16]  OK: Backlog is null
node6  ping         [last change: 2011/09/24 20:20:16]  OK
node6  mysql        [last change: 2011/09/24 20:20:16]  OK
node6  rep_threads  [last change: 2011/09/24 20:20:16]  OK
node6  rep_backlog  [last change: 2011/09/24 20:20:16]  OK: Backlog is null

And following is the result of mmm_control show command;

[root@node4 ~]# mmm_control show
  node5(49.212.101.135) master/ONLINE. Roles: reader(49.212.101.133), writer(49.212.101.140)
  node6(49.212.101.136) master/ONLINE. Roles: reader(49.212.101.132)
  node7(49.212.101.137) slave/ONLINE. Roles: reader(49.212.101.139)

My question is for writer vip(49.212.101.140), i.e, I do ping to the writer vip ,but fails,, Why???

[root@node4 ~]# ping 49.212.101.140
PING 49.212.101.140 (49.212.101.140) 56(84) bytes of data.
From 49.212.101.134 icmp_seq=1 Destination Host Unreachable

As for reader vip, pinging is OK;

[root@node4 ~]# ping 49.212.101.133
PING 49.212.101.133 (49.212.101.133) 56(84) bytes of data.
64 bytes from 49.212.101.133: icmp_seq=1 ttl=64 time=3.02 ms
64 bytes from 49.212.101.133: icmp_seq=2 ttl=64 time=1.35 ms

Ping to the writer vip is failed,why?? However, ping to the reader vips is OK.

Please advise me whats wrong. Thank you sir.

Added following test status later;

As for the above bug(?), I tried followings commands.

[root@node4 ~]# mmm_control move_role writer node6

OK: Role 'writer' has been moved from 'node5' to 'node6'. Now you can wait some time and check new roles info!

[root@node4 ~]# mmm_control show node5(49.212.101.135) master/ONLINE. Roles: reader(49.212.101.132) node6(49.212.101.136) master/ONLINE. Roles: reader(49.212.101.139), writer(49.212.101.140) node7(49.212.101.137) slave/ONLINE. Roles: reader(49.212.101.133)

[root@node4 ~]# ping 49.212.101.140 PING 49.212.101.140 (49.212.101.140) 56(84) bytes of data. 64 bytes from 49.212.101.140: icmp_seq=1 ttl=64 time=2.05 ms 64 bytes from 49.212.101.140: icmp_seq=2 ttl=64 time=1.09 ms 64 bytes from 49.212.101.140: icmp_seq=3 ttl=64 time=1.52 ms

As for node6 writer pinging is OK, However when changing writer node to node5 #mmm_control move-role node5, mmm_control show status seems Ok, but pinging is not reachable;

[root@node4 ~]# mmm_control move_role writer node5

OK: Role 'writer' has been moved from 'node6' to 'node5'. Now you can wait some time and check new roles info!

[root@node4 ~]# mmm_control show node5(49.212.101.135) master/ONLINE. Roles: reader(49.212.101.132), writer(49.212.101.140) node6(49.212.101.136) master/ONLINE. Roles: reader(49.212.101.139) node7(49.212.101.137) slave/ONLINE. Roles: reader(49.212.101.133)

root@node4 ~]# ping 49.212.101.140 PING 49.212.101.140 (49.212.101.140) 56(84) bytes of data. From 49.212.101.134 icmp_seq=2 Destination Host Unreachable

I think this seems to me a fatal bug of mmm2.2.1.

share|improve this question
    
I think mmm2.2.1 has a bug that can't set writer vip under some conditions. So I set writer vip manually; ip addr add 49.212.101.140/32 dev eth0 arping -c 2 -U -I eth0 -s 49.212.101.140 49.212.101.129 ip a s ping 49.212.101.129 –  nn2480 Oct 5 '11 at 1:09
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.