mmm-monitor failed to start, no error in logs

[nemd]

Hello,

I have problem with mmm-mond. It seems to not start. After proper
configuration (tested on another machine) i'm doing service mysql-mmm-
monitor start, and after that i can see only one process listed (in
ps):

9197 ? S 0:00 mmm_mond
9198 ? S 0:00 \_ mmm_mond
9199 ? S 0:00 \_ perl /usr/libexec/mysql-mmm/monitor/
checker ping_ip

After turning on debug log (in file mmm_mon.conf), daemon is not
starting:

service mysql-mmm-monitor start
Starting MMM Monitor Daemon: 2010/08/23 16:44:48 INFO STARTING...
2010/08/23 16:44:48 DEBUG Created pid file '/var/run/mysql-mmm/
mmm_mond.pid' with pid 9175
2010/08/23 16:44:48 INFO Waiting for network connection...
2010/08/23 16:44:48 INFO Spawning checker 'ping_ip'...

this is only info in log file.

mysql_control show/ping shows: ERROR: Can't connect to monitor daemon!

Environment info:
Centos 5.5 (x86_64), latest updates
Epel repository (which from mysql-mmm was installed)
Two NICs bonded together (i check also without bonding, same problem)
firewall/iptables off
Versions installed:
mysql-mmm-agent-2.2.1-1.el5
mysql-mmm-2.2.1-1.el5
mysql-mmm-monitor-2.2.1-1.el5
mysql-mmm-tools-2.2.1-1.el5

I have to add, that i have another working cluster, with same versions
and configs (also centos 5.5, only different hardware and IPs) and
everything is perfect.

[nemd]

Update to previous message:

link to my config files: http://pastebin.com/AYux2y6n

--
Best regards,
nemd

[István Pödör]

Hi There,

Well at the moment I can't test it and I'm not sure, but did you tried to remove all the IPs from ping_ips and add only 127.0.0.1 instead?

Istvan

--
You received this message because you are subscribed to the Google Groups "MySQL Multi Master Manager Development" group.
To post to this group, send email to mmm-...@googlegroups.com.
To unsubscribe from this group, send email to mmm-devel+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/mmm-devel?hl=en.

[nemd]

hi,

thanks for answer. I removed IPs like you wrote and after that mmm-
monitor is able to start. So it's not mysql-mmm problem. There is
quite complicated network/routing configuration (and not so
professional i guess) so i think that's the case. Strange thing is
that ping directly from shell is ok, same for remote mysql
connections. I'm not sure in what way mmm-monitor checks hosts (i
assume it's not simple ping or it's packets correction sensitive). I
will try to check whole network config (switches including) and i'll
leave info about progress here.

--
Best regards,
nemd

> > mmm-devel+...@googlegroups.com<mmm-devel%2Bunsubscribe@googlegroups­.com>
> > .

[István Pödör]

Hi there,

I'm glad it's working now :)

Anyhow that was my first guess, some complicated routing. If you could share your routing table that would be useful to see what went wrong. by the way, every time when you can see that one of the mmm processes stucked like check ping_ips this time, you can try to figure out what's wrong because that will be the name of the module it's executing too.

Bests,
istvan

To unsubscribe from this group, send email to mmm-devel+...@googlegroups.com.

[nemd]

hello again,

This is local test network, before i send whole rack cabinet to
client. Routing is messy and is causing more problems, so it's
pointless to fix this at this time. But anyway, what can i expect
after removal real IPs from ping_ips line ? What is exact purpose of
this particular checks ?

--
Best regards,
nemd

[Pascal Hofmann]

Hi nemd,

> This is local test network, before i send whole rack cabinet to
> client. Routing is messy and is causing more problems, so it's
> pointless to fix this at this time. But anyway, what can i expect
> after removal real IPs from ping_ips line ? What is exact purpose of
> this particular checks ?

The ping_ips are used to determine whether the network connection of the monitoring host is OK. If it isn't OK, no actions will be taken when agent hosts are not reachable (via ping, mysql).
This is to prevent mmm_mond from detecting all hosts as "down" where in reality mmm_mond just can't reach anything because of a broken network connection.

Regards
Pascal

[nemd]

hi,

So what this means in practice ? Could you describe some simple case
scenario with empty ping_ips line ?

--
Best regards,
nemd

[Istvan Podor]

Well, this means that when your connection from mmm_mon to all node is fail because of network issues, but it can reach the passive master, probably will fail over to it even if there is no need of it.

Istvan

> --
> You received this message because you are subscribed to the Google Groups "MySQL Multi Master Manager Development" group.
> To post to this group, send email to mmm-...@googlegroups.com.

> To unsubscribe from this group, send email to mmm-devel+...@googlegroups.com.

[nemd]

And this is exact what i want to know :).Thank you for your help.

--
Best regards,
nemd