Hosts status

[Martin Pala]

OK, it thus seems that monit was able to contact m/monit and the  
configuration should be correct.

The problem will be most probably caused by very short read timeout  
(10ms) which was set in m/monit. This can be too short depending on  
network latency. We have fixed it and will release new beta beginning  
next week.

Can you please add the content of M/Monit log? (you can find it in  
logs/mmonit.log) This can confirm the root cause.

Thanks,
Martin


On Aug 8, 2008, at 3:26 AM, Antonello Alonzi wrote:

> Tank you for all information.
>
> After install new versione of monit (5b2) I haven't insert the "set
> mmonit" in monitrc.
>
> Now from mmonito I see the monit host, but are diplayd as "inactive".
> I see into syslog of monit host:
> -----------------------
> message sending failed to http://10.113.15.193:8080/collector --
> HTTP/1.1 500 Internal Server error
> -----------------------
> The mmonit version is 2.0 beta3.
>
> Have you any idea about these error?
>
> Thank you,
> Antonello
>
> 2008/8/5 Martin Pala <martinp at tildeslash.com>:
>>
>> On Aug 5, 2008, at 1:06 AM, Antonello Alonzi wrote:
>>
>>> Hi to all,
>>>
>>> 2008/7/31 Rafael G. <rgo at aspgems.com>:
>>>> Jan-Henrik Haukeland escribió:
>>>>>
>>>>>
>>>>> I suspect that you have an old version of Monit. M/Monit require
>>>>> Monit
>>>>> version 5, beta 2, which you can download from
>>>>> http://tildeslash.com/monit/dist/beta/monit-5.0-beta1.tar.gz
>>>>>
>>>>>
>>>>
>>>> I suspect the same :)
>>>>
>>> thank you for suggest.
>>>
>>> Now I have install monit 5.0_beta2 and some othe function (eg reload
>>> service on monit host from mmonit) run correctly, but the "Hosts
>>> status" page on mmonit don't display the correct status of monit  
>>> host.
>>> I see:
>>>
>>> red point -  apache - 0 - 0,0 - No report from monit. Last report  
>>> was
>>> Thu, 01 Jan 1970 01:00:00 - 0
>>>
>>> Now I have monit 5.0_beta2 and mmonit 2.0_beta 3
>>>
>>> do you have other ideas?
>>>
>>> Thank you,
>>> antonello
>>
>> Hi,
>>
>> according to message the M/Monit didn't received any message
>> (heartbeat) from monit (the timestamp of the host entry in M/Monit is
>> 0 which is 1.1.1970 - beginning of unix epoch).
>>
>> To fix this, you'll need:
>>
>>
>> 1.) add configuration to monit:
>>
>>  set mmonit http://<mmonit's user>:<mmonit's user
>> password>@<mmonit's ip>:<mmonit's port>/collector
>>
>> example (will work out of the box if you didn't changed mmonit  
>> account):
>>
>> --8<--
>>  set mmonit http://monit:monit@192.168.10.10:8080/collector
>>  set eventqueue basedir /var/monit
>> --8<--
>>
>> If mmonit is running on different port, you'll need to modify the
>> configuration accordingly. By default the following users were  
>> created
>> in M/Monit (you can add/remove users and change passwords as you like
>> in M/Monit):
>>
>>  admin:swordfish
>>  monit:monit
>>
>> Note that the eventqueue is optional, but recommended - it will retry
>> the delivery of alerts and status messages if mmonit will fail.
>>
>>
>> 2.) after you added the "set mmonit ..." line, reload monit
>> configuration (or just restart monit):
>>
>> --8<--
>>  monit reload
>> --8<--
>>
>>
>>
>> ... that's it.
>>
>>
>> If the messages still won't go through, the problem will be most
>> probably on the network (for example firewall dropping the conections
>> to mmonit). You can fix the network settings accordingly.
>>
>> To see what monit does in detail you can run monit with the -v  
>> option,
>> which will show whether and how the messages to mmonit failed.
>>
>>
>> Martin
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> mmonit-user mailing list
>> mmonit-user at tildeslash.com
>> http://www.tildeslash.com/mailman/listinfo/mmonit-user
>>
> _______________________________________________
> mmonit-user mailing list
> mmonit-user at tildeslash.com
> http://www.tildeslash.com/mailman/listinfo/mmonit-user