[smokeping-users] smokeping's server capacity

[Dez Cadena]

Tobias Oetiker wrote:
> Dez,
>
> if it is snmp latency then this is probably right, the question is
> if you cisco is happy with this ...
>   
Tobi,
Cisco is ok so far... I've tried increasing forks but load average of 
the box is rising fast (as expected)

> you might want to look into using the remote smokeping
> functionality
>
>   
Are you talking about the slave/master setup? How would that help?
cheers

> cheers
> tobi
>
> Today Dez Cadena wrote:
>
>   
>> Tobias Oetiker wrote:
>>     
>>> Dez,
>>>
>>> what makes you think the process is diskbound ? do you have high iowait ?
>>>
>>>
>>>       
>> Tobi,
>>
>> I thought that all storing to RRDs is performed right after all polling is
>> done but I guess I'm wrong. By doing a smokeping --debug, I realised that
>> RTTMon operation is set for first x devices (x=number of forks for probe) and
>> then the response for those devices is stored to the appropriate RRDs... and
>> so on
>>
>> So, the only way to further improve the "performnce" is to increase the number
>> of forks per probe?
>>
>>
>>     
>
>