In the scenario detailed in a previous mail:
"...my application is a set of binaries, say bin1, bin2, ..., binN; lets
imagine machine mach1 runs bin1 and bin2, and mach2 runs bin1 and bin3; so
maybe the solution is to have mib1 for bin1, mib2 for bin2 and so on,
and query mach1 for mib1 and mib2 and query mach2 for mib1 and mib3"
does it make sense to have a single machine concentrating the
information of all of them?
I undestand SNMP is "agent centric", so I guess the idea is to query an
agent for information about it and not other agents, what do you think?
Sebastian-
Sebastian Bello escribió:
> Thanks to all!
>
> Dave Shield escribió:
>> On 07/09/06, Sebastian Bello <[EMAIL PROTECTED]> wrote:
>>> 2- ok, so I just should choose a version?
>>
>> If you are writing a MIB, you should use SMIv2 (i.e. NOTIFICATION-TYPE).
>> There is no reason for using SMIv1 any more, unless you are forced to
>> by ancient software. SMIv2 has been defined for over six years now,
>> so there is very little excuse for anything not to support it.
>>
>> This has nothing to do with the version of SNMP that you use to query
>> the agent.
>>
>>
>>
>>> 4- my application is a set of binaries, say bin1, bin2, ..., binN;
>>> lets
>>> imagine machine mach1 runs bin1 and bin2, and mach2 runs bin1 and
>>> bin3; so
>>> maybe the solution is to have mib1 for bin1, mib2 for bin2 and so
>>> on, and
>>> query mach1 for mib1 and mib2 and query mach2 for mib1 and mib3, is
>>> this a
>>> good approach?
>>
>> Not really.
>> You should take a step back from the individual applications, and
>> think about the information that you're wanting to monitor/control.
>> If there are areas of overlap between the various applications, then
>> it makes sense to define a standard table (or tables) to cover the
>> common information. This can then be supplemented by extra tables to
>> cover any information that is specific to individual apps.
>>
>> But the more general you can make the basic MIB framework, the less
>> work you'll have to do later when your bosses come along with yet
>> another application they want you to support.
>> And if you can use existing standard MIBs for some/all of this - so
>> much the better.
>>
>> Note that the division between MIB files is purely for your
>> convenience. The agent (and mgmt applications) effectively merge them
>> all into a single tree before using them. So it makes no difference
>> in practise whether everything is defined in one MIB file, or each
>> table (or even scalar object!) is defined in a separate file.
>> Somewhere in between is probably a good idea :-)
>>
>> But be driven by the logical structure of the abstract information,
>> rather than the particular running architecture that you're currently
>> looking to monitor.
>>
>> Dave
>>
>
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Net-snmp-users mailing list
[email protected]
Please see the following page to unsubscribe or change other options:
https://lists.sourceforge.net/lists/listinfo/net-snmp-users
