sorry, nopow the line clear,
hi all,,, PoE= Polling Engine
please , i have to know 2 things:-
1- what different between Hot Standby Engine and additional Polling Engine??
i get these 2 informations :
**Orion Hot Standby Engines are not currently capable of collecting data and computing network performance statistics for EnergyWise, VMware ESX, or wireless devices,,,, but (additional PoE only support this, what this mean ??????????? it is just monitor IPphones,VM ,, not swithes and routers).
**Orion Hot Standby Engines one NPM server active and the other passive, but PoE both r working at same time so i can say 250 node will polled from one , and the other 250 from the other(load balancer polling engine).
*********i want same idea as additional PoE , but without any limitation**********
2-we have now NPM and we will take newer one, and now we monitor 250 node and 10,900 interface
and after one year god willing , will be twice 500 node , so please which one of the above design is proper ?
Note: we will take addition to NPM,, NCM, Netflow and Engtoolset,
so, please i need the the answer, and the proper .
and so sorry for this long Q,
thank u ,
Hot standby will take over if one of your "operating" polling engines fails
if you are monitoring 250 nodes and 10,900 interfaces currently, you will want an additional polling engine at a minimum. SW recommends max 8000 elements per engine (element is node, interface, volume, etc..) In addition to that, you have to consider the frequency of polling and any custom pollers...if you poll more than default frequency, that will add load to the box.
that said and based on what you've said, you currently have 11150 elements on that one server.
The Hot standby would be based on your companies need for redundancy or high availability of the monitoring system
hope that helps
Great post, Larry, thanks.
fu2000-here are some links to the online Orion NPM Administrator Guide, which describe further what Larry has said.
thanks nioylif & MarieB,,,
nioylif,,,,,ok i get your idea,,,it is proper to use hot standby ,, but the idea of this is :
if one server fail the other come active, but this not solve the problem of stress on one server, (i.e , the active server is the active polling engine, so still we work with one polling engine ),
at same time i can conclude from your explain , that we need another polling engine, do u mean just as failover, ,,,,,,,, for example : the active server (polling engine) dosent fail for one year, so this server represent our case now, we work with one server,,,,,,,,,,, except that u mean at same server we create more than one polling engine, so they will work at same time as load balance.
please i need your help to fix this,
thank u very much
Sounds like you've got it.
you would want to add a server running the "additional polling engine" software package for the load balancing.
If in addition, you want redundancy, then you would get yet another server running the "hot standby engine" software.
It sounds like you need either 3 or 4 servers
For the FailOverEngine solution
or for the Hot Standby Solution
The Hot Standby solution will only support one (out of many) failed engines for basic NPM monitoring,
But the Failover Engine will fully support all of the modules running on its dedicated Active engine and is the best option.
thank u very much, (for all)..
nioylif, be Patient with me please,,
((you would want to add a server running the "additional polling engine" software package for the load balancing))
if i did this , , , i dont want redundancy,,,
can i let one polling engine monitor half of the nodes, and the other one monitor the rest? (both r active at same time (load balance))
1- what will be happen if one of the polling engine fail, , is the other one will be responsible for all nodes immediatly?
2-can i do this load balance, between two servers each one represent polling engine?
thank u so much
and last thing :-
i get this from NPM administration guide
Orion Hot Standby Engines are not currently capable of collecting data and computing network performance statistics for EnergyWise, VMware ESX, or wireless devices,,,, but (additional Polling Engine only support this, what this mean ??????????? it is just monitor IPphones,VM ,, not swithes and routers?????????).
Sorry for being late to the party here.
First off, lets take Hot Standby Engine or HSE off the table. We are no longer selling that to customers.
Regardless, both of these solutions primary use case if for High Availability. These solutions are not meant to be automated load balancer of elements. Now with that being said if you do have multiple pollers, you can load balance elements between the pollers manually today with some of the utilities we provide out of the box.
We have multiple pollers within Orion grabbing different info. The HSE only protected and watch the standard poller, which gather node, interface and volume information. We poll and gather VMWare, Wireless and EnergyWise info via other polling means. The HSE did not watch the services on the Orion server, which is one of the primary reasons we released the Failover Engine. The Failover Engine is a proper, complete High Availability solution to protect all of Orion, including the modules.
To answer this question - can i let one polling engine monitor half of the nodes, and the other one monitor the rest?
The answer is yes.
The answer to these other two:
1- what will be happen if one of the polling engine fail, , is the other one will be responsible for all nodes immediately? The answer here is no, you would have to manually assign them to the running poller today2-can i do this load balance, between two servers each one represent polling engine?
thank u bshopp,,
please i want to know this,
1- is Failover Engine idea is not same as Hot Standby Engine??? ok, i understand from your explain that failover engine better because it has more fetures, but we still work with one active and the other is passive, or what???????
i didnt find book explains failover engine , but hot dtandby engine on NPM administration guide.
2- from all the above explains , can i conclude that, for our company , we can use server for database and one for first polling engine and one for second polling engine, and we make polling engine load balance . is it enough ?, (i am with this),
OR addition to that , we have to use failover engine as redundancy.
thanks , and sorry for this long story!!!!
The Fail Over Engine is the replacement product for the Hot Standby Engine and the Hot Standby Engine will NOT be suppprted in the next release of NPM.
If one polling engine can support your entire load you could run it on a single engine or optionally spread the load across multiple servers, but if one server will not support the entire load, you will not be able to migrate nodes from a failed server to the remaining server and expect it to still work. Sizing is very important and with 11000 elements, you will need two polling engines (one primary and one additional). but that will be un-resillient. To make it resilient, you will need to add two Fail Over Engines (one dedicated to each of the polling engines).
You may also want to do a full resilence/DisasterRecovery/ContinuityOfBusiness check and look at the resilience of your SQL and your connectivity.
Thank U Very Much,,