topic Re: CPsizeme on Full HA in Firewall and Security Management

CPsizeme on Full HA

Wei_Soon_Heng — Tue, 25 Aug 2020 07:25:21 GMT

Hi All,

My client has a plan to replace the current FULL HA of model 12400 with new model and deploy it in normal ClusterXL deployment. Before that, we need to do appliance sizing on current FULL HA server. So will it be any over-sizing if we based on output of cpsizeme from current Full HA appliance?

Will running CPSizeme have any impact on performance?

Thanks

Re: CPsizeme on Full HA

Arik_Ovtracht — Tue, 25 Aug 2020 07:45:07 GMT

Hi,

There shouldn't be any over-sizing based on what you described, but please pay attention to the generated report, and consult the results with Check Point professional services if something seems off.

CpSizeMe does not impact the machine's performance, since it is intended to collect performance-related data, and that would skew the results.

Re: CPsizeme on Full HA

Wei_Soon_Heng — Tue, 25 Aug 2020 10:00:09 GMT

Hi Arik,

Just want to confirm again, there is no any increase of performance load like CPU load , memory consumption & etc , right?
Other than CPSizeme tools, could we able to estimate the sizing based on output of CPVIEW ?

Thanks

Re: CPsizeme on Full HA

Arik_Ovtracht — Tue, 25 Aug 2020 10:04:07 GMT

There should be no notable increase. Of course, CpSizeMe does take a bit of memory and CPU, but it shouldn't affect the results of the sizing process.

Yes, you can use CPview, but that will only give you metrics, not a recommendation.

Re: CPsizeme on Full HA

Wei_Soon_Heng — Tue, 25 Aug 2020 12:38:43 GMT

I still got some doubt on sizing part.
Normally, the CPU & memory usage of standalone server will be higher than distributed. If we take CPSizeme from standalone server, and purchase the next new machine which intent to deploy as purely gateway only, it is inappropriate sizing method in this scenario from my perspective which lead to oversizing.

Thanks

Re: CPsizeme on Full HA

Arik_Ovtracht — Wed, 26 Aug 2020 11:36:50 GMT

I did some checking, and you are right.

Running sizing on a standalone, will provide inaccurate result as the MGMT side also consumed CPU resources.

You can overcome this (almost entirely) by running the sizing tool on the Active GW which is acting as the BACKUP MGMT.

This will give more accurate information and “true” sizing.