X

An Oracle blog about Oracle Enterprise Manager and Oracle Management Cloud

Enterprise Manager Ops Center - Ops Center EC/PC sizing

Counting Access Points Vs Assets

When designing a customer's Ops Center environment, one of the first questions I get asked is how many servers can an Enterprise Controller (EC) or a Proxy Controller (PC) support.  Customers think of sizing, as being the number of servers,  LDOMs, etc. and these are lumped together into a total number of assets. In the real world, the answer to the "How many assets can my Enterprise Controller /Proxy Controller support?" question is, "It Depends".  How you place your Proxy Controllers in your network and what you manage from each Proxy Controller can impact the total number of assets that can be supported from an Enterprise Controller.

The amount of load placed on an Enterprise Controller or Proxy
Controller is not directly determined by the number of assets it manages, but by the number of Access Points
(AP) it is supporting. A simple view of Access Points is that they are a connection. But it is a little more complex in that they are also a reflection of the way Ops Center internally models its data structures. The actual load on any given Enterprise
Controller or Proxy Controller closely links to the number of Access
Points it is managing. The number of Access Points seen by the
Enterprise Controller can differ depending on whether an asset is
connected via a single or multiple Proxy Controllers. Therefore, the Access Point count on the Enterprise Controller will be higher than the actual number of managed
assets, if assets are managed by multiple Proxy Controllers. When the documentation is referring to Enterprise Controller /Proxy Controller capacity and uses the term "Asset", it is actually counting "Access Points".

Let's look at a couple examples of this to make it clearer:

Physical Server ILOM

In this example, a single physical server has multiple data collection methods. A Proxy Controller can gather LOM data by accessing the service processor and by using the agent running inside the operating system.


If the same Proxy Controller is used to access a single asset, the two data feeds are consolidated into a single Access Point. The Enterprise Controllers total load will be a single Access Point.


If different Proxy Controllers are used to access a single
asset,  each Proxy Controller will record it as a separate Access Point. The Enterprise Controllers total load will be the sum of both Proxy Controllers and will be a total of 2 Access Points.

LDOMs

Another example would be an LDOM guest where we obtain data from the
Control Domain agent and the agent running inside the LDOM guest OS.

Once again, if  both data feeds are via a single Proxy Controller they only count as 1 Access Point on both the Proxy Controller and the Enterprise Controller.

Where as, if each data feed is via a separate Proxy Controller they each count as 1 Access Point on each Proxy Controller and the total Access Point count on the Enterprise Controller will be 2.

With the release of Enterprise Manager Ops Center - 12.2.2, we have updated and renamed the Enterprise Manager Ops Center Sizing and Performance Guide (https://docs.oracle.com/cd/E40871_01/doc.122/e57052/toc.htm). This revised document is easier to understand and has been updated to reflect current releases of Ops Center. Note: Since the term "Access Point" was not considered
to be commonly understood, the term "Asset" has been used in the documentation. 

To make the counting of Access Points easier, we have added a program to the OCDoctor toolbox. This program "AssetCount.sh" is new in OCDoctor version 4.45 and can be found on any EC/PC or system that has an agent deployed. The path to the OCDoctor toolbox is /var/opt/sun/xvm/OCDoctor/toolbox.

# /var/opt/sun/xvm/OCDoctor/toolbox/AssetCount.sh
        AssetCount.sh [-v] { standard | machine | agent | all }
        standard - show Asset count
        machine - show machine output with Assets names
        agent - show information about agents, agentless Assets, Service Processors
        all - show all the information together
        -v : verbose output, present additional notes around the counting algorithm
        (Version 1.0 2015/01/08)
#

Let's just look at the standard output (we are running this on the Enterprise Controller). The output shows:

  • The total Access Point count for the EC (72)
  • The number of Access Points for each type of asset, for each Proxy Controller ( Note: the total Access Point count for each Proxy Controller is labeled as Assets )
root@ec1:~# /var/opt/sun/xvm/OCDoctor/toolbox/AssetCount.sh standard
EC 72
Proxy Assets Zones Ldoms OVMGuests Servers Storages Switches ExadataCells MSeriesChassis MSeriesDomain
------------------------------------------------------------------------------------------------------
pc4 32 5 25 0 2 0 0 0 0 0
pc1 28 0 26 0 2 0 0 0 0 0
pc0 12 2 4 0 6 0 0 0 0 0
Use option '-v' to see additional notes on the counting algorithm.

#


You can also use the "machine" option to list out which asset is attached to which proxy.

# /var/opt/sun/xvm/OCDoctor/toolbox/AssetCount.sh machine
EC 72
Proxy 32 pc4
Zones 5 S11zone101 S11zone102 S11zone100 S11zone103 S11zone104
Ldoms 25 stdldom21 stdldom34 stdldom36 stdldom22 stdldom45 stdldom47 stdldom25 stdldom43 stdldom49 stdldom30 stdldom23 stdldom29 stdldom42 stdldom20 stdldom32 stdldom46 stdldom44 stdldom26 stdldom31 stdldom27 stdldom48 stdldom40 stdldom35 stdldom41 stdldom28
OVMGuests 0
Servers 2 10.187.70.169 pc4
...
Proxy 28 pc1
Zones 0
Ldoms 26 stdldom21 stdldom34 stdldom36 stdldom22 stdldom45 stdldom47 stdldom25 stdldom43 stdldom49 stdldom30 stdldom23 stdldom29 stdldom42 stdldom33 stdldom20 stdldom32 stdldom46 stdldom44 stdldom26 stdldom31 stdldom27 stdldom48 stdldom40 stdldom35 stdldom41 stdldom28
OVMGuests 0
Servers 2 10.187.70.171 pc1
...
#

You can see this would be incredibly verbose on a large environment ( I have truncated it here). You can clearly see that the LDOMs (stdldomXX) are being reported by both "pc1" (via Control Domain) and "pc4" (via OS agent). Note: the differing LDOM count on "pc1" and "pc4"  as "stdldom33" has no OS or agent on it, so it only reports against "pc1".

You can also use the "agent" option to display agent/agentless/LOM totals for each Proxy Controller.

# /var/opt/sun/xvm/OCDoctor/toolbox/AssetCount.sh agent
EC 72
Proxy Agents Agentless SPs
--------------------------
pc4   25     2         0
pc1   1      1         0
pc0   5      5         5
Use option '-v' to see additional notes on the counting algorithm.
#

In addition, OCDoctor.sh --troubleshoot will also be checking for the number of configured Access Points and comparing this number against the recommend Access Point count. It will print a warning at approximately 90% of the recommended capacity and a warning if you exceed the recommended capacity.

Armed with the information above, you should be able to better design your Ops Center deployments and to monitor the growing load on the Enterprise Controller/Proxy Controllers as you add more assets.

Regards,

Rodney

Be the first to comment

Comments ( 0 )
Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.