How to manage the CAS Array?

It is one thing to know that a single Client Access Server is up and running, however in most Microsoft Exchange 2010 environment engineers are deploying multiple servers configured as an array for high availability. The purpose of the CAS Array is to balance the user’s connections on several CAS that deliver the protocol needed by the user.

GSX Monitor & Analyzer main features

Real Time monitoring of the service provided by the CAS Array

Warning

Managing the Array service delivery

Real Time statistics

Load Balancing Monitoring

Real Time Statistics

Troubleshoot performance and availability issue in the CAS Array

GSX Monitor measures the performance of your underlying server at the system, role and Array level, providing proactive performance. Additionally it takes also a step forward and makes the statistics available for reporting and trending in GSX Analyzer.

CASArrayscreenshots


CASArray CASDesignation

Real Time monitoring of the service provided by the CAS Array

GSX acts as the bridge between different servers by providing real-time monitoring of the service provided by the Array. GSX considers the CAS Array as a platform to deliver all of the protocols set among the different servers. When protocols are set on different servers in the Array, if one of these server fails the Array can still deliver the protocols and GSX still consider the service as delivered.

When four servers are part of the Array with only one different protocol configured on each server, when one server is down, the CAS Array is down.

CASArraysLED

If the Array is down, you will see it directly in the Exchange overview with a Red color on top of the group server.

Serveruptime

 

Warning

GSX also ensures that the administrator is aware of all errors – both critical and non-critical. By detecting and sending warnings if a non-critical error is detected on the array, the system ensures that the administrator has a chance to address these errors before the error becomes more serious.

For example, if the hour changed on one of the servers it leads to real-time replication. With GSX, this problem won’t lead to server failure, and the protocols will still be available.

GSX also alerts the administrator when of the CAS servers is down but the protocol can still be delivered by another server. In this case, the Array is not down but the administrator will be alerted on this non-critical error so they can address it at this non-critical juncture.

Managing the Array service delivery

GSX enables the administrator to stay on top of the system by displaying in real time the critical information related to the CAS Array. Critical information includes:

  • Protocol availability across the CAS Array: Shows the performance of the Array to deliver the service.
  • CAS Array load balancing by number of connections: Shows the repartition of the load between servers and hence shows if the Array is really performing its load balancing task and how.

piechart

 

GSX Analyzer goes a step further providing CAS Array availability statistics for trending and forecasting. Based on the ability for the Array to provide the protocols configure within the Array these statistics give you the measure of the global service that your Exchange CAS infrastructure should deliver to your end user.

Real Time statistics

GSX enables the administrator to stay on top of the system by displaying in real time the critical information related to the CAS Array. Critical information includes:

  • Protocol availability across the CAS Array: Shows the performance of the Array to deliver the service.
  • CAS Array load balancing by number of connections: Shows the repartition of the load between servers and hence shows if the Array is really performing its load balancing task and how.

Load Balancing Monitoring

GSX gathers information through WMI and performs PowerShell tests to check the status of the nodes.

Through GSX, the NLB status is displayed and based on the services in charge to keep the NLB up and running;. The convergence of the node is also displayed.

GSX checks each node and send two types of alerts:

  • WARNING if Node is converging or draining
  • DOWN if Node is not converged, converging or draining.

GSX checks also the node at the CAS Array level and send 2 different types of alerts:

  • WARNING if a node is in warning
  • DOWN if a node is down or can't resolve or ping the VIP

If the NLB is down (or warning) the CAS Array is also down (or warning).

networkloadbalancer

Real Time Statistics

GSX gives real-time statistics on the critical points of the NLB. In one view, you gather all the information related to your CAS role, CAS Array and Network Load Balancer.

Some Common Statistics:

  • NLB: Successful: Status and time to access the VIP.
  •  X port rule(s) configured: Port rule(s) that determines what is proceeded by the NLB.
    Example: Between the port 80 and 85 only TCP protocol will be dispatched across the node.
  • Status of the NLB Nodes: Node name, Convergence status, NIC availability (GSX tests the availability of the Network Interface Card dedicated to the NLB), Net connection (GSX checks the NIC and the network).

Troubleshooting performance and availability issue in the CAS Array

GSX Monitor & Analyzer check the Array as an entity and the CAS at the role level, providing statistics of availability, performance and usage.
Thanks to the Environment Health view of GSX Analyzer and its forecasting capabilities, you can now compare the availability, performance and usage of your CAS within your Array, quickly identifying issues on server and bottlenecks.
You can compare the usage (connections: peak and average), end user performance (Outlook connectivity, Active Sync, OWA response time, end to end protocol performance tests), system performance (CPU, RAM, Disk), availability (uptime with our without maintenance, during business or off hours) of all your CAS within the Array quickly identifying issues.

EHV