Health Check Monitoring

1
Health Check Monitoring

• Ivan M. Redondo
• Technical Relation Manager

2
What is Health Check Monitoring?

• a service to perform health checking on various
  subsystems
• performs self-healing actions.
• tests can be configured and extended by the user
• HCA will rely on the IMA Advanced service to
  perform privileged actions
• stopping IMA
• rebooting a server

3
Typical Problems to be addressed

• Enumeration of published applications
• XML service hangs (Black Hole Issue)
• IMA service hangs
• IMA Ticket requests fail
• XML Ticket requests fail
• Terminal Service hangs

4
Issues NOT addressed

• SSL Relay bad certificate test
• SG authentication test
• Queued ASP Requests on the IIS WI server
• AAC Authentication test
• IMA slowness to enumerate user permissions
• End user client cannot obtain a valid Citrix
  license
• No ICA listener test

5
Farm Properties
6
Server Properties
7
Four Tests

• IMA Service test
• IMATest.exe
• Terminal Services test
• CheckTermSrv.exe
• XML Ticket Request test
• RequestTicket.exe
• Logon/Logoff test
• LogonMonitor.dll

8
IMA Service Test
9
IMA Service Test

• Performing an app enumeration on the IMA service
  to ensure it is up and running
• Description IMA Test
• Interval 60
• Filename imatest.exe
• Threshold 2
• Timeout 120
• ActionOnError Alert only

10
Terminal Services test

• Enumerates list of sessions running on the server
  and session user information
• Interval 30
• Filename checktermsrv.exe
• Threshold 2
• Timeout 120
• ActionOnError Alert only

11
XML Service test

• Requests a ticket from the XML service running on
  the server.
• Interval 60
• Filename requestticket.exe
• Threshold 2
• Timeout 120
• ActionOnError remove server from LB

12
Logon Monitor test

• Monitors the logon/logoff cycles to determine if
  there is a problem with session initialization
• Interval 30
• Filename LogonMonitor.dll
• Threshold 2
• Timeout 120
• ActionOnError Alert only

13
Logon Monitor test parameters

• Three parameters unique to Logon Monitor
• Must be updated manually through command line
• SessionTime 5
• SessionThreshold 50
• SampleInterval 600

14
Recovery Actions

• Performed when a test fails
• Intended to self-heal the server
• Configured by admin
• Regardless of action an event will be triggered
• Event will cause an alert in SMA

15
List of recovery actions

• Alert Only
• Remove Server From Load Balance Tables
• Shutdown IMA service
• Restart IMA Service
• Reboot Server

16
Alert Only

• HCA will send an error message to the event log
• HCA will continue running its tests but will not
  post more event log on subsequent failures
• If the test starts passing again another event
  will be sent to the Application log indicating
  that the test passed

17
Remove Server From LB Tables

• HCA will communicate to the local IMA service
• IMA service will remove server entry out of the
  dynamic store server load table
• ZDC will stop resolving incoming connections to
  the server
• Existing connections continue to run
• Disconnected sessions will continue to resolve to
  this server
• Direct connections can be made (bypassing LB)

18
Farm Exclusion Limit
19
Shutdown IMA service

• HCA will send a message to the IMA Advanced
  service to shutdown IMA
• IMA Advanced service tries to shutdown IMA
• If not successful it will kill the IMA process
• HCA executes will continue running its tests
• Any test that relies on IMA will fail, triggering
  and event and the configured action

20
Restart IMA Service

• HCA will send a message to the IMA Advanced
  service to restart IMA
• IMA Advanced service tries to shutdown IMA
• if not successful it will kill the IMA process
• IMA Advanced service will then attempt to start
  the IMA service
• Obviously only shutting down IMA is guaranteed to
  succeed, starting IMA is not
• Any test that relies on IMA will fail, triggering
  and event and the configured action

21
Reboot Server

• HCA will send a message to IMA Advanced service
  to reboot the server
• The IMA Advanced service will make a system call
  to reboot the machine
• After reboot, HCA will operate normally.

22
Questions?
23
(No Transcript)