Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 3 Next »

Discovery Manager Checks

The Abiquo Discover Manager (NARS) is an Abiquo remote service. The Abiquo Infrastructure check verifies the platform infrastructure at intervals. It does this by connecting to all of the hypervisors on all of the Abiquo racks and all of the public cloud regions, and their VMs, to compare them against the information recorded in the Abiquo database. 

It checks the following components of physical machines:

  • Virtual CPU
  • RAM
  • NICS
  • Datastores
  • Initiator IQN

Any changes are recorded in the Abiquo database.

You can configure the discovery manager as described below using the Abiquo Configuration Properties#nars

Multiple Threads

The Abiquo Discovery Manager has a pool of plugin-clients. Each plugin-client connection checks different machines in the infrastructure. You can configure the number of plugin-clients with abiquo properties. This number should be related to the number of cores of the machine running discovery manager service. 

Timeout

If a physical machine does not respond within the timeout period, Abiquo will 

  • mark the machine as a candidate for HA, if Abiquo HA is enabled on the rack OR
  • discard the machine

The default timeout for the infrastructure check is 5 minutes (30000 milliseconds). 

You can configure the timeout with the abiquo.ha.infcheck.timeout property.

Discarded Machines

If a machine cannot be contacted within the timeout period, and HA is not enabled, it is discarded by the infrastructure check. Machines that are discarded are:

  • logged in the Abiquo logs (nars and catalina.out) with the message: 
  • not checked again

Maximum Duration

The maximum duration of the infrastructure check is the same as the infrastructure check timeout value.

Delay Between Checks

After the check has completed, the Discovery Manager waits for the delay period before starting a new check. You can set the delay with abiquo properties. The default delay period is 5 minutes.

Check Frequency

The time between checks is equal to the duration of the check plus the delay between checks. Therefore you may be able to increase the frequency of the infrastructure checks by decreasing the timeout value.

Configure HA in Relation to Infrastructure Checks

Abiquo High Availability is triggered after a certain number of failed infrastructure checks on a machine. You can configure the number of failed infrastructure checks to wait before starting HA using the abiquo.ha.infcheck.cycles property. 

Note that the abiquo.ha.check.delay is the delay between checks of machines that already have HA in progress in order to determine if HA movements are complete.

Configure Virtual System Monitor for Public cloud

To ensure performance and scalability of the remote services in managing many public cloud regions, Abiquo introduced the following properties to tune the virtual system monitor.
abiquo.vsm.pollingfrequency.<cloudprovider>
The delay between VSM polling of virtual machines in milliseconds in a cloud provider plugin
abiquo.vsm.vmsyncfrequency.<cloudprovider>

The delay between virtual machine synchronizer operations in milliseconds in a cloud provider plugin

abiquo.vsm.actor.creation.delay.<cloudprovider>

The delay in milliseconds added between the creation of actors for a cloud provider plugin type on VSM startup.

Abiquo SaaS example

An example of values that could be used in a SaaS system.

abiquo.vsm.pollingfrequency.amazon = 600000

abiquo.vsm.vmsyncfrequency.amazon = 300000

abiquo.vsm.actor.creation.delay.amazon = 2500


  • No labels