Configuring Scalable Chassis High Availability

Use these settings to configure Active/Standby Chassis.

Configuring Chassis Weights (Chassis High Availability Factors)

Each hardware component in a Chassis has a quality weight factor, which sets its relative importance to overall Chassis health.

For example, ports are more important than fans and are typically assigned a higher weight value.

The Chassis grade is the sum of all component weight values.

In a High Availability environment, the Chassis with the higher grade becomes Active and handles traffic.

The grade for each component is calculated based on this formula:

(Unit Weight) x (Number of components in the state "UP")

To see the weight of each component, run in Gaia gClish:

asg stat -v

Description

Use the "set cluster configuration high-availability factors" command to configure a hardware component's weight.

Syntax in Gaia gClish of the Security Group

set cluster configuration high-availability factors member <SGM Factor>

set cluster configuration high-availability factors port {other <Other Port Factor> | standard <Standard Port Factor> | mgmt <Management Port Factor> | bond <Bond Port Factor>}

set cluster configuration high-availability factors sensor {cmm <CMM Factor> | fans <Fans Factor> | power_supplies <PSU Factor> | ssm <SSM Factor>}

For more information, see:

Parameters

Parameter

Description

<SGM Factor>

Weight factor for a Security Group Member.

Valid range: integer between 0 and 1000.

<Other Port Factor>

High grade port factor.

Valid range: integer between 0 and 1000.

<Standard Port Factor>

Standard grade port factor.

Valid range: integer between 0 and 1000.

<Management Port Factor>

Management port factor.

Valid range: integer between 0 and 1000.

<Bond Port Factor>

Bond interface factor.

Valid range: integer between 0 and 1000.

<CMM Factor>

Weight factor for a CMM.

Valid range: integer between 0 and 1000.

<Fans Factor>

Weight factor for a fan unit.

Valid range: integer between 0 and 1000.

<PSU Factor>

Weight factor for a Power Supply Unit.

Valid range: integer between 0 and 1000.

<SSM Factor>

Weight factor for a SSM.

This factor applies to all SSMs.

Valid range: integer between 0 and 1000.

Examples

[Global] HostName-ch01-01 > set cluster configuration high-availability factors sgm 100

[Global] HostName-ch01-01 > set cluster configuration high-availability factors port other 70

[Global] HostName-ch01-01 > set cluster configuration high-availability factors port standard 50

[Global] HostName-ch01-01 > set cluster configuration high-availability factors sensor cmm 40

[Global] HostName-ch01-01 > set cluster configuration high-availability factors sensor fans 30

[Global] HostName-ch01-01 > set cluster configuration high-availability factors sensor power_supplies 20

[Global] HostName-ch01-01 > set cluster configuration high-availability factors sensor ssm 45

Configuring the Chassis ID

You must make sure that the Chassis IDs are different before you start to configure the software.

Chassis IDs are configured on the CMM and should be 1 for the first Chassis and 2 for the second Chassis.

Important - If the Chassis is up and running, change the Chassis ID on the Standby Chassis. You must perform a Chassis failover.

Step

Instructions

1

Pull out the first CMM from the Chassis.

2

Connect to the remaining CMM with a serial cable (baud rate - 9600).

3

Log in with these user name and password: admin / admin

4

Edit the /etc/shmm.cfg file:

vi /etc/shmm.cfg

5

Search for:

SHMM_CHASSIS=

6

Set the correct Chassis ID:

  • For Chassis 1:

    SHMM_CHASSID="1"

  • For Chassis 2:

    SHMM_CHASSID="2"

7

Save the changes in the file and exit the editor.

8

Remove the current CMM and insert the second CMM.

9

Repeat Steps 2 - 6 for the second CMM.

10

Insert both CMMs into the Chassis.

11

Attach the correct identification labels to the Chassis and CMMs.

This step is required if the Chassis has already been configured (after the First Time Configuration Wizard).

12

Pull out all SGMs from the Chassis.

Insert all SGMs into the Chassis.

Important - This step causes a hard reboot of the Chassis.

Configuring the Quality Grade Differential

Description

Use the "set cluster configuration high-availability failover" command in Gaia gClish to set the minimum quality grade differential that causes a failover.

Syntax in Gaia gClish of the Security Group

set cluster configuration high-availability failover <Trigger>

Parameters

Parameter

Description

<Trigger>

Minimum difference in Chassis quality grade to trigger a failover.

Valid range: Integer between 1 and 1000.

Configuring the Failover Freeze Interval

Description

A Standby Chassis cannot failover a second time until the specified failover freeze interval expires.

The default failover freeze interval is:

  • For the "Active Up" chassis configuration - 30 seconds

  • For the "Primary Up" chassis configuration - 150 seconds

  • For VSX Virtual System Load Sharing (VSLS) configuration - 150 seconds

If the Standby Chassis grade changes to a value greater than the minimum quality grade gap for a failover, the Standby Chassis fails over and becomes a new Active.

The failover does not start until the freeze interval expires. This confirms that the Standby Chassis quality grade is stable, before it becomes a new Active.

For example, a Standby Chassis quality grade can become unstable if a fan speed increases and decreases frequently.

Syntax in Gaia gClish of the Security Group

set cluster configuration high-availability freeze_interval <Freeze Interval>

Parameters

Parameter

Description

<Freeze Interval>

Minimum time in seconds to wait until the next Standby Chassis failover.

Valid range: integer between 1 and 1000.

Notes:

  • When you run the "asg stat" command after Standby Chassis failover, the output shows the freeze time.

  • The <Freeze Interval> value is 5 fold greater, if the setup is configured to work in VSLS or "Primary Up" mode.

    Example: If the freeze time must be 250 seconds, you must enter the value 50.

Configuring the Chassis Priority

After you configure the High Availability to work in the "Primary Up Mode" with the command "set cluster configuration high-availability mode 1" (see How Dual Chassis Active/Standby Mode Works), you must configure the chassis priority:

set cluster configuration high-availability vs site_priority "<ID of Primary Chassis> <ID of Secondary Chassis>"

Example - set Chassis 2 to be the Primary over Chassis 1:

[Expert@HostName-ch0x-0x:0]# gclish

[Global] HostName-ch01-01 > set cluster configuration high-availability vs site_priority "2 1"

[Global] HostName-ch01-01 >