This article describes the issue of the SRX chassis cluster not failing over, when the ae aggregate interface is configured under interface-monitor.
The chassis cluster may not failover, if one of the physical interfaces goes down; even though they have 255 as the defined weight and the
ae interface is also being monitored.
redundancy-group 1 {
node 0 priority 254;
node 1 priority 245;
preempt;
interface-monitor {
ge-0/0/2 weight 255;
ge-0/0/3 weight 255;
ae1 weight 255; < ae1 is being monitored for failover
}
}
interface-range ae1-members {
member ge-0/0/0;
member ge-0/0/1;
enable;
traps;
gigether-options {
802.3ad ae1;
}
}
- The chassis cluster interface failover supports only physical interfaces.
- The logical interface should not be defined in interface-monitor. This will cause confusion in the failover logic and the weight will not decrement from 255; so failover will not occur.
- Remove the ae1 interface from the interface-monitor and add the child interface of ae1 in there.
The final configuration is:
redundancy-group 1 {
node 0 priority 254;
node 1 priority 245;
preempt;
interface-monitor {
ge-0/0/2 weight 255;
ge-0/0/3 weight 255;
ge-0/0/0 weight 255; ==> Child interfaces of 'ae1' are added here; instead of the 'ae1' logical interface
ge-0/0/1 weight 255; ==>
}
}
For more information about the supported configuration on the chassis cluster, refer to the following link:
http://www.juniper.net/techpubs/en_US/junos11.1/information-products/topic-collections/security/software-all/security/index.html?topic-47219.html For additional information, refer to
TN10 - SRX High Availability Deployment Guide (Chassis Cluster configuration examples).
For more information about the aggregate interface on an EX switch, refer to
KB22474 - Link aggregation (LACP) supported/non-supported configurations on SRX and EX.