SRX

here is the cluster :

  node0---node1

rg1 is primary in node1 since there is one reth down in node0(unhealthy)

rg0 is primary in node0 since the higher priority

1:now if contril link between node0 and node1 is down,what will happen(fabric link is good)?

does node0 change inactive redundancy groups 1 to priamy even there is an interface down in node0?

2:if control link is good ,but fabric link down

what will happen?

Hi, 

From SRX HA Deployment Guide:

Control Link communication loss only (Fabric communication is still successful)
The RG0 secondary node will transition to an Ineligible state then transistion to a Disabled state

Fabric Link communication loss only (Control communication is still successful)
SRX-HE
Secondary node will transition to Ineligible then to Disabled state (10.4R3 and lower)
No action taken(10.4R4 and higher)
SRX-Branch
Secondary node will transition to an Ineligible state then to a Disabled state

----------------------------------------------------------------------------------------------------------------------------------------------------

1:now if contril link between node0 and node1 is down,what will happen(fabric link is good)?

does node0 change inactive redundancy groups 1 to priamy even there is an interface down in node0?

2:if control link is good ,but fabric link down

what will happen?

----------------------------------------------------------------------------------------------------------------------------------------------------

1. Only RG0 on secondary node will transition to disabled. I believe RG0 will not be able to communicate to PFE on node1.

"The RE communication with the remote node, such as communication with PFE, kernel state synchronization with RE, configuration synchronization, and Juniper Stateful Redundancy Protocol (JSRP) heartbeats between the nodes is handled via the control link."

In this case reth in RG1 will go down.

2. I believe failure of fabric link only would affect Z-Flow traffic or session sync for active-active cluster [10.4R4 higher]

Cheers,

Ashvin

hi

thanks for your answer.

for control link failure:

rg0 : primary node0(one interface is down in node0 dataplane ,so rg1 is active in node1)

rg1:primary node1

even in this case, when control link failure happens, node0 will still take primary role for rg1?

Hi, 

I believe so as the RE [RG0 on node0] will not be able to communicate with the PFE of node 1 when control link is down.

Cheers,

Ashvin

Hi Robbie,

The following statement answers your query :-

In the event of a legitimate control link failure, redundancy group 0 remains primary on the node on which it is currently primary, inactive redundancy groups x on the primary node become active, and the secondary node enters a disabled state.

Redundancy group 0 remains primary on the node on which it is presently primary (and thus its Routing Engine remains active), and all redundancy groups x on the node become primary.

This can be found in the following link :-

http://www.juniper.net/documentation/en_US/junos12.1x44/topics/concept/chassis-cluster-control-link-failure-recovery-understanding.html

This essentially means that when there is a control link failure, the RG0 would remain primary on the same node 0 in your case, and all the other RGs would also become primary on the node 0 itself. The node 1 would go into the disabled state to prevent a split brain situation in the network.

Hope this helps.

Regards,

Sahil Sharma

---------------------------------------------------

Please mark my solution as accepted if it helped, Kudos are appreciated as well.