SRX

HI experts,

can anybody give me  proper steps of rebooting SRX firewall in cluster?

thanks

Hi bc456 ,

I understand that you want to reboot the devices in cluster:

It depends on which node you want to reboot first.

If the traffic outage is fine ( generally used during upgrades) then reboot the device in given order:

Note: Ensure all RG groups are primary on one node:

1. Reboot Primary node first
2. within few seconds , reboot secondary node.

If you do not want traffic outage , then follow the given steps:

Note: Ensure all RG groups are primary on one node:

1. Reboot secondary node (Node1)first
2. wait for the device is reboot and come back online.
3. check the cluster status and confirm that priority of both nodes for both groups should have configured values.
4. Failover RG0 and RG1 and other RG groups to Node1
5. verify the traffic flow and if everything works fine, then
6. Reboot Node0
7. Once the device comes online , check the cluster status and also verify the priority for RG groups.

By default , when you do Manual failovers of RG groups , 255 will be priority of that primary node.

You need to reset it using the following command:

request chassis cluster failover reset redundancy-group x

x- redundancy group (either 0 0r 1 0r 2)

Manual failover commands:

request chassis cluster failover redundancy-group x node x

x- redundacy group number and also Node id either 0 or 1

Regards,
rparthi
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

[Please Mark My Solution Accepted if it Helped, Kudos are Appreciated Too] .....

Hi,

Thank you for your reply. I tried the 1st method. Rebooting the primary node followed secondary node become a split brain. both node 0 and node 1 become primary node.

I did something similar to method 2 in the following order

1- failover RG 1 and RG0 to node 1

2- from node 1, log in to node 0

3- reboot node 0

4- wait till node 0 come back

5- reset the manual failover of RG 1 and RG 0

6-failover RG 1 and RG 0 to node 0

7- from node 0, log in to node 1

8- reboot node 1

9- wait till it comes back

Thank you,

Hi bc456,

Thanks for the update.

Rebooting both nodes at the same time within few seconds after interval will not cause split brain.

Split brain situation can happen when both control and fab links are down.

Your second method was also a good way to reboot the devices.

Thanks and Regards,

rparthi

[Please Mark My Solution Accepted if it Helped, Kudos are Appreciated Too] .....

rparthi,

Thank you for your explanation. I was also looking for the correct way to reboot a cluster without service interruption.
Just out of curiosity: what would happen if you just reboot the active node without doing the manual fail-over of the redundancy groups first? Does this cause a service interruption or would the second node take over immediately?

Regards,

Steven

Hello Steven

When you reboot primary node, automatic seamless fail-over should happen, and there will not be any service interuption.

Regards,

Raveen

Thanks, Raveen