1400 Manual Cluster Failover

[biskit]

Hi all,

I'm struggling to remember or find the answer to how you manually fail over a 1400 cluster.  The normal "clusterXL_Admin down" doesn't work on 1400's.

I've set the other member to the higher priority in SmartDashboard and set ClusterXL to switch to higher priority member, then installed, but it hasn't switched.

Any ideas how I can manually fail over between cluster members on 1400's?

Thanks,

Matt

[Matt_Ricketts]

Are you able to physically unplug one of the Ethernet cables in the 1400 and force fail over?

[Mark_Mitchell]

The SMB devices are a little bit different as they run Gaia embedded. Performing the following will perform the failover in the same manner as the Gaia full devices (3000 series and above).

# Complete in Expert mode

cd $FWDIR/bin

clusterXL_admin.csh down

[G_W_Albrecht]

That is all documeted here: sk65060: How to cause a manual fail-over in ClusterXL on SG-8x appliances.

[Maarten_Sjouw]

The actual simplest way to do so, is by using the priority in the SmartConsole, as it is centrally managed, that is where it should be, open the cluster object, Cluster members and adjust the priority and push the policy.

For the quick switch the only way is to use the CLI as described in the previous posts.

[G_W_Albrecht]

If it is centrally managed - even 14x0 SMBs can be used locally managed.

[Mark_Mitchell]

When using the priorities (with central management with SMS) if the failover settings within advanced settings are set to "Maintain Current Active Cluster Member" then the unit won't failover based on the priorities. 

If you want to perform failover using the priorities you will need to change to "Switch to higher priority cluster member" and then install policy. 

[Gaurav_Pandya]

Correct. You need to change option to "Switch to higher priority cluster member" in clusterXL setting. 

[biskit]

Thanks everyone.

I've already set to highest priority and installed (as mentioned in the original post) but this did not switch members.  I don't know why not.

I'll try the clusterXL_admin.csh command tonight during a maintenance window

Matt

[Mark_Mitchell]

Not a problem at all. 

Just to confirm within your SMS have you set the "Switch to higher priority Cluster Member" option as pictured earlier?

If it's easier, feel free to post a screenshot and we can see if we can see anything that may be causing the issue.

[Mark_Mitchell]

Hey Matt, how did you get on within your maintenance window and the failover with the command clusterXL_admin.csh?

[biskit]

Here's my settings:

However, even after multiple installs, FWA is still Standby.  It should now be Active.

[Mark_Mitchell]

Very Interesting, it looks ok to me from a clusterXL configuration point of view. What firmware version are you running? 

I have a pair of 1450 for lab use so I'll see if I can replicate the same results. 

[biskit]

Thanks.  I'm currently running on R77.20.70.

[JozkoMrkvicka]

There is no option in SmartView Monitor to perform failover ? Try to right click on active member within Monitor.

[Mark_Mitchell]

Please see below for SK for Best Practices on Manual Check Point failover. 

Best Practices - Manual fail-over in ClusterXL 

I would always recommend this method of a fail over. As from the CLI you can also obtain more detailed data quickly around clustered interfaces and cluster health etc. 

[biskit]

The clusterXL_admin.csh down worked a treat

One day the commands might be consistent between CP appliance models 

[Martin_Valenta]

in expert mode, ip link set dev LANx down/up