Things to verify after migration!

Hi,

First of all I would check that everything is working while the 3600 is online. This way you know every application is working as it should. It takes a lot of time and effort to focus on a problem that was already there before migrating.

You mention the 4000 is not online yet. Very hard to check if everything is correct. Is a different Management IP configured on the 4000? If so, you can compare the configuration by having both GUI's open at the same time. That's the easy way. If this is not possible because Management IP stays the same, connect another computer to the 4000 and have both GUI's open on different computers.

You can also create a Singe Configuration File on both appliances and run a compare of both files. Maybe you see differences you can discard, but maybe you see differences that might cause the migration to fail. Just a extra check before migrating.

What is your migration plan? Disabling switch-ports connected to the 3600 and enable switch-ports connected to the 4000? The roll back procedure is very simple this way.

Is the 3600 a cluster setup. In the past we have migrated to new hardware by adding the new hardware to the device group, do a configuration sync and remove the old hardware. Software version must be the same and you need free IP-addressen on the production network for the new hardware, but it is something to consider.

In the end you need a good test-plan and involve application owners and users for testing. The GUI is the first indication if everything is OK. Are interfaces up? Are VS's UP? Are nodes up? Look at the statistics of VS's. If the counter increases, you know the F5 is processing traffic.

If things are failing, tools like tcpdump can be very usefull. Also check the /var/log/ltm files.

Good luck with you migration.

Martijn

Hi,

First of all I would check that everything is working while the 3600 is online. This way you know every application is working as it should. It takes a lot of time and effort to focus on a problem that was already there before migrating.

You mention the 4000 is not online yet. Very hard to check if everything is correct. Is a different Management IP configured on the 4000? If so, you can compare the configuration by having both GUI's open at the same time. That's the easy way. If this is not possible because Management IP stays the same, connect another computer to the 4000 and have both GUI's open on different computers.

You can also create a Singe Configuration File on both appliances and run a compare of both files. Maybe you see differences you can discard, but maybe you see differences that might cause the migration to fail. Just a extra check before migrating.

What is your migration plan? Disabling switch-ports connected to the 3600 and enable switch-ports connected to the 4000? The roll back procedure is very simple this way.

Is the 3600 a cluster setup. In the past we have migrated to new hardware by adding the new hardware to the device group, do a configuration sync and remove the old hardware. Software version must be the same and you need free IP-addressen on the production network for the new hardware, but it is something to consider.

In the end you need a good test-plan and involve application owners and users for testing. The GUI is the first indication if everything is OK. Are interfaces up? Are VS's UP? Are nodes up? Look at the statistics of VS's. If the counter increases, you know the F5 is processing traffic.

If things are failing, tools like tcpdump can be very usefull. Also check the /var/log/ltm files.

Good luck with you migration.

Martijn

1. Run commands before and after the maintenance to identify the status of the VS and Pools.

    

2. Check logs in the new LTM after upgrade to make sure nothing is broken.

    

3. Check application function after upgrade.