Viprion 2400 -- migrating from 2150B to 2250B

[Dexter]

Hi All,

I am planing to migrate current 2150B to 2250B on 2400 Chassis. Since its being a while, here are the steps I am planning to perform in order to do successful migration. Kindly guide me if any point is missed or advise for a successful migration :

Note : 2 2400 Chassis with single 2150B

vCMP

1. Create similar Host/Guest configuration on 2250B with same VLAN, Trunks and Interface configuration
2. Assign same amount of resources to each Guest

Guest

1. Create similar Vlans, Trunks, Interface Configurations
2. Migrate Archives from Old 2150B to new 2250B
3. Since 2250 is new and most probably with new version, I'll have to downgrade the version to match the same version as running on 2150B in order to successfully migrate.
4. Verify all configurations under LTM, ASM, APM and GTM

One thing, since mixing of different blades aren't possible under same Chassis, who will I configure the Blade with above mentioned steps. Moreover, do i need to configure the same 2400Chassis once 2250 is installed, apologies as i'm newbiw ^_^

[NAG]

Hi,

"vCMP

1. Create similar Host/Guest configuration on 2250B with same VLAN, Trunks and Interface configuration
2. Assign same amount of resources to each Guest

"

There is no need to recreate Host config with VLANS and Guests. You can take a UCS Archive backup of Host on old blade before replacing it. And then, after the old blade is replaced with new one, you can simply activate the license and restore config using UCS archive.

Following articles help you in the processes:

https://support.f5.com/csp/article/K82540512

https://support.f5.com/csp/article/K4423

"Since 2250 is new and most probably with new version, I'll have to downgrade the version to match the same version as running on 2150B in order to successfully migrate."

It depends on which version you are currently using. New blades support certain TMOS version and above. You cannot downgrade beyond certain version.

"One thing, since mixing of different blades aren't possible under same Chassis, who will I configure the Blade with above mentioned steps. Moreover, do i need to configure the same 2400Chassis once 2250 is installed, apologies as i'm newbiw ^_^"

I am not sure if I understood well. You cannot run 2 different blades at once.

Very High level POA::

1) Take UCS backup of Host and all its gusts

2) Swap the blades

3) reactivate the License

4) Restore Host UCS and validate

5) Restore Guest UCS files and validate

Hope this helps.

Nag

[Dexter]

thanks   for the detailed info. I know I cannot use 2 different blades at a single time from a same chassis, this means that I'll have to remove the one in production (all business services are down) and then do the migration. Sounds risky.

One thing more, once I swap in the new blade, is there any thing i need to configure/install ? I mean do I need to configure anything from Viprion side or simple connect the console with the 2250 and start migration. Also, I believe I have to go through initial configurations on F5 before I can reach the UCS part, correct?

[NAG]

1) you said you have 2 viprion chassis. I assume you are using them to host active-standby setup for guests. If that's the true, all the guests are active on one Viprion host so you can carry out change on other Viprion chassis without disrupting prodution services.

2) once you swap the blades, configure management IP and activate license. And then upload Host UCS archive from previous setup and restore it.

Hope it helps,​

Nag​

[Dexter]

Sadly thats not the case. Both Chasiss are standalone, i know its a design issue but I'm stuck now. I believe the only option is to shutdown the complete chasiss and install the new blade and start from scratch, this will result in huge downtime which won't be accepted​

[NAG]

vCMP hosts running on Viprions will always be standalone.. it's the guests which run upon them are setup as HA pairs. If A and B are in HA pair, if A is active on Host 1, B will be standby on Host 2.

Check if this is your setup, you can manage without downtime.​

[Dexter]

That too sadly isn't the case either. Both Chassis, Hosts and Guests are seperate environment from each other. Looks like no way around other then to shut the production down :/​

[NAG]

Looks like, there are no measures to minimise down time are built into the Architecture. Unless you use a new chassis with new blade, there is no way you can avoid downtime.

[Dexter]

Yeah, that sucks. One more thing, i believe all I need to do is to first configure the new Blades with Interfaces (IP's, Trunks e.t.c.) and then simply restore the UCS taken from old vCMP to new vCMP.

This will I believe restores the Guests as well with Network Configurations, right? Or do I need to setup Guest same as deployed in current production and then only restore UCS on guest level once the Network configurations are done?

[Maithili]

  can you please let me know what issues have you seen when you replaced the blades. Did you follow the process that   mentioned above. Also had few questions

1. is the base registration key tied to the chassis/blade and do we need to do a license transfer from 2150 to 2250.
2. i saw that on f5 documentation they mention about using modify sys crypto master-key prompt-for-password for encrypted password. Also saw f5mku -K for ssl profiles. Should we use both or is using the master-key is good for ssl profiles and user passwords?

if anybody has used this please let me know your experience with it. I am a newbie, Thank you for your help