Splitting 2 production Viprion clusters(active-stdby) into two..

When you are changing back the config setting to REdundantPair you got all the settings back in place ? and cluster redundancy works again without any other changes necessary, just as it was before the change to SingleDevice ?yes.

Great, but another question is what will happen if I change only to SingleDevice and leave the other bits of config ? I think it is better to delete all the failover/failsafe trigger events ? I can imagine it can try to failover even though there will be no peer/stdby cluster ?

 

can we be 100% sure that just changing to SingleDevice will disable any failover actions from occuring, eventhough some bits of config for it will be still there in conf files ?

if i were you, i would restore ucs if i want to revert back to active/standby configuration rather than changing to single device and leaving other ha setting.

Posted By nitass on 01/31/2012 06:45 AM

 

if i were you, i would restore ucs if i want to revert back to active/standby configuration rather than changing to single device and leaving other ha setting. no this is not something I'm aiming to achieve.

 

I'm considering some option how to do this split the best way.. You've only changed the HA to SingleDevice and that's all. The GUI will hide probably all the HA/Redundancy/FAilover options once that changed, but you shown me that in conf files and probably in SCF all that HA config is still kept and only some flags has changed, so this is not the cleanest way especially thinking about someone else doing changes anytime later. This config might be a surprise on a SingleDevice system type for someone else later.

 

 

On the other hand if I start to remove everything I might casue production traffic interruption. I wonder for example how would all the floating IPs behave once you changed to SingleDevice ?

 

and If I would like to decomission them I would need to probablt do something like this:

 

- export SCF and then delete all of them, i.e from that SCF section:

 

 

shell write partition Common

 

self 1.2.3.4 {

 

netmask 255.255.255.0

 

unit 1

 

floating enable

 

vlan TEST_VLAN

 

allow all

 

}

 

(...)

 

 

and add a normal self IP for each deleted floating one, ie.:

 

 

self 1.2.3.4 {

 

netmask 255.255.255.0

 

vlan TEST_VLAN

 

allow all

 

}

 

and import that new SCF.

 

But I guess this may casue also service interruption since IP will be reconfigured (deleted and added again) ?

 

 

Thanks for continued support,

 

Tom

 

 

 

I wonder for example how would all the floating IPs behave once you changed to SingleDevice ?i understand floating selfip works just find even it is standlone. the different is it won't be able to float to another device.

 

 

But I guess this may casue also service interruption since IP will be reconfigured (deleted and added again) ?i agree. it will interrupt traffic.

 

 

by the way, do you really need to remove floating selfip?

no I don't have to remove them. It is up to me to decide. The reason I wanted to do this is to have as clean config as possible and I wasn't sure how the floating IPs will behave in a singleDevice mode? For example looking into conf files I found out that normal self IPs are in bigip_base.conf whereas the floating ones are kept in bigip.conf which is shared config in redundant mode I think. So was wondering if they will not be deleted when changing to SingleDevice?

 

Also looking into conf guides:

 

 

"

 

The Floating IP setting appears on the screen only when the BIG-IP system

 

is configured as a unit of a redundant system. For more information on

 

configuring a redundant system, see Chapter 20, Configuring High

 

Availability.

 

"

 

so not sure what will be shown on the GUI when switch to standalone done? I would like to avoid situation when GUI shows one thing to the user and SCF/conf files have something more or different. Additionally software is usually written to work in a predictable configuration sets. Keeping floating IPs with a SingleDevice mode looked to me as a very non standard thing and I'm still having concerns if this will not unveil some software bugs or issues later. Also If config gets messy then it can fog some thigns and comlpicate life later.

 

But because of this IP reconfiguration traffic loss will occur I might get into more seriuos problems (stp recalculation,arp caches to expire, ipmp failovers on back-end servers etc.) I think I would leave those flaoting IPs as they are if I only know for sure that they will work fine in SingleDevice mode and will not cause any config/upgrade issues later ?

 

 

Thanks very much for continued support on this,

 

Tom

as i see, floating selfip is still shown in webui after changing to single device. however, it cannot be newly created via webui since floating ip option is gone. anyway, creating using command line is working fine.

 

 

I think I would leave those flaoting IPs as they are if I only know for sure that they will work fine in SingleDevice mode and will not cause any config/upgrade issues later ? i cannot guarantee this and not sure if support engineer is able to answer. anyway, you may try to open a support case and see what they say.

 

 

cheer!