Skip to main content

There’s another very lengthy post on building LiveSync with floating names where the suggestion was “don’t do it”, then branched into more complex things like separating webservices from the CS, etc. It’s closed for replies now though.

So, we have 3 CommCells all doing LiveSync but no floating names for the CS. We’re needing to refresh many of these servers so my thought would be to implement the floating CS name since it’s overdue, but I’m a bit puzzled since most documentation seems to speak to setting this up “at install” time, not post, many years after the fact. Is https://documentation.commvault.com/2022e/expert/104936_setting_up_failovers_using_floating_commserve_name.html relevant for refreshes and such?

Ideally sure I’d like to split the webservers out from the CS’s per best practices but given the amount of potential change here (new CS’s both primary and standbys), new locations on network (new IPs), new OS’s, etc., where’s the logical place to begin - move things via refresh steps first, then break things out or could the move to floating names be done first thereby easing the moves to the new infrastructure? We’ll have all 3 on 11.28 before any of this takes place if it matters.

thanks in advance

Using sCSINTERFACELIST gives you a bit of flexibility of what to do here.

 

My recommendation would be to add all the unique floatingnames for each machine to sCSINTERFACELIST that way all the clients will have the settings ahead of time.

 

You can either install the CS with or without the floatingname.

 

For webservers, you can split off before refresh to avoid installing webservers on all the standby CS.

Also use TPPM to configure access to the. See doc here - https://documentation.commvault.com/2022e/expert/95761_setting_up_standby_commserve_using_port_forwarding_gateway_configuration.html

 

Hi @Leo E 

I did see that in the docs, and each CommCell has that setting but only with the FQDNs of the primary and standby names. I guess the question is, this seems to have been added at some point along the way (SP14 as noted?) to negate “needing” floating names or aliases like say many years back? Meaning, with that key in place and all clients in the special Clients For Commserv LiveSync group which it then applies, clients should know who to talk to once the “switch” has been made in either direction? For some reason I had it stuck in my brain that floating names were “always” wanted so just wanted to confirm. The IP’s shouldn’t need to be in there if DNS is correct (correct?)

thanks

Yes, the clients will connect to the running /active node after failover.

No need to add the IP if the FQDN is correctly resolved to an IP via DNS.

Floatingname is a good thing and recommended but must be implemented as virtual names. Each CS must have a unique virtual name.

Reply


Terms & ConditionsCookie settings