08-03-2015 09:21 AM - edited 03-01-2019 12:18 PM
Is anyone actually using UCS Central in a production environment?
As a whole it is buggy and has lots of quirks.. If you are using in a production environment can you give details on how many UCSM are connected to it?
Thanks in advance.
08-04-2015 02:52 AM
Hi Eric
This is a valid question ? and there are many options that may impact the answer, e.g.
- green field / brown field setup ?
- do you require global service profiles (and therefore pools....)
I always recommend customers to introduce UCS Central in a non business critical fashion (brown field) ; e.g. consolidation of inventory, errors, maybe global NTP, DNS configuration, UCS backups. No global SP, pools,....
Once they feel comfortable, they could extend the functionality, eg. loading firmware and distributing to UCS domains, global SP and pools,....
UCS Central 1.3 introduced html v5, which is nice, but I was seeing a few caveats; going back to the old interface.
Walter.
08-04-2015 03:26 AM
you answered the question, even you don't have full confidence in ucs central to push it as a global production solution from the start. =)
08-10-2015 09:20 AM
I do, and I finished a Greenfield project last year where it was instrumental in the success. I had 4 sites in 3 countries with 150+ pieces of Cisco gear to run VMware on.
The kicker was that I used the UCS Emulator with UCSC to do the testing of all the environments and define the profiles and taxonomy. It was sensational.
I literally had ever site up and operational for OS deployment within 24 hours (northbound and SAN connectivity) due to the fact with UCSC you just allocate the UCSM to the right leaf and voila - all policies, VLANs, profiles etc are pushed.
Absolutely brilliant.
The NTP bug was also a point where touching the NTP settings in a single spot made it easier to add and remove rather than the fat Java client or CLI.
I do not use Firmware Catalogs or upgrades via UCSC - i prefer to keep those entirely subject to local change management. Everything else is enabled via UCSC.
One caveat - It's a bear if you need to alter a fabric IP for a cluster. Still working around that in one site since the 'Official' option is basically to completely remove the config from a UCS (stupid requirement). So i'm looking at kludges to that to avoid.
But overall, really a great tool if you are managing a multitude of sites in locations remote, or many many stacks with requirements for consistent conventions for infrastructure policies (which we all should have!)
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide