EMCCloud

Most of the hypervisors support some form of vm migration so that you can properly maintain your hardware, the end result is more dependent on architecture choices

In my opinion, it is not "out of the box" ready. There are significant design considerations that include load balancers, compute redundancy, and layer-3 fault domains that would allow you to design a highly available infra.

correct. Your HA-ness will be almost entirely dependent upon your architecture.

there are lots of options. We generally apply incremental updates online/daemon restart but are more cautious with the changes from one release to another.

congrats @noggin143 and team!

thx Danny: that is a great video. Nice share

Congrats @noggin143 and the CERN team! :-D

I didn't go to Paris but I heard that was the focus

At the beginning there was so much to the command line that you needed to know. From a pure "L1" operator standpoint, the advances in the 'GUI' have helped tremendously!

while in the early days Ops was clearly an afterthought, I do like the work being done in the last year or so to engage ops: full days at the summits, mid-cycles, etc. Chance to share war stories/best practices.

also it seem that most (if not all) of the projects are taking a hard(er) look at what it means to run this thing.

sat in on a design session about logging for instance: we know it is broken, and we want to get it fixed.

@craig_tracey @relaxed137 Any upcoming changes in Kilo+ that you are looking forward to as operators?

Unless your average 'NOC' person can run and support your cloud, it won't penetrate fully. integration into common DC tooling for logging/monitoring needs to continue. It's come a ways, but needs more.

Upgrade experience has significantly improved in the last release. Since we went production in July 2013, we've done two upgrades without impact on running VMs.

Some details at http://openstack-in-...

It's gotten easier to manage over time but still a long way to go especially in the areas of logging and monitoring. Menesca introduces some good ideas, but I'm not sold yet.

@ShamailXD looking forward to better logging, more robust policy management and a common CLI.

Many people are skeptical of openstack based upon the social networks but I'm not an expert

I don't understand EMCs connection to openstack? Can you explain for me who is in tech but not in the deep insider #openstack #cloud world

@noggin143 what would you say are 'must have' tools in your ops toolchain?

@mikeonsing They make storage and own VMware and IT enterprises all want openstack not pure AWS cloud

OpenStack is loosing / has lost the hyperscale folks who one w/d think w/d want the 100% API driven infrastructure of OpenStack. And is too tough for mainstream enterprise. Or am I off base?

@mikeonsing like many tech leaders, #EMC is interested in #Openstack

@mikeonsing #EMC joined #Openstack consortium back on 2012

Defiinition of mainstream enterprise? Which customers driving this? Finance, hyper-enterprise, retail, Operators?

@vmtyler There are lots of options. The user survey at http://superuser.ope... gives the common ones.

@epowell101 there are enterprises who need to build their own and enterprises who want to buy it "in a box" which IMHO is why we see a lot of "in a box" players right now. It does need to "JUS WORK" for your average enterprise.

@mikeonsing #EMC first contributed to help integrate leading storage platforms to #OpenStack releases

@epowell101 I am hearing tons of FUD on openstack and hearing great stuff. Someone is getting signals crossed.

@mikeonsing leading enterprise storage platforms like @EMCVMAX and @EMC_VNX are fully integrated with #cinder #openstack storage driver

@relaxed137 Maybe... I think there is a moving target. If we converge OpenStack towards today's enterprise operations we will miss the shift towards "real" DevOps which is a real (and complicated) trend.

@vmtyler For CERN, we selected Puppet, Kibana/Elastic Search as our config management and monitoring. At mininum, config mgmt is a must. We also use RDO as the base for our installation of OpenStack.

@mikeonsing #EMC object storage strategy is compatible with #swift and #glance #openstack services

@noggin143 How do you wire together the components? mCollective plus scripts plus Puppet and so forth? Very useful info.

@epowell101 Dev ops though is iterative and requires infrastructure that allows you to iterate. Code based 'Dev ops', dynamic security controls, and dynamic 'cloud' updates are all required. but it must all work

@mikeonsing as #EMC sees a lot of development around #Openstack the company continues to invest

@mikeonsing #EMC has recently acquired @cloudscaling a leading #OpenStack solutions provider

@relaxed137 Agreed. DevOps = far bigger trend than simply cloud X or Y imo. Cloud X or Y or tool Z are just enablers for actually achieving step function agility boost.

@mikeonsing case you haven't seen https://www.emc.com/...

@epowell101 We use the puppetlabs OpenStack modules and rpms based on the RDO packages. some care is needed to avoid all the cloud upgrading at once so we stage updates through QA and cell by cell.

@mikeonsing and in the words of @randybias http://www.cloudscal...

@mikeonsing #EMC has announced integrated #HybridCloud @OpenStack flavored solutions for 2015

@mikeonsing another good blog post from @hui_kenneth http://cloudarchitec...

@vGazza @mikeonsing #EMC announced solutions that will span the VMWare, OpenStack and Microsoft cloud solutions! Hybrid cloud is evolving.

@seanmwinn @mikeonsing after all #Cloud should be about choice!

@epowell101 You're off base. OpenStack has no value for the true hyper scale folks. Just smaller web/tech outfits.

@epowell101 OpenStack and devops are only indirectly related. Mostly orthogonal.

@noggin143 What about the significant changes you made in the networking stack?

remember to comment in threads under questions for maximum virality and votes