Hello my friends, It has been quite a while since I last blogged but I wanted to take some time to share some of my experience over the past couple of years. I have had the opportunity to work with some great companies, people, and it has definitely been a very enlightening experience.
I had the privilege of being apart of a special project nearly 5 years ago which began my career in the cloud. I got to engineer and deploy one of the nations first ever GSA clouds which was a great experience. As time rolled on and cloud was adopted many things came into the light. Being a VMware savvy guy I really didn’t have all the time to spend learning all these new technologies which were directly competing. At this time, Amazon was getting big, VMware was about to release VRA, and the market stood still… or so it felt.
Microsoft had launched their On-Prem cloud and before we knew it we had to start getting serious about the cost of our delivery and compute. If you have never had the pleasure of working for service providers let me tell you – its all about cost. So we put Azure to the test, compared it, vetted it, did anything we could to ensure it could be operationally supported. It was a very interesting time and nice comparison to our existing IaaS architecture. We definitely had our work cut out for us.
Since then the challenges of hybrid cloud have become real. Although some vendors had good solutions like UCS Inter-Cloud Fab or vCloud Connector… (insert whatever else here) we always seemed to have unique enough requirements to disqualify it. Needless to say we still deployed, stood them up, tested them, and found great value it still wasn’t justifiable enough for us to warrant a change. Being a service provider isn’t about offloading to another cloud… it’s about how you can upsell your services and provide more value for customers.
As time grew on people adopted Cisco UCS into their infrastructures and eventually it seemed like updating and maintaining infrastructure became critical and the speed of delivery is only hindered by how fast we can adopt new offerings.. If we cannot seamlesly update, migrate, or refresh to new then what can we do?
“Its so old its not even supported!”
“Wow, no new firmware for 5 tears?!”
“Support for VMware has lapsed :(”
You can automate this pain away easily. Just because one vendor doesn’t support a feature or a new version does not mean you have to still burden your IT staff. If you can standardize operational processes between your cloud(s), Visibility, Integration, and Support – would you?
The biggest challenge is getting out of the old and into the new. Most legacy infrastructure runs on VMware and you can do this with Turbonomics and a variety of other tools. One of the benefits of going 3rd party is that you don’t have “lock-in” to any infrastructures or software. You can size it, optimize it, price it, and compare it to ensure things run as they should. Versioning, Upgrades, and these things will always be challenge but as long as you can ensure compliance, provisioning, optimization, and performance it won’t be an after thought. I found Turbonomics to always get the job done and always respond in a way that provided a solution and more than that… at a push of a button.
Some of the benefits:
– Agnostic Integration with a large set of vendors
– Automated Provisioning for various types of compute
– Easily retrofit existing infrastructure for migration
– Elastic compute models
– Cost Comparison, Pricing Existing, Etc…
– I.e. Amazon AWS, Azure
– Track and exceed your ROI Goals
– Eliminate Resource Contention
– Automate and Schedule Migrations between Compute Platforms (Iaas > DBaaS)
– Assured performance, control, and automated re-sizing
– Not version dependent and can be used in a wide variety of scenarios – I.e. I can elaborate if needed.
– Get rolling almost instantly with it…
5 years and I still think Turbonomics is a great product. I have used it extensively in the early days and also worked with it during the vCloud Integration piece. The free version is also amazing and very helpful. Spending time checking capacity, double checking data, ensuring things are proper and standard, all that stuff you can forget about it. Configure your clouds; private, public, or dedicated into Turbonomics quickly.
You just have to trust proven software especially if its been 7 years in the making and exceeds capabilities that most tools require significant configuration for. Also, always keep in mind that TURBONOMICS can learn your environment and the value of understanding the platform and providing insight can be huge. You have to admit that some admins may not understand or know other platforms. This simplifies all that by simply understanding the workload and infrastructure that it runs on.
Other Great Information or References:
Cisco One Enterprise Suite – Cisco Workload Optimization Manager:
CWOM offered with
Turbonomics and BMC:
“Running it Red Hot with Turbonomics”
I am sure this topic has been beaten to death and I am sure I am not going to say anything I already may have said. I don’t want to cover how to perform the upgrade but I really wanted to provide some feedback on some post task implementation plains when doing your upgrade. I hope this is helpful to someone out there. For others this may not be something you have to worry about because you may use a standard installation of the components. For me however I find this invaluable because it puts you a position of what you can expect when doing the upgrade. In my experience I have run into every component being configured differently and sad to say it was impacting to time and stress…
The Post Checks:
- Back-up the stuff.. 😉
User Accounts and service accounts for the following:
VMware Update Manager
- Service account usually..
- Web Manager Log in
- Appliance Console log in
vCloud Director Cells
- Root passwords
- Connectivity accounts in the configuration of the portal to vCenter
vCenter Server services
- Service Accounts
- I don’t use it but the same goes – get the accounts and IDs.
Database configurations and user names and access
Vmware Update Manager Database
- ODBC configuration
- SQL Permissions and Access (Use SQL manager to test!!!)
- SQL service account
vCenter Server Database
- ODBC Configuration (64-bit)
- SQL Permissions and Access (Use SQL manager to test!!!)
- SQL service account
VCD Oracle DB
- I’m not an oracle dude but you should just need to make sure its backed up.
I was able to clone and use a test upgrade on my vCenter to check all my configurations. I would highly recommend considering a standard configuration which if you follow VMware documentation you should be fine. Just remember to document things when you have to do them a bit different. One last thing I would recommend is definitely performing backups, snapshots, and even clones if you DON’T have 100% for sure backups. This is invaluable…
One Last Thing…
As a side note I don’t think I would add any of the other features of vCenter 5 unless you’re going to configure them and use them right then. There really is no point installing the other components because if you follow the VMware recommendations on vCenter sizing for each component consumes additional CPU. An example would be like installing the Web Client. It’s important to know what you are going to install and how you are going to configure it. If you are not going to configure it I would not install it. Standardize your implementation and then move forward.
It seems I keep running into things throwing my in a loop and I am sure someone else out there in the VMTN worlds knows that I am talking about. I love vCloud, I enjoy the product, but man some things are still pretty vague although it continue to grow more and more each day I definitely want to do my part to help contribute. So I will make this a short and quick post.
Well take a peek at this VMware Knowledgebase and it will describe in detail the issue I ran into:
“Upgrading vCloud Director 1.5 with an Oracle database to vCloud Director 1.5.1 fails with the error: CALL create_missing_index()”
Although it is hard to say why this is a bug in the database or duplicated entries most oracle DBA’s can knock it out fairly easy. However let me explain to you how this happens and provide some further insight on how to resolve it without having to guess.
Now I want to point out that someone forgot to use the call-management tool (which you can find here) but one would think that this is mostly a database issue because it’s referencing a duplicate table entry. Also this seems to happen with updating vCloud from 1.5 to 1.5.1.
If you run into this issue you need to do the following:
- Restore the Database
- Run the KB Fix of the following:
“DELETE FROM object_condition WHERE object_id IN (SELECT object_id FROM object_condition GROUP BY object_id, object_type, category, condition HAVING count(*) > 1);”
- Then perform you Database upgrade
I know this seems to be pretty simple but the KB doesn’t tell you specifically what you need to do. The Oracle piece was vague in that one would think you need to run the snippet above to fix it. You should just know that this is something that should be ran AFTER the database restore. This may save you some time if you run into it.
So the sad truth.
Regarding my post from a few days ago allow me to elaborate.
VMturbo takes what vCOPS AKA VROPS 6.0 just introduced. Prior it was based on integration with VCO.
Long story short VMturbo has been doing this for a long time. Its time to trust something that can both – automate capacity and properly balance any workload based on SLA.
I have mad respect for VMware and VMturbo but they have been at this for so long. What took me months in vROPS – takes me seconds or hours in VMturbo. I will say VROPS 6.0 is almost NEARLY caught up to VMTurbo…
When I decided to trust VMturbo to manage my workloads the ROI became real. In my experience I have worked for service providers and enterprises but VMturbo deserves a fair chance with any business. When you automate DAY 1 operations.. that is pretty darn nice.
I won’t get into the nitty gritty but it comes down to patented analytics that recommend actions based on performance data…In other words KPIs mixed with Super Metrics and “Insert VROPS expert here” tag…
May the vForce be with you..
VMturbo or vROPS?
I’d go with something that would easily tie in ROI into your capacity. vRealize Operations is nice but as I continue my journey into tooling and reviewing end to end solutions there are various challenges and complexities.
In short, the more you understand about your workloads (Or LEARN through AI) the more optimization and efficiency you can attain.
The key distinction I find is the development and ease of use Turbonomics (VMturbo) brings. It drasticially simplifies your compute management by eliminating dependencies and ensure standard process/features are executed in the most efficient way.
If you want to stay in tune with Industry trends and cutting edge features I would definitely give Turbonomics a fair shot.
#CISCOUCS #WINNING #BIGDATA #UCSDIRECTOR