Wednesday, March 25, 2015

Dual VIC or not to Dual VIC - The case against redundant UCS VICs

I answer lots of questions internally via email, and I hate that it's trapped behind walls. In the spirit of doing something a little different and to lob it over the walls, I'm just going to start posting them. 

I'll call this #RealTalk - I'll redact anything that could point back to the real customer - but if you're on Twitter I'll leave your information intact. 

This first one seems to come up often. The case for redundant VICs on Cisco UCS. The main reason I'm against it??.... if your application is so damn important, why is it running on a single host?! It adds more complexity, I'm against complexity. 

Anyway, here's the raw email .....
Hey Scott, 
Thanks this helps a lot. The old team is almost 100%. <redacted> left for EMC, <redacted> is a PSS for another team and @Vallard bailed to the cloud team. Back fills are being hired and we should have a full team by the end of the month. Its like having a whole new team.  I don’t think  we will make our numbers. The K12’s are tough. 

From: "Scott Hanson (scohanso)" <>
Date: Tuesday, March 24, 2015 at 7:03 PM
To: <redacted>
Subject: Re: VMware Host

I've seen customers do it when less than life's are on the line. 

I generally talk customers out of doing it - if the app is that important, it's not running on a single host anyway, there's some type of application cluster, OS cluster, or VMware Fault Tolerance involved. 

Also the failure rate of anything without moving parts rounds to zero. In the past I've pointed people to Chris Aitkinson's blog. He's an admin for Travelport and at the time he wrote this blog they had 1900 servers and 1 VIC failure. Not sure if he still blogs, but there's a Google cache version still -

For those reasons, I think it's overkill. 

However, in this case, maybe it's not. Not being able to book a trip is different than a trauma unit. 

Ultimately their call, but I'd lay it out and let them decide. 

If they do it, make sure they have an understanding of placement policies, as to avoid accidentally placing the redundant vNICs on the same adapter -

How's the old team doing? You guys making your numbers and getting some $$$ .... It's actually a little light over here, hopefully picks up at the end. 

Scott Hanson - @CiscoServerGeek
Consulting Systems Engineer
US Enterprise - Data Center

On Mar 24, 2015, at 6:37 <redacted> wrote:

Hey Scott, 
Hope you are doing well. Your thoughts on the question below . Should they get the 1380 VIC for redundancy ? My gut feeling tells me yes for best practice since it is a single point a failure and these guys are a TIER 1 trauma center.  Any reason why we should not add another VIC for redundancy ? Will the cluster be sufficient for redundancy ? 

From: <redacted>
Date: Tuesday, March 24, 2015 at 4:22 PM
To: <redacted>
Subject: VMware Host



If you were going to build an ESX cluster with UCS B200 M4 blades and were going to put Tier-1 applications on this cluster would you rely solely on the built-in 1340 VICs?  Or would you purchase the 1380 VICs as well in order to create VIC redundancy in VMware as shown below?


As in:


1340 VIC        1380 VIC       Purpose

vmnic0         vmnic3          Mgmt

vmnic1         vmnic4          VMotion

vmnic2         vmnic3          Production VMs


This way if a VIC failed you would still be up and connected to Ethernet and Storage.


Or do you think this is “overkill”.


Monday, October 27, 2014

The only 3 things you really need to know about Cisco UCS

I talk to many customers about Cisco UCS. It always boils down to these 3 talking points. If you've only got 10 minutes to talk about UCS, have this conversation.

Unified Fabric

Cisco UCS leads in the simplicity and flexibility of our fabric design. Through innovations in our ASICs on our Virtual Interface Cards (VICs) we are able to create up to 256 virtual ethernet and/or fiber channel interfaces. These connections are created in software giving you the flexibility to simply change connections without changing cables.

Down the road your applications might require a different connection architecture, and UCS allows you to do that in software. Our design is based of a pair of fabric interconnects and either a 4 port or 8 ports IO Module in the chassis. Other designs offer "choices" for the switching fabrics. In the case of Dell, they have a 93 page guide that walks you through the choices ( I'd argue that those are the choices you don't need to make these days. How you choose to connect things should be determined by application demands, not some piece of hardware.

Do we really want to go back to the days of choosing Token-Ring, ARCnet, FDDI, Thicknet, etc .. and then choosing NetBIOS, NetBEUI, AppleTalk, 3270SNA, IPX/SPX, etc ?? Just give me the IOPs to meet my application needs and a pathway to connect the communication.

Simplified Management

With Cisco UCS we don’t have the concept of management consoles and agents. It's all built into our hardware from the ground up. There's no Dell OMSA or, HP Insight Management agents. No Dell Management Console (DMC), Dell ITA, or HP Insight console for Cisco UCS. All the functionality is built into the platform.

When you want to upgrade firmware, there's no complicated matrix of what needs to be updated first or what order things should be updated. It's all handled by the system.

I have a customer that used to plan an all weekend outage to upgrade their HP blade systems with Virtual Connect. The first time we upgraded their Cisco UCS system, we were done in 3 hours, and that's including the time it took to hand hold them through the process. They now have more UCS blades than they had HP systems and it still only takes a couple of hours to perform upgrades. I've got another customer that does them during the day and experiences ZERO downtime during the process.

Stateless Computing

That last thing that separates us is the statelessness of our systems. In most servers the things that define the server live on the server. Things like MAC address, UUIDs, WWPNs, WWNNs, Firmware, BIOS settings, number of Ethernet and FC interfaces, etc. If you want to examine or change the state of those items you have to boot the server, hit F1 to go into the BIOS, change things, then F10 save and exit.

Not the case with Cisco UCS.

In UCS everything is stored in the system as either a Resource Pool, or a Policy. Those are then used to define a Service Profile. The Service Profile is then "associated" with a blade or rack server and it's programmed with the state information contained in the Service Profile. 

Service Profiles are portable, I can associate them with different blades or rack servers. The "identify" of the server remains the same no matter where I move it. This means other things in your infrastructure like zoning on switches, mapping of LUNS, network items tied to MACS don't have to change.

If you'd like to see Cisco UCS in action, I'd recommend watching this 1 hour demo I did a few months ago. No PowerPoint slides, just some whiteboarding and a demo of the system showing the power of Cisco UCS Unified Fabric, Simplified Management and Stateless Computing.

Monday, March 31, 2014

RIP Server - Rise of the Service

I woke up this morning to an article written by Peter Coffee - RIP, the server. It's time to breathe the air of cloud connection.

It's definitely a sensational eye-ball catching headline when you read "RIP, the server." You immediately think, "Did Peter just say the server is dead?!" ... and certainly many of the readers did as you browse through the comments.

  • "Excuse me, but what else is the Cloud made of but servers!"
  • "Last time I checked, "the cloud" was running on servers...and lots of them"
  • "What does he think those Cloud environments run on....air?"
  • "you idiot. the cloud runs on servers."
  • "the moron who wrote this doesn't know tech. Cloud is a bunch of servers -_-"

Of course servers aren't dead. This is the challenge when a visionary piece hits the street and lands in a zombie herd of technical people. (Sorry too much Walking Dead last night) :-) ) Big shifts in how we do things don't happen overnight. They happen slowly, and while you're buried in your day to day tactical activities you can't see the shift. You poke your head up from time to time and "poo-poo" anyone that says something different from the way you are doing it. 

Back in the early 80's, I was a BBS kid. I grew up on 300/1200 baud modems and trading software on underground BBS's. One night while browsing through the boards I can across a file name "Madonna_Like_A_Virgin_Digitized.sound" ... well not exactly, this was early 80's so it was probably named MDLKVIRG.IFF .. don't you all miss the 8.3 days ? :-) Anyway, it was a HUGE file at the time and was going to take all night to download. I setup my YMODEM transfer and went to bed. (I actually worked with Ward Christensen later at IBM - It was kind of a big deal to meet him)

The next morning I woke up, super excited to play the file. I loaded it up, pressed play, and for a full 10 seconds I could barely make out the screeching and scratching over the PC speakers of Madonna singing Like a Virgin. It sounded like shit. I remember banging my fist on my homemade computer desk and screaming "Who the F#!K is going to do this?!"

Fast forward 10 years. I'm now a co-op at IBM. Taking classes part time and working mostly full time. This was the early 90's. We were college students on a corporate network in the early days of the internet. What do you think we were doing? Yup, downloading files from sites we shouldn't have been visiting ... hello newsgroups, and hello no filters and firewalls. One of the brighter amongst our group, name of Steve McKay (hopefully some statute of limitations has passed), came running down the hall to show us the .MP2 files he had just burned to a CD. You have to remember during this time CD burners cost $600 and media was $20 each. You had to be careful to not do anything else on your PC while burning a CD, or you made a $20 coaster.

I shrugged, "Hmmm, music files. I've seen these before. How many can you fit on the CD? ... only 20? ... How much to burn the CD? ... I can only play them from this application, I can't play them in a *regular* CD player? ... That's interesting, but who the F#!K is going to do this?! .. I can buy a music CD for $10."

It was only a few short years later that .MP3 was out, costs came down, and digital music was everywhere. Plus you have to remember that Apple had not even entered the market. Yet another lesson that you don't have to be first to be best.

Why did I tell this story?

First to remind myself to not get too old and crusty in my thinking and stop yelling "Get off my lawn!" Secondly, IMO Peter is trying to make people think to the future and much larger implications. He used some analogies that gave some knee-jerk reactions the comments, but if you stop and think, he's on to something.

I'm less and less competing with HP, Dell, Lenovo (the artist formerly known as IBM System x) and more and more with Amazon, Rackspace, Google, etc. My customers want services, not servers. Where I used to be competing against other hardware manufacturers on the best MS Exchange design and the merits of my technical solution, I'm now losing the whole deal to Office 365. Peter is simply pointing out new choices where customers can deploy their applications and services. Does it cost more in the public cloud now? ... Can you build a private cloud better? ... What about hybrid solutions? Questions that have certain answers now, but different answers in the future.

Lots of things to think about, instead of simply saying, "Hey you idiot, servers run the cloud." That's not really the point.

But then again, what the hell do I know? I "poo-poo'd" digital music not once, but twice.

Thursday, March 27, 2014

Support our troops through Operation Supply Drop

What is Operation Supply Drop?
Operation Supply Drop (OSD) is a military gaming 501(c)(3) charity designed to build video game filled care packages for soldiers (America and Her Allies) both deployed forward to combat zones as well as those recovering in military hospitals.
Now you might be asking yourself, "wait a minute .... how the heck does giving video games to the troops help them?" You've got to remember the ages of these men and women who are serving us. They've grown up with gaming and games promote teamwork and camaraderie, as well as help combat the stress and boredom that comes with a deployment.
I've never been much of a fund raiser type of guy, but I know with your support, you can help me raise a simple personal goal of $500.
I'm also impressed with Operation Supply Drop's mission. It started with one man and a simple crazy idea, but they now have a much bigger goal to help the troops after deployment. Read about it directly from the website :
As the United States and her allies end of formal combat operations and continue the full transition of authority to the Afghan government scheduled in 2015, OSD will also transition from wartime to peacetime support of the military. This involves increased on-base activities stateside, further supply drops to peacekeeping and humanitarian missions worldwide, as well as OSD's newest initiative, the "Thank You Deployment" (kicking off mid-2014).
In the future, OSD will continue to act as an intermediary between video game developers/publishers and the United States military, working with soldiers leaving the service, into adult education programs to facilitate them into entry-level gaming developer jobs.
So join me in helping our troops. Thanks!
One Team, One Fight!
Donate directly by clicking here

Monday, February 17, 2014

Whiteboarding Cisco UCS and Cisco UCS Director

No, not water boarding. Although remind me to tell you the story of a customer that had their datacenter running under the jailhouse. Didn't turn out so good when an inmate stuffed the toilet and the sewage pipes burst, which were running directly OVER the racks ..... mmmmmmm, hopefully you're not reading this at lunch :-)

Lately I've been doing #SlideFree presentations, using a whiteboard followed by a demo.

If you're reading this in March, April, or May of 2014 - Click and register for more upcoming events.

I challenge you to do the same. When I worked at IBM, my last gig was in the Executive Briefing Center. I was mentored by a guy named Ron .. Ron Jeremy, Ron Burgandy ? .... heck I can't remember his last name. Anyway, he told me something I'll never forget. He said,
"Scott, these people are going to remember maybe 5% of what you told them. What they will remember is, did they believe you, and did they like you. If they like you and believe you, we're half way there. So stop putting so much damn stuff on your slides!"
Ron was right. Every preliminary presentation I've done with customers that includes too much detail, fails miserably. They end up being more confused than when I started.

Here's links to two recent whiteboarding sessions I recorded. If we haven't had a chance to meet in person, this is how I typically present the topics of Cisco UCS and Cisco UCS Director. Enjoy!

Thursday, February 6, 2014

Read all about it! HP charges - Cisco UCS updates remain FREE

It was an interesting night last night. I'm sitting in my normal poker game downtown. Between hands I always "check the streams" to see what's happening in the world. I look down and don't see pocket AA, but I do see a tweet that HP is going to start charging their server customers for firmware updates! Whoa, seriously, how can I not jump all over this. Being in the x86 server world my whole life, it was just a given that shit breaks and you can always go to the vendor's support site and download a fix. Support contract or not.

It's just the world we live in. When I worked at IBM, a wise man told me, "All hardware eventually fails ... and all software eventually works" ... if you deliver that line with perfect timing, you always get smiles. We all know this, we have to patch, maintain, fix all the time. We simply go to the Microsoft, Dell, IBM, Cisco, etc support sites and download the updates we need.

Well apparently not anymore for HP. This is good for us (Cisco) as most customers evaluate a new purchase when their support contracts run out. Well the one's in my segment sometimes push it a little longer. But now they are forced to renew contracts or look to a new vendor. It's good for me, so hell yeah I sent a tweet pointing this fact out.

I knew I would get a few replies to the tweet of the nature, "But yeah, Cisco has contracts, entitlements, SmartNet, etc that have to be in place to download your software". You're likely coming from the router/switch world of Cisco, in which that statement is absolutely true.

It is NOT TRUE FOR CISCO UCS. You can use any ID that you registered fresh off the street to download Cisco UCS firmware and software updates. 

Now I wouldn't be a proper geek if i didn't prove it. You have to show your work. Which is a problem with the other interesting thing that popped up in the streams last night. Go read Tony's great post on the report that shows how awesome FC is compared to FCoE. Seriously great read, you don't want to have this man pick apart your technical "marketing" whitepaper.

So here's my work.

First, create a gmail account from a computer on a public network.
 Next, create a free account.

Check that I don't have Service Contracts Associated.

Check I have no serial numbers attached to my account

Download any Cisco UCS updates I need

Anything I can't DL is marked with a "lock" icon - Router, switch, etc
Lock icon highlighted

All my freshly downloaded Cisco UCS updates - no registration, free account, wide open.

Friday, October 26, 2012

Quick and Dirty Disjoint Layer 2 Configuration on Cisco UCS

Quick! ... Someone tell me how to get to the LAN Uplinks Manager configuration page on Cisco UCS .. anyone, anyone? Yeah, it's a little hidden and you don't use it that often, but when you need to configure disjoint layer 2 in UCS, you need it.

Here's some quick and dirty notes on configuring Disjoint Layer 2 for Cisco UCS.

  1. Create VLANs on UCS on LAN Tab
  2. Configure Uplink ports on FIs
  3. Optionally put them in Port Channel Group
  4. Click LAN tab, LAN Object (at top), then “LAN Uplinks Manager” hyper link
  5. In LAN Uplinks Manager, click on VLANs tab, then VLAN Manager tab
  6. For Fabric A, Select the Uplink interfaces or Port channels on the left (you can CTRL select for Multiples) .. then select the VLANs on the right, then click the “Add to VLAN button” on the bottom.
  7. Repeat for Fabric B