Beyond Lift & Shift

Controlling Cloud Migration Cost

Cloud is expensive !! Certainly, no company would step into the cloud journey when they hear their bottom line may be impacted. While there is no apple to apple comparison between on-premises infrastructure and cloud, no one can decline that the cloud as-a-service model is much more reliable, available, scalable, and agile. The flexibility of cloud services is unmatchable, and cloud portfolio, especially in the SaaS domain, is sort of everything-as-a-service. For most customers, there is 99% fit in the cloud; they can run all their apps and workloads in the cloud. Then what? Just that the higher cost (myth?) is keeping many customers away from starting their journey into the cloud.

Recently someone presented CloudPhysics assessment report (HPE acquired CloudPhysics in Feb-2021) and it looked like this:

CloudPhysics Assessment: Cloud cost comparison. Source: https://www.cloudphysics.com/

What’s highlighted in this assessment? This assessment captured the current utilization and configuration metrics from on-premise hosts and estimating the cost if these existing hosts had to run into the cloud (Microsoft Azure and AWS). A comparison of cost running the same workload on-premise vs cloud. But just like other general assumptions about cloud migration this assessment is missing a major consideration about cloud migration, that is the infrastructure and application modernization. Cloud migration offers tremendous opportunities to modernize application and infrastructure and avail the new generation of suites available in the cloud offerings. While modernization can offer many benefits of the latest features and tools available in modern apps (as-a-Service), this transformation can improve the application performance and reduce the cost at the same time. Cloud native applications are surely optimized and on the other hand lift-and-shift migration of on-premises workload can inherit existing issues and run unoptimized in the cloud, surely to grab unnecessary resources while running in the cloud. Customers who adopted the virtualization model will easily be able to relate this to virtual machine cleanup after physical-to-virtual (P2V) migration in order to run clean and efficiently on a virtualized platform. This goes the same for cloud migration whereas the benefits are realized in direct relation to the dollar cost.

So, what’s the best way to modernize the cloud workload? A simple answer is that a well architected cloud migration strategy is the only way you can save money while making the journey to the cloud. A much more comfortable ride without much hassle and bumps in the road. At a very high level cloud migration roadmap will include – 1. Assess, 2. Prioritize, 3. Plan and Plan, & 4. Execute. And in order to map out the best path to modernize the infrastructure, Cloud Architects will evaluate essentials of the cloud migration i.e. 6 R’s of Cloud Migration.

6 R’s of Cloud Migration

Gartner first came up with the 5 R’s of cloud migration in 2010, later revised by other cloud providers and added one more to the list –

  1. Rehost (Lift & Shift)
  2. Replatform (Lift, Tinker & Shift)
  3. Repurchase (Replace)
  4. Refactor (Rearchitect/decoupling of apps)
  5. Retain
  6. Retire

Rehost (Lift & Shift) –

Simply described, Rehosting is redeployment of the workload (typically a VM or host) to a cloud IaaS (Infrastructure-as-a-Service) hardware environment. By far this is the most assumed method for most who are new to the cloud and haven’t started their journey yet.

Example: Migrating VMware or Hyper-V VMs to cloud hosted virtual machines.

Pros – Fastest, Reduced risk, Use of free migration tools from Cloud Service Providers (CSP), No configuration or code modifications required.

Cons – Doesn’t uses exorbitant cloud native features, inherit native on-premises problems.

Replatform (Lift, Tinker & Shift) –

Replatforming requires tiny bit modification of platform to adjust according to the cloud requirements. This is essential for application or databases to work natively in the cloud. This is typical to most IaaS or PaaS migration scenarios.

Example: Migrating on-premise relational database system to a cloud native Relational Database Service (RDS)

Pros – Minimum modification needed to code, OS independent, opportunity to use modern platform, reduction in shared responsibility model with CSP.

Repurchase –

As the name suggest this requires abandoning of existing software platform and buying a more suitable suite in the cloud. SaaS solutions provided by CSPs have plenty of choices to select a replacement software for customer need with scale for SMB to enterprise customers.

Pros – Quick provisioning and migration to new software with least efforts and abandoning of shortcoming of existing software.

Cons – New dependencies, new learning curve, loss of perpetual licensing cost.

Refactor (Rearchitect/decoupling of apps) –

This is a trend in most enterprise where applications are being refactored for improved performance, scalability and availability. Cloud migration offers golden opportunity for such initiatives where software applications can be rearchitected to be modernized.

Example: Migrating from monolithic Virtual Machines to serverless compute architecture.

Pros – Significant transformation of applications, true migration to cloud with most benefits of availability, scalability and availability.

Cons – Time consuming, expensive and can be interruptive to business critical applications.

Retain –

If the assessment shows drawbacks of moving into cloud or the cost is not justified or the workload can not be moved into cloud then it’s best to retain the workload on-premise. This is where cloud migration assessment will play a significant role.

Example: If OS or application is not supported in cloud.

Retire –

Applications that can be turned off or not in use can be retired so it doesn’t incur cost of migration.

It is obvious that cloud migration is not easy and it shouldn’t be because the technological shift that comes with it makes every customer ready for the future and the benefits of cloud are enormous. While no one can predict whether the cost of cloud services will be lower or higher in the future, one thing is sure that cloud is the future. Many organizations can justify today by not moving into the cloud but in future they won’t be left with any options. Supply chain issues, global pandemic, semiconductor shortage and expensive personalized software development will force the infrastructures to move to something reliable, available and sustainable i.e. cloud.

Microsoft Azure Fundamentals (AZ-900) Certification – Unofficial Study Guide

Anyone starting fresh journey in to the cloud learning should consider Microsoft’s Azure Fundamental certification. AZ-900 is the single exam needed to get the Azure Fundamental certificate. Enormous resources are available on the web and during the pandemic many more came to exist. Although it’s not wrong to pick any one of the those provide it should match your learning style and should give you guidance and confidence to attempt and pass the certification test.

First get your free voucher from Microsoft (worth $99 towards the exam) by attending free two days class. It runs around 2 – 2.5 hrs each day. Find the schedule and registration link at Microsoft Virtual Learning Days

Here is my take on study material for Azure Fundamental (AZ-900) cert exam. Preparation should be done in three ways; first, go through a detailed and easy to understand course (please don’t put time as the constraint because fast paced course might be just waste of time if it doesn’t match your catching pace). Secondly, revise it with something fast paced, this will bring back all the concepts that you learnt in longer course. Thirdly, learn the exam technique on how to understand the questions and attempt those with confidence and when to revise.

I suggest these three YouTube videos to 1. Detailed course 2. Exam cram, and 3. Exam preparation.

  1. www.Youtube.com/watch?v=lm0sdkylqvs (8 Hrs of comfortably paced full course)
  2. www.youtube.com/watch?v=gH3pwWO0Q9Y (Exam crams in 2 Hrs)
  3. www.Youtube.com/watch?v=53LO_rJz6E (55 Questions explained with details and solving skills)

Moreover, I have uploaded an Unofficial Study Guide to prepare for the cert.

Azure Fundamentals – AZ-900 Unofficial Study Guide – DOWNLOAD

Good Luck all !!

HPE BladeServer BL460c FCoE on RedHat Linux

A quick procedural reference to configure HPE BL460c bladeserver with FlexFabric Virtual Connect. RedHat Linux doesn’t present the Fiber Channel FlexNIC as Fiber Channel ports. These ports (LOM1b and LOM2b) are presented as NIC ports with default MacAddresses. In order to configure the Fiber Channel ports, FCoE configuration needs to be enabled.

Two configuration changes are required from default configuration –

1. /etc/sysconfig/network-scripts/ifcfg-enoXX
2. /etc/fcoe/enoXX

In order to configure the correct network adapters need to figure out which adapters are configured for Ethernet and which are for Fiber Channel.

Look through the Virtual Connect to determine the configured Ethernet adapters –

From Virtual Connect Command Line Interface (CLI via SSH to Virtual Connect Manager) –
VC –> show profile C7000_Bay1
(Replace the C7000_Bay1 with the server profile)

Sample output:
Ethernet Network Connections
===================================================================
Port Network Status PXE/IP Boot MAC Address Allocated Name Order Speed (min-max)
===================================================================
1 Multiple OK UseBIOS/Auto 00-17-A4-73-00-30 3Gb-10Gb Network
————————————————————————-
2 Multiple OK UseBIOS/Auto 00-17-A4-73-00-32 3Gb-10Gb Network
————————————————————————-
3 Multiple OK UseBIOS/Auto 00-17-A4-73-00-34 3Gb-10Gb Network
————————————————————————-
4 Multiple OK UseBIOS/Auto 00-17-A4-73-00-36 3Gb-10Gb Network
————————————————————————-

FCoE Connections
===================================================================
Port Connected Name Type Status Allocated WWPN MAC Address To Bay Speed (min-max)
===================================================================
1 1 Fabric SAN OK 4Gb-8Gb 50:06:0B:00: 00-17-A4-75–A 00:C2:65:04 00-0E
——————————————————————————
2 2 Fabric SAN OK 4Gb-8Gb 50:06:0B:00: 00-17-A4-75–B 00:C2:65:06 00-0F
——————————————————————————

Not that FCoE connections have WWN as well as the MAC Addresses for each port. Note down the MAC Addresses and search for them in Linux –

On the RedHat server please ensure that FCoE util is loaded. Verify the install with following command –
# yum list | grep -i fcoe
fcoe-utils.x86_64 1.0.32-1.el7 @anaconda/7.5

If not install it using following command –
# yum install fcoe-utils.x86_64

Now, configure the FCoE on RedHat server

Sample RedHat Linux commands –
# ip addr
4: eno31: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
link/ether 00:17:a4:75:00:0e brd ff:ff:ff:ff:ff:ff
5: eno32: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
link/ether 00:17:a4:75:00:0f brd ff:ff:ff:ff:ff:ff

There will be many adapters but in this case eno31 and eno32 are matching the MAC Addresses with the one we noted down in Virutal Connect output.

Once identified, configure these two adapters.

1. Configure the two NIC adapters in /etc/sysconfig/network-scripts/

/etc/sysconfig/network-scripts/ifcfg-eno31
=====================================
# TYPE=Ethernet
# PROXY_METHOD=none
# BROWSER_ONLY=no
# DEFROUTE=yes
# IPV4_FAILURE_FATAL=no
# IPV6INIT=yes
# IPV6_AUTOCONF=yes
# IPV6_DEFROUTE=yes
# IPV6_FAILURE_FATAL=no
# IPV6_ADDR_GEN_MODE=stable-privacy
# NAME=eno31
BOOTPROTO=none
UUID=1ff88eae-3300-41d7-91b8-490110ca8df2
DEVICE=eno31
ONBOOT=yes
NM_CONTROLLED=no
===========================================
Most importantly, ONBOOT=yes will ensure that adapter is set to FCoE adapters at each boot.

2. Configure both NICs in /etc/fcoe

/etc/fcoe/eno31
====================================
## Type: yes/no
## Default: no
# Enable/Disable FCoE service at the Ethernet port
# Normally set to “yes”
FCOE_ENABLE=”yes”

## Type: yes/no
## Default: no
# Indicate if DCB service is required at the Ethernet port
# Normally set to “yes”
DCB_REQUIRED=”no”

## Type: yes/no
## Default: no
# Indicate if VLAN discovery should be handled by fcoemon
# Normally set to “yes”
AUTO_VLAN=”yes”

## Type: fabric/vn2vn
## Default: fabric
# Indicate the mode of the FCoE operation, either fabric or vn2vn
# Normally set to “fabric”
MODE=”fabric”

## Type: yes/no
## Default: no
# Indicate whether to run a FIP responder for VLAN discovery in vn2vn mode
#FIP_RESP=”yes”
=========================================
Note that MODE=”fabric” and DCB_REQUIRED=”no” are the most important settings and the file name should be matching the one we set in /etc/sysconfig/network-scripts/ifcfg-eno31 and eno32 files.

That’s it, the FCoE adapters should be enabled and they should be logged in to SAN Switches.

Why HPE invested in Synergy, “The Composable Infrastructure”?

HPE is positioning Synergy “The Composable Infrastructure” to slowly phaseout c7000 “BladeSystem” that was announced in 2006 and remained unmatchable since then. Synergy introduces industry’s first Composable Infrastructure” that echos HPE’s modern IT slogan “Hybrid IT Transformation” and replaces traditional c7000. Recently, many of the c7000 customers are asking questions around HPE’s new product line announcement – Synergy BladeSystem – “Why do we need Synergy”?

Synergy

Answer to the question lies in the question that HPE asked themselves before investing huge efforts in research and development of Synergy. Why HPE needed to come up with new product line when c7000 was keeping up with the development of new SAN, Network, SAS, Virtual Connect, Flex-Fabric technologies? Since its launch, c7000 BladeSystems adapted to most of the new technologies with a minor firmware code upgrade and delivered unprecedented performance and uptime to its customers. Still, HPE decided to invest in new product to deliver an unmatchable experience to the customers. Undoubtedly, HPE’s efforts paid off with Synergy that simplifies the environment for a quicker deployment Synergy poses several benefits to IT –

  • Industry demand is to move forward with Composable and Synergy is fully Composable.
  • Quicker to deploy with least components configuration.
  • It is Simple to deploy and manage.
  • Costs much less to upgrade than any other product in market.
  • Interconnected upgrade options are integrated as local resources with the introduction of re-timers in satellite fabric modules.

While developing the Synergy platform, HPE kept in mind that be it a traditional customer with simple applications (web hosting, fileshares, simple virtualized environment, database and home grown or legacy applications) or a customer with new breed of complex applications and services (DevOps, containers, managed services, hosting), HPE ensures that the “Composable Infrastructure” will be able to meet the requirements and demands of modern IT managers and administrators.

Composable infrastructure offers –
– Composable Frame
– Composable Management
– Composable Computer
– Composable Storage
– Composable Fabric

Composable Infrastructure benefits from its underlying components –
– Fluid Resource Pool
– Software-defined Intelligence
– Unified Programming Interface

HPE Synergy significantly reduces number of components on a purchase order and that is made possible with the introduction of new interconnect modules and compute nodes. With new generation compute modules, Synergy offers infrastructure to any size environment and makes it future proof with a promise to accommodate future technologies. While HPE is yet to release more information on its future product line it is a very good guess to assume that HPE’s latest undertaken projects will be first take option on Synergy, such as – Photonics, The Machine, Simplivity, More variants Storage and Fabric interconnect modules.

(More information can be found at http://www.hpe.com/info/synergy )
Posted in HPE

VMware Update Manager Installation with SQL Express (.Net framework3.5)

Installing VMware Update Manager (version 6.5) from the same installation ISO as vCenter there are two options; install on existing database (typical when VUM is installed on a vCenter server) or to install with embedded SQL Express. The Embedded SQL Express option is only available when Auto-run is selected from the installation ISO.

Graphic – Autoplay from ISO

2017-03-25 20.38.47

Graphic – Autorun from root folder

2017-03-25 20.40.29

If you run the VMware Update Manager installer (.exe or .msi) from the “Update Manager” directory, option for Embedded SQL Express is not there and a message is displayed like this –

Graphic – .EXE from the folder

2017-03-25 21.22.14

Graphic – VMware embedded db unavailable from autorun2017-03-25 20.52.41

Remeber, VUM uses 64-bit DSN that needs to be set if selected to use existing database. Once the database option is selected and passed the screen, VUM installs .NET framework 3.5 as pre-requisite. There are two problems encountered while installing the .NET framework 3.5. First, VUM installer tries to install the .NET framework 3.5 and the installation fails. If this happens you may try to enable the .NET framework 3.5 from the control panel –

Graphic – enable .NET framwork 3.5 from server manager console2017-03-25 21.29.38

or, another option is to use CLI command (with administrator privileges). DISM expample –

DISM /Online /Enable-Feature /FeatureName:NetFx3 /All

or (if you have source mounted as D:\ )

DISM /Online /Enable-Feature /FeatureName:NetFx3 /All /Source:D:\sources\sxs /LimitAccess

Graphic – install-net35-disminstall-net35-dism

Troubleshooting: The .NET framework installation may fail if installation media is missing or Windows Server Update Services (WSUS) can’t be reached.

Feature installation gets error

Graphic – .NET framework installation error using server managerdotnet_feature_install_faial

DISM installation stuck at 66.5% and subsequently fails with error code 0x800F0906

Graphic – .NET framework installation error using DISMinstall-net35-dism-error

This error is because Windows host isn’t allowed to download online updates (this is separate from Windows Update). To solve this change the group policy settings to allow optional component installation to contact Windows update.

Run the gpedit.msc and navigate to

Computer Configuration > Administrative Templates > System > 

and open the “Specify settings for optional components and component repair” by double-click and then check the box “enable” and “Contact Windows update …”

Graphic – gpedit navigation2017-03-24 16.28.03

Graphic – enable windows update2017-03-25 21.14.46

All is well from here. Run the DISM CLI and .NET framework 3.5 will be installed online without requiring Windows source DVD.

First blog post

I am excited to start this blog to write about various infrastructure configurations that I come across. I am an IT consultant and deal with various configuration and issues, I will post most interesting facts that I encounter on typical configurations involving HPE, EMC, VMware, Veeam, Microsoft and Brocade.