HPE Discover – Thoughts

hpe_pri_grn_pos_rgbIt’s Sunday morning and I’m sitting in London Heathrow Terminal 3, on my way to HPE Discover.  If you follow my blog, you will have noticed that my focus has been on Microsoft Azure, with this in mind, it I wasn’t entirely sure if HPE would see the value in bringing me to HPE Discover this year.  However the legend that is @CalvinZito believes I can bring something to the table

It seems strange if not alien for me to be thinking about hardware, I have been decoupled from the murky depths of DAC cables, storage performance/capacity, servers and networking items such as virtual connect for over twelve months.  Yes you still have architectural considerations in Microsoft Azure, however these are higher up the stack and are geared towards limits and quotas rather than hardware boundaries.

Meg Whitman has proved that she has the vision and belief to shape and change HPE.  With the split into Enterprise and Ink in 2015 and the recent announcement that HP Enterprise Services will be acquired by CSC (Computer Sciences Corp).  This leaves HPE as a global hardware and software manufacturer.  I’m interested to know how HPE is adapting to the future of ‘hybrid cloud’ and how they are going to show value in a ‘commodity hardware’ market.

When you have a company that turns over $53 Billion dollars and has 252,000 employees, it’s a gigantic task to point the ship in a new direction.

Let’s see what HPE announce at Discover and what their thoughts are in an ever changing market place.

HP ConvergedSystem 200-HC StoreVirtual System – Questions Answered

Background

HP released two offerings of the HP ConvergedSystem 200-HC StoreVirtual System last year.  Essentially they have taken ESXi, HP StoreVirtual VSA, OneView for vCenter and automated the setup process using OneView Instant On.

HP Converged System 200-HC Diagrams v0.1

Two models are available which are:

  • HP CS 240-HC StoreVirtual System, this has 4 nodes each with:
    • 2 x Intel E5-2640v2 2.2GHz 8 Core Processor
    • 128GB RAM
    • 2GB Flash Backed HP Smart Array P430 Controller
    • 2 x 10GbE Network Connectivity
    • 1 x iLO4 Management
    • 6 x SAS 1.2TB 10K SFF Hard Drives
    • Around 11TB of usable capacity
  • HP CS 242-HC StoreVirtual System, this has 4 nodes each with:
    • 2 x Intel E5-2648v2 2.2GHz 10 Core Processor
    • 256GB RAM
    • 2GB Flash Backed HP Smart Array P430 Controller
    • 2 x 10GbE Network Connectivity
    • 1 x iLO4 Management
    • 4 x SAS 1.2TB 10K SFF Hard Drives
    • 2 x 400GB Mainstream Endurance SSD
    • Around 7.5TB of usable capacity

These are marketed with the ability to provision virtual machines within 30 minutes.

What Does Provision Virtual Machines Within 30 Minutes Really Mean?

To answer this question you need to understand what HP have saved you from doing, which is:

  • Installing ESXi across 4 x Hosts
  • Installing vCenter to a basic configuration
  • Installing HP StoreVitrual VSAE to a basic configuraiton across 4 x Hosts
  • Pre-installed Management VM running Windows Server 2012 Standard that has OneView for vCenter and CMC for StoreVirtual Management

So after completing the initial setup, you do have the ability to upload an ISO and start deploying an OS image.

What About The Stuff Which Marketing Don’t Mention? AKA Questions Answered?

Database

  • SQL Express is used as the database (local instance on Management VM).  I have real concerns around the database if logging levels are increased to troubleshoot issues and/or the customer doesn’t perform an kind of database maintenance
    • I’m waiting on confirmation from HP as whether you can migrate the SQL database instance to a full blown version

Host Profiles

  • Grey area, these can be used.  However HP rather you stay with the base configuration of the nodes (much like the networking see below).

Licences

  • The solution is only supported using Enterprise or Enterprise Plus VMware licenses, with the preference being HP OEM.
  • Windows Server 2012 Standard is supplied as the Management VM.  Initially, this runs from a local partition and is then Storage vMotioned onto the HP Converged Cluster.  Windows licensing dictates that when a OS is moved across hosts using Standard Edition you cannot move the OS back for 90 days or you need to license each node for the potential number of VM’s that could be ran.
    • HP have confirmed that you receive 2 x Windows Server 2012 Standard licenses and DRS Groups Manager rules are configured to only allow the Management VM to migrate between these two ESXi Hosts.

Management Server

  • You are able to upgrade the Management Server VM in terms of RAM, CPU and Disk Space and be supported.
  • You cannot add additional components to the Management Server VM and be supported e.g. VUM, vCenter SysLog Service
    • I’m waiting on confirmation from HP around what is and isn’t supported, I would air on caution and not install anything extra

Networking

  • The 1GbE connections are not used apart from the initial configuration of the Management Server.  My understanding is that these are not supported for any other use.
  • HP prefer you to stay with the standard network configuration, this causes me concern.  10GbE Network providing Management, iSCSI, Virtual Machine and vMotion traffic.  How do you control vMotion bandwidth usage on a Standard vSwitch? You can’t a Distributed vSwitch is a much better option, but if you need to reconfigure a node, you will need to perform a vSS to vDS migration

Updates

  • You can upgrade individual components separately, however you must stay within the HP Storage SPOCK for the Converged System 200-HC StoreVirtual (Note a HP Passport login is required)

Versions

At the time of this post, the latest supported versions are as follows:

  • vSphere 5.5 U2 , no vSphere 6
  • vCenter 5.5 U2
  • HP StoreVirtual VSA 11.5 or 12.0
  • HP OneView for vCenter Storage/Server Modules 7.4.2 or 7.4.4
  • HP OneView Instant On 1.0 or 1.0.1
  • PowerCLI 5.8 R1

Final Thoughts

HP have put together a slick product which automates the initial installation of ESXi and gives you a basic configuration of vCenter.  What it doesn’t give you is  design to say that your workloads are going to be suitable on the environment and or a solution that meets a client requirements.

HP 3PAR Streaming Remote Copy Replication

The replication in 3PAR Arrays has always been mediocre.  In the older versions of 3PAR Inform OS if you choose ‘sync replication’ for a single remote copy group, you could not use ‘a sync’ for any other remote copy groups.

This feature was addressed in a newer version of 3PAR Inform OS, however your lowest RPO using ‘a sync’ was bottlenecked at 15 minutes regardless of available bandwidth.

With the release of the HP 3PAR 20000 Series, comes a new feature which is streaming ‘a sync’ replication.

What Is Streaming ‘A Sync’ Replication

Essentially, if you have the bandwidth and cache available the source 3PAR will stream replication across to the target 3PAR reducing your RPO below 15 minutes.  I like to think of it as a best endeavours.

Replication Modes

When designing a replication infrastructure it’s important to know the transport method as well as the thresholds in terms of bandwidth and latency between source and target arrays.  This ensures that you are not only within a supported SLA, but also to ensure that write performance of the source array is not effected.

The table below shows supported thresholds.

Replication Modes

Architecture

The source array uses a local cache to maintain host write transactions in memory.  A concept known as ‘delta sets’ are used.

Source 3PAR Array

  • I/Os are transferred from primary array to secondary array as part of a delta sets
  • I/Os on the primary array that belong to a particular remote copy group are grouped together into delta sets
    • A delta set is made up of sub-set of I/Os, where sub-set represents I/Os owned by a remote copy group on each given node

Target 3PAR Array

  • A delta set is applied on the secondary RC volume group only after:
    • The entire delta set has been received in the secondary array cache
    • And the previous sets that this delta set depends upon have completed.
  • A secondary RC volume group is always in a crash consistent state, before or after the application of a delta set. It is not crash consistent during the application of a delta set.
    • If the delta set fails to apply on the secondary volume then the group stops and a fail back to the last coordinated snapshot is required

Remote Copy Architectuer

What About Write Bursts?

A write burst is when the array receives a significant number of writes which could last for a few minutes.  If the inter site link between source and target array is sufficient this has no impact.

It is when the inter site link cannot cope or the write cache gets filled then the source 3PAR will choose a random remote copy group to stop and a snapshot is taken.

Note: You have no control over which remote copy group

Once stopped these groups will start again at the next sync period.

Final Thoughts

This is a great feature set being added to the 3PAR 20,000 Series.  I’m sure when the next .1 release update is received you will be able to select which remote copy groups you would want to stop either due to a write burst or cache overflow.

With most 3PAR updates, I expect the streaming ‘a sync’ replication to find its way into the 7×00 series within a short period of time.

HP StoreVirtual Multi-Path Extension Module (MEM) for vSphere

With the release of LeftHand OS 12.0, HP have introduced a Multi-Path Extension Module to replace the previously recommended path selection policy being ‘Round Robin’ and Storage Array Type Plugin VMW_SATP_DEFAULT_AA

By default Round Robin would send 1,000 I/O down each path.

Note: That this can be altered as per my previous article entitled ‘How To Change Default IOP Limit‘ but should normally only be done under the guidance of the manufacturer.

The issue with this is how a volume is created within a StoreVirtual node.  Let’s imagine we have two cluster nodes, A and B. When the first volume is created it is ‘owned by a master node’ in this case ‘node A’.  When the second volume is created it is owned by ‘node B and so on, as shown below.

  • Volume 1 – Node A
  • Volume 2 – Node B
  • Volume 3 – Node A
  • Volume 4 – Node B

Using the recommended bonding method of ‘Adaptive Load Balancing’ SCSI read and write commands are issued on all NIC’s in the bond which can result in data being accessed from a non-authoritative node, which means a trip across the network to the authoritative node.  I think a picture is in order!

VMFocus HP StoreVirtual DiagramThis was rather inefficient and meant that random reads (which are served from disk) could be accessed from a non-authoritative node.  This is where the StoreVirtual Multi-Path Extension Module (MEM) steps in.  It has knowledge on where the data resides and ensures that all:

  • Read I/O’s are always serviced by the storage node that holds the authoritative data.
  • Write I/O’s are always serviced by the storage node that receives a mirror copy of the data ensuring data integrity.  Remaining copies are then forwarded to the remaining storage nodes.

This results in the following data flow architecture.

VMFocus HP StoreVirtual MEM Diagram

The installation guide for HP StoreVirtual Multipathing Deployment Guide is fairly straight forward.  The only issue I ran into was ‘could not find trusted signer’ when trying to install the vib.

VIB Install

This was resolved by adding –no-sig-check to the end of the command esxcli software vib install -v /HPMEM.vib –no-sig-check

Once installed, you will need to change the path selection policy for your datastores to HP_PSP_LH

HP_PSP_LH

 

 

How To: Rehost HP StoreVirtual Licenses

I’m not sure exactly when, but HP changed the licensing portal from ‘Poetic’ to a new portal named ‘My HP Licensing Portal’.  All of the information looked exactly the same, however you could not rehost HP StoreVirtual Licenses.

The purpose of this blog post is to assist anyone who was in the same situation as me, scratching their head trying to figure it out!

Problem

You have an existing HP StoreVirtual license which ties the feature set to the first NIC MAC address on your StoreVirtual VSA.  You have changed, upgraded or redeployed your VSA and you need to rehost the license onto the new MAC address.

Solution

Browse to myhplicensing.hp.com and login for your email address and password that you use for your portal account.

Note: If you are not sure what email address is tied to your account login to HP Licensing for Software and select Administration > My Profile which will show your email address.

Once logged into My HP Licensing select Rehost Licenses

StoreVirtual License 01

Next select Rehost Licenses and click on the MAC Address you want to update

StoreVirtual License 02

Select the tick box to confirm this is the license that you want rehosting and then click Rehost.

StoreVirtual License 03

Select ‘Enter New Locking ID’ and enter the MAC Address of the first network adapter on your StoreVirtual.  Then click Next

StoreVirtual License 04

This bit takes a while but once done you will receive the license file which can be saved or emailed

StoreVirtual License 05