Monday, 15 September 2014

Our SBS (Small Business Solution) Options with Standalone and Cluster Hardware Considerations

We’ve received a number of questions about the “How” we present our SBS to prospective and existing clients.

Our primary focus is on what we have provided with Small Business Server starting with SBS 2003 Standard.

  • Active Directory permissions based security
  • Remote Web Access (RWA/RWW) Portal
  • Remote Desktop access via RD Gateway (since SBS 2008 Standard)
  • RemoteApp access via RD Gateway for LoBs (since SBS 2008 Standard)
  • E-mail services access via Outlook, Outlook Anywhere, Exchange ActiveSync, and Outlook Web Access
  • Remote Folders and Files access
  • SharePoint based document management system
  • SQL backend for LoB, SharePoint, and other needs

We focus on the services the prospect would require while our existing clients are already used to them.

Once we have an understanding of the prospect’s needs, since we already know our client’s business really well, we move forward with a proposal that would be geared towards their business size and sensitivity to downtime.

On the services front where we are installing into a standalone host we would have two options:

  1. Base
    1. Requires two Windows Server OS Licenses
    2. DC, Exchange, RDS, and LoB (WSUS and LoBs)
  2. Premium Add-On
    1. Requires one Windows Server OS License
    2. SQL and SharePoint

Obviously the server and CALs would also be needed for the various components that will be installed into the guest OS.

If we are setting up a cluster then one needs to consider the number of VMs running on one or more of the nodes in the event of a node failure.

On the hardware side we would have a number of options:

  1. Entry-Level Single
    1. E3-1270v3, 32GB ECC, Hardware RAID, 8x 2.5” 10K SAS
  2. Mid-Level Single
    1. Single Socket 1U R1208JP4OC, E5-2600 series, 128GB ECC, Hardware RAID, 8x 2.5” 10K SAS
  3. High-Level Single
    1. Dual Socket 2U R2208GZ4GC, E5-2600 pair, 128GB-256GB ECC, Hardware RAID, 8x 2.5” 10K SAS or 16x 2.5” 10K SAS
  4. Entry-Level Asymmetric Cluster
    1. Pair of 1U R1208JP4OC or 2U R2208GZ4GC and an Intel JBOD2224S2DP
  5. Mid-Level Cluster
    1. Four 2U R2208GZ4GC and an Intel JBOD2224S2DP
      • Two Scale-Out File Server cluster nodes
      • Two Hyper-V cluster nodes
  6. High-End Cluster
    1. Six 2U R2208GZ4GC and three Intel JBOD2224S2DP units
      • Three Scale-Out File server cluster nodes
      • Three Intel JBODS with Two-Way or Three-Way Mirror and Enclosure Resilience
      • Three Hyper-V server cluster nodes

Within the above hardware configurations we would have a lot of flexibility that allows us to customize to the specific needs of the prospective client or our clients.

We work with a number of different firms that are prime candidates for at least an asymmetric cluster setup to minimize the possibility of downtime. The cost associated with these entry-level clusters versus a single larger server for the host platform makes them very attractive.

The basic VM configuration would involve fixed VHDX files unless the files are installed on dedicated partitions/LUNs. Note that we would use a shared set of partitions/LUNs if there are around 10 or more VMs as things get to be a bit of a bear to manage otherwise.

Our base VM configurations would be as follows:

  • DC: 4GB, 95GB OS VHDX, and 1TB Data VHDX
  • Exchange: 8GB, 95GB OS VHDX, and 250GB + 20GB/Mailbox Data VHDX
  • RDS: 4GB+, 95GB OS VHDX, and 100GB + 20GB/User Profile Disk
  • LoB: 8GB, 95GB OS VHDX, and 1TB Data VHDX Minimum
  • SQL: 16GB, 95GB OS VHDX, and 250GB+ Data VHDX
  • SharePoint: 16GB, 95GB OS VHDX, 200GB Data VHDX

We have a set of PowerShell steps and scripts that we use to configure these environments. PowerShell helps to greatly reduce the amount of time required to set things up. It also gives us consistency across all of our client deployments which is vital to troubleshooting if the need arises.

Shameless Plug: We’ve spent some time on the above in our SMBKitchen ASP Author Chats. If you are looking for more information the Author Chat is one of the best ways to do so.

Philip Elder
Microsoft Cluster MVP
MPECS Inc.
Co-Author: SBS 2008 Blueprint Book
Chef de partie in the SMBKitchen ASP Project
Find out more at
Third Tier: Enterprise Solutions for Small Business

Friday, 12 September 2014

MANDATORY: Intel JBOD2224S2DP Firmware Update for Same Enclosure ID Storage Spaces Problem

Intel has released a new firmware for the Intel JBOD2224S2DP storage enclosure that deals with the enclosure delivering the same Enclosure ID to Windows Server 2012 R2 Storage Spaces.

Why is this firmware mandatory?

Because up until now when two or more Intel JBOD2224S2DP units were connected to Scale-Out File Server nodes and one ran the Get-StorageEnclosure PowerShell command one would get the same ID back for every one.

The firmware problem killed Storage Spaces enclosure resilience. What is that you ask?

In Storage Spaces, with a Two-Way or Three-Way mirror one can have three enclosures set up to allow one to drop out completely and things keep going.

If one has configured five enclosures with a Three-Way Mirror then the Storage Spaces setup can tolerate two enclosures dropping out.

If the Intel JBOD unit is already in a production setting with plans to add more enclosures at a later date then it is important to note that this firmware update would be required prior to adding the new units.

For new setups, this firmware update should be a part of the preparation steps for the JBOD prior to implementation or baseline testing.

Philip Elder
Microsoft Cluster MVP
MPECS Inc.
Co-Author: SBS 2008 Blueprint Book
Chef de partie in the SMBKitchen ASP Project
Find out more at
Third Tier: Enterprise Solutions for Small Business

Tuesday, 9 September 2014

Hyper-V VM: Set Up PDCe NTP Time Server plus other DC's time service

When we are running a greenfield deployment, a migration, or a domain reconfiguration one of the key configurations to get right is the Domain Time setup.

We've had situations where anomalies in the time setup have brought down clusters (node's time was out of whack).

In Hyper-V Management disable time integration for the DC(s).

On the primary DC (PDCe) VM that will be the time authority for the domain we run the following steps (please copy and paste into NotePad to clean up):

w32tm /config /syncfromflags:manual "/manualpeerlist:0.ca.pool.ntp.org,0x1 1.ca.pool.ntp.org,0x1 2.ca.pool.ntp.org,0x1 3.ca.pool.ntp.org,0x1" /reliable:yes
w32tm /config /update
net stop w32time && net start w32time
w32tm /resync /force
w32tm /query /source

If you hit ENTER after the last line above and copy from the line underneath to the beginning of the first line and paste into an elevated CMD all of the steps will run automagically. Note that the NTP.ORG servers are Canadian localized. There are servers for most regions in the world.



On every other DC in the domain we run the following steps in an elevated CMD:
w32tm /config /syncfromflags:domhier /reliable:no /update
net stop w32time && net start w32time
w32tm /resync /force
w32tm /query /source

Note that the above step(s) may need to be run a couple of times to get the settings to seat.

Once the above setup has been completed the domain's time should remain on time. If there are any issues with time, that is it starts to slip or jumps ahead, check in Hyper-V Management to see if the IS setting for Time Sync has not somehow become enabled again.

If it has run the above steps again to reseat the settings.

NOTE: For SQL or high load VMs that tend to skew time, we've seen situations where the VM's time skews before our eyes, then set up a BATCH file with the following in it:
w32tm /resync /force

Set it to run in the VM's Task Scheduler every five minutes. If time still gets too far out of whack then decrease the time increment until the VM keeps its time as much as possible.

Philip Elder
Microsoft Cluster MVP
MPECS Inc.
Co-Author: SBS 2008 Blueprint Book
Chef de partie in the SMBKitchen ASP Project
Find out more at
Third Tier: Enterprise Solutions for Small Business

Friday, 22 August 2014

Blogging Lull and Business News

Hay all, it's been either really busy or quite quiet around here. During the lulls I've been taking the time to spend at home with the family. Thus the lack of blog posts for the last while.

We are now in the process of flipping our entire setup into a new 2012 R2 DC/HV Cluster setup. Our existing 2008 R2 RDS server dropped a drive, so rather than risk losing data (did not want to waste time on an RMA and we do not have anymore of them in stock) on the old RDS everything is being transferred over to the new system.

This includes Windows Live Writer (all posts have been RemoteApp WLW posted for a number of years now). We've not quite finished setting the RDS server up yet which includes installing WLW into 2012 R2 RDS and delivering via RemoteApp again.

Our SBS 2011 Standard server will also be deprecated in favour of 2012 R2 DCs with failover DHCP configured and Exchange 2013 CU5 providing our mailbox experience. One of those DCs will reside on the cluster while the other is a physical server (we _always_ set up a physical DC in cluster environments).

In the midst of all of that our primary APC 6KV 208/220V UPS and step-down transformer went full stop on us. Unfortunately the 2 year warranty was over by a good 10 months so no go for replacement. We negotiated a decent deal to replace both via the TRADE-UPS program APC offers plus toss in an extended warranty for a total of 5 years coverage.

We looked to Eaton as a possible replacement but unfortunately economics took precedence in our decision to accept APC's deal (we're in a costly PoC for a new business endeavour).

What's on the agenda for this coming fall?
  • Our Proof of Concept (PoC) testing will start up again
    • Scale-Out File Server and Storage Spaces SSD, HDD, SSD Tiers, and resilience testing
    • Hyper-V VM storage thrashing (VHDX via SMB3 share on SOFS)
    • Hyper-V Network Virtualization for Multi-Tenancy
    • Dell SonicWALL NSA2600 HA Testing (we have a pair on the way)
We are excited about the new business we are in the process of  building. We plan on jumping into the Cloud Services mix with a product offering no one out there can come close to offering at this time!

We are also keeping our clients very happy with their very stable I.T. environments. :)

My Third Tier involvement, both with the Help Desk and the ASP Project, has been awesome. I will continue to work with Third Tier clients and ASP subscribers to the best of my abilities. The clients are great and it's a great gig.

Well, that's about all for now. Thank you for reading and supporting us over the years!

NOTE: The Third Tier Brain Explosion slated to happen in September as a GFI conference Pre-Day has been cancelled due to the lack of sign-ups. Perhaps all y'all could comment on some of the conferences you attend, why, and where they are located. That would help us to pick a venue that would work better for you.

Philip Elder
Microsoft Cluster MVP
MPECS Inc.
Co-Author: SBS 2008 Blueprint Book
Chef de partie in the SMBKitchen ASP Project
Find out more at
Third Tier: Enterprise Solutions for Small Business

Friday, 11 July 2014

Third Tier Brain Explosion: Technical Training Sept 7-8 in Orlando

If you’ve been to one of Third Tier’s Brain Explosion (BE) then you know to expect a day packed with intense tech. This is the 4th annual Brain Explosion. We’re splitting the BE into two days so you can properly digest the content and have time to socialize in between. This makes it an intimate affair where you’ll get to know the speakers and your fellow attendees both during the event and afterwards socially.

We’re very social; we’re very technical and this event is going to be as fun as learning new tech can be. Our staff dedicates their days to supporting IT Professionals in the SMB market space and they run their own small IT firms too! The content of the Brain Explosion is designed to be boots on the ground issues of today that you can go and implement right away.

Read all about it: http://www.thirdtier.net/brain-explosion/

You can register one of two ways. Register for the Brain Explosion as a pre-day event at the GFI conference. This option will allow you to automatically extend your hotel stay too. Or registered directly with Third Tier for the Brain Explosion only. For this second option, register an account at http://www.thirdtier.net/helpdesk, press Payments, Buy Credit, Purchase more Support, add the BE to your cart and proceed to checkout. It’s $99 which includes Content, food and a pub meet up. Questions? Email Amy@thirdtier.net
 
Onward,

Amy Babinchak, Susan Bradley and Philip Elder
SMBKitchen ASP

Thursday, 26 June 2014

CRITICAL: Seagate 1200 SSD Firmware Update Required for 2012 R2 Storage Spaces

We got hit with this today:

image

  • Get-PhysicalDisk

Since this was our second run at standing up this Scale-Out File Server cluster with things not working as expected we began to dig in.

During a Space creation process the symptom of the Volume Format phase hitting an error happened. We jumped into PowerShell to poll the disks right then and saw the above.

The Disks:

image

  • Get-PhysicalDisk | where MediaType -eq SSD | ft Model,FirmwareVersion -AutoSize

The PowerShell to get the above information along with the serial numbers:

  • Get-PhysicalDisk | where MediaType -eq SSD | ft Model,FirmwareVersion,SerialNumber -AutoSize

Go to Seagate's Support site and choose the Download Finder.

image

Enter the serial number, just the first set of digits before the four 0000 pattern as the number was repeated twice for us, your country, and then under Certificate click on the Click here link and _not_ the Email Me link.

image

The highlighted link downloads the actual firmware ZIP file.

A copy of Seagate's SeaTools is required to update the drive's firmware.

If the SSDs are in a cluster setting, as they are here, make sure to properly drain the nodes and shut down the cluster (TechNet). Then shut down all nodes but one.

Run the firmware update and reboot that node. Bring the Cluster back online and fire up the other nodes one by one.

NOTE: We have _not_ tested this firmware yet.

A Microsoft Forum post brought us into the right direction: Clustered Storage space degraded and SSD disks "starting".

Remember how it has been mentioned ad nauseam on this blog about how we are very careful about testing our deployments? Well, in this case we were a part of the planning phases for this cluster but did not have access to these particular SSDs prior to deploying in a Michigan Data Centre.

This situation sure brings home the point that we _always_ need to test our setups before deploying them at client sites or on behalf of our clients.

EDIT: Brain was five steps ahead of fingers so the "Firmware Update" in the title never made it into the original post! :)

Philip Elder
Microsoft Cluster MVP
MPECS Inc.
Co-Author: SBS 2008 Blueprint Book
Chef de partie in the SMBKitchen ASP Project
Find out more at
Third Tier: Enterprise Solutions for Small Business

Monday, 23 June 2014

Third Tier Brain Explosion: Our New SBS!

I will be presenting along with our Third Tier team as a Pre-Day Brain Explosion (BE) for the GFIMax conference in Orlando, Florida on the afternoon of September 7th and the morning of the 8th.

We are splitting things into two days to give folks a chance to both digest all of the content but also to allow BE attendees the chance to meet and chat with us.

Cost for the Pre-Day BE is a very reasonable $99.

Registration for the event can be done on the Third Tier Portal under the Purchase Support option (windowed scroll bar in there that's a bit difficult to see) or on GFI's site (Pre-Day Landing Page).

The content I will be working with will be the technical aspects of delivering our new SBS.

What is that you might ask?

S = Small

B = Business

S = Solution

It is our very complete solution set that we have been deploying to our client sites for about six months to a year now.

  • RWW/RWA Replacement
    • RDWeb and RDGateway
  • Exchange/OWA/EAS
    • Exchange 2013 CU4 (SP1)
  • SharePoint
    • SharePoint Foundation 2013
  • Remote Desktop Services
    • Remote Desktop Session Host and RemoteApps
    • RD Endpoint access

Our last ASP SMB Kitchen subscriber chat was spent walking through SBS and how it is essentially seamless to our clients. One WAN IP is all that is required just as it was with Small Business Server.

The beauty of the solution is the simplicity with which things change for our end users: Virtually not at all. :)

We supply an on-premises solution that gives everything Small Business Server has given them. Plus, we clearly demonstrate that our product is on par with or better than any Cloud based solution out there just as SBS has been to small businesses for the last ten or more years.

Looking for that on-prem replacement? Please subscribe to our ASP Project as our monthly chats and content supplied by our team will facilitate that search.

Want the technical pearls to deploying this kind of solution for your clients? Then please do register for the Brain Explosion and wear a brain bucket! :)

Thanks for reading.

Philip Elder
Microsoft Cluster MVP
MPECS Inc.
Co-Author: SBS 2008 Blueprint Book
Chef de partie in the SMBKitchen ASP Project
Find out more at
Third Tier: Enterprise Solutions for Small Business