AI Adsense

Thursday, October 06, 2016

Windows Server 2016 Licensing Explained

Microsoft Windows Server 2016 became officially available recently on 9/26/2016 and as I'm sure you have heard the licensing changed from a per processor model to a per core model.

Before you read any futher here's a simple thing to keep in mind when purchasing this new licensing model:

Once you have enough cores purchased (16 core minimum) to cover the metal, every 8 packs (16 cores) of licensing cover 2 more virtual machines.

The first thing that people have had a hard time understanding is the new "Per Core" licensing and even with Microsoft's own explanation it seemed impossibly hard to understand so I'm going to try to break it down simply so everyone can get a quick grasp on how it works.

Server 2016 licensing is sold in 2 core packs.  The very minimum no matter how small your server is a 16 core purchase.  Microsoft says this would be the same cost as Server 2012 R2 Standard licensing.

You purchased a small server for a business that has only 4 cores and 8G RAM.  The correct licensing for Server 2016 would be 16 cores purchased.

You MUST cover all cores on the server prior to taking into consideration the number of virtual machines if it is a VM host.

Your server has 2 processors that are 12 cores each.  Total of 24 cores.  You plan on only running 2 virtual machines on this server.  You must purchase 12 x 2 packs of Server 2016 core licensing.  If you move up to 4 virtual machines guess what?  You're going to need to purchase another 4 packs of licensing to cover the next two.  This is a total of 16 packs and every 8 packs = 2 VMs.

After you have covered the number of cores on the physical server, every 8 packs (16 cores of licensing) after that will give you two more virtual machines of no more than 8 cores each.

Your plan is to run 5 virtual machines on a server with 2 x 8 core processors.  You must purchase the initial 8 packs (16 cores) to cover the server's physical processors.  Then from there you must purchase another 8 packs for the next 2 virtual machines and then another 8 packs for the 5th virtual machine leaving you 1 more virtual machines you can build without needing more licensing.

So far this news isn't extremely terrible but there are two more things on top of all of this you need to know.

1. Once you approach the 8th virtual machine, it becomes more cost effective with the new licensing model to purchase the Datacenter edition.  The cost is going to be very high for customers so virtual server sprawl is going to start being a real issue.

2.  Microsoft says you MUST license a virtual host to carry all of your virtual machines in the event of a failure of one host.  This is a massive cost purchase in the event of a multi-host virtual environment.

Let me explain that one a bit better.  You have a decent sized business with three VMware hosts.  Each host has 2 x 10 core processors and you're running 24 virtual machines evenly spaced out over all three hosts so you have a 3 x 8 scenario with your virtual machines.

Microsoft now says you must purchase not just a total of 96 packs (remember 2 cores per pack) but you must purchase a total of 96 packs x 3 hosts in the event that two of them were to go offline and you have to run all of your VMs on one host and because you can vMotion them around.  This simply means a business would now need to purchase 288 cores of licensing to be properly licensed.

I know this is a bit wordy but I hope it makes a little better sense than all of the other confusing information out there about it.  I'm confident this is correct as our licensing supplier has been through the offical training and this was how it was explained to me.

Good luck.

Monday, September 19, 2016

ESXi 6 - The Primary GPT table states that the backup GPT is located beyond the end of the disk

I was attempting to reinstall ESXi 6.0 on a server that had a previous datastore on a RAID 5 array.  The array was missing a disk so I destroyed and recreated the array minus the missing disk and decided to just do without the storage of the one disk.

When I attempted to add the array into my fresh ESXi 6 load I got this error:

Call "HostDatastoreSystem.QueryVmfsDatastoreCreateOptions" for object "ha-datastoresystem" on ESXi "" failed

Not knowing how to fix this I did some research and found out a quick and easy repair.

1. Enable SSH on the VMware host.
2. Connect and run "ls -lha /vmfs/devices/disks" and this will list your disks with their disk ID.
3. Run the following command on the disk "partedUtil getptbl /vmfs/devices/disks/naa.5000c501234597a333"

This will return the following output if you have chosen the correct disk:

Error: The primary GPT table states that the backup GPT is located beyond the end of disk. This may happen if the disk has shrunk or partition table is corrupted. Fix, by writing backup table at the end? This will also fix the last usable sector appropriately as per the new reduced size. diskPath (/dev/disks/naa.5000c501234597a333) diskSize (286748000) AlternateLBA (570310655) LastUsableLBA (570310622)
Warning: The available space to /dev/disks/naa.5000c501234597a333 appears to have shrunk. This may happen if the disk size has reduced. The space has been reduced by (283562656 blocks). You can fix the GPT to correct the available space or continue with the current settings ? This will also move the backup table at the end if it is not at the end already. diskSize (286748000) AlternateLBA (570310655) LastUsableLBA (570310622) NewLastUsableLBA (286747966)
Error: Can’t have a partition outside the disk!
Unable to read partition table for device /vmfs/devices/disks/naa.5000c501234597a333

Apparently deleting the RAID array didn't fully erase all of the previous partition information.  Now it needs to be cleared manually.

4. Run the following to clear it:  "partedUtil setptbl /vmfs/devices/disks/naa.5000c501234597a333"

That's all there is to it.  By creating a msdos partition on the disk it will clear the previous error and allow esxi to create a datastore there with no errors.

Good luck!

This is ve

Friday, July 01, 2016

[SOLVED] Blank Screen after ECP Login Exchange 2013/2016


After logging into the Exchange ECP with the correct credentials you get just a blank screen with no errors messages.  This will almost always happen if you remove a certificate using the certificates MMC or even sometimes if you remove one using the ECP.

Below is what you need to check to fix this problem:

Since this change is not reflected in the backend website you have to make sure that the Exchange ECP site is looking at the same certificate in both locations.

IIS Default Site SSL Certificate

IIS Back End SSL Certificate

The certificate choice in both of these locations have to match exactly then the issue is resolved.

Good luck!

Thursday, April 28, 2016

[SOLVED] Windows 10 File Explorer Not Responding, Freezing, Locking Up

So lately this problem for me has gotten out of hand.  After some research on the web and a couple of Microsoft articles here's what you can do to fix this issue:

1. Open Control Panel
2. Select File Explorer Options
3. Click on CLEAR at the bottom beside of "Clear File Explorer History".

Apparently bad data in this windows can cause Explorer to freeze up.

One more thing.  If this doesn't work then change the option at the top of the File Explorer Options to open to "This PC" instead of the default which is quick access.

Hope this helps if you're running into this issue.

Monday, January 04, 2016

UCS login Error: Invalid HTTP Response

So trying to check our UCS server I ran into this issue today.

If you aren't running the latest UCS software then certain versions of Java will toss this error.  If you need to get logged in and don't have the option for updating the UCS software right away then here's how you can get around it.

Just enable HTTP via SSH and you can then login.  Remember to disable this once you have your UCS software and Java versions current because until you do everything is transmitted in clear text.

UCS# scope system
UCS /system# scope services
UCS /system/services# enable http
UCS /system/services# disable http-redirect
Warning: When committed, this closes all web sessions
UCS /system/services# commit

Once this is done now you can connect to the UCS Manager without the error.