I won’t go into great detail here but in short, if VSSAdmin is not letting you delete shadow copies and is throwing this message:
Error: Snapshots were found, but they were outside of your allowed context. Try removing them with the backup application which created them.
There is another (less safe…) program called DiskShadow which will.
It’s built into windows server… and using it to get rid of pesky shadow copies that won’t otherwise goes away looks like this:
DISKSHADOW> Delete Shadows All
and away they go…
One of the best blog articles I came across that details both VSSADMIN and DiskShadow tool and why VSSAdmin sometimes falls short is here:
After working with RDS (Remote Desktop Services, previously known as “Terminal Services”, also referred to “The biggest pain in the rear and the only way to get more than two remote desktop sessions on a server because Microsoft either hates admins, hates this product, or both”) I have come to the conclusion that Microsoft really needs to make something which should be simple, simple. Alas it isn’t simple. And today’s topic…
You have an RDS licensing and management server… you have several Session Host servers as part of your deployment. At some point one of those servers died and was henceforth removed from the domain. It’s gone, never to return. You login to manage Remote Desktop Services licensing… and are told that a bunch of servers are missing from the management pool.
Something along these lines:
The following servers in this deployment are not part of the server pool:
The servers must be added to the server pool.
Instead of MS just letting you manage RDS, you have to fix this first. Fine… you go to add them back in but wait…. you have that server which was a session host which died a horrible death and therefore cannot be added back in… What now?
I regularly use the Microsoft Windows sysprep tool to create template Windows Server 2012 R2 systems for wider deploy using cloning. Sysprep is used to modify a pre-configured Windows system and create an image or “template” so that you can create unique copies of it for faster system deployment. Failure to use syprep before cloning a windows machine can cause odd issues, especially in an enterprise environment with Active Directory.
Sysprep is a wonderful tool but it has a few quirks. One such quirk is that it routinely wipes out all of the mounted drive information other than the system drive. This means every time you create a clone from a sysprep’d system image you have to go through and re-assign drive letters. This is fine on a simple server with one extra data server. This is a hassle on a database server which might have five extra drives.
After dealing with this again recently I finally decided to do a google search and came across a simple solution: (more…)
I will keep this short and sweet. We have servers in our environment that have multiple IP addresses assigned to a single NIC. That’s normally just fine. However on occasion I will have very strange issues occur where essentially all networking appears to be working and yet web browsing won’t work. I can ping my default gateway, ping other systems in the same subnet, telnet out on port 80 and 443, etc, etc. But the network connectivity still behaves oddly. What’s the issue?
It all has to do with networking logic decisions made many years ago (I believe as far back as Windows Server 2000) by someone at Microsoft. (more…)
I have been fiddling about with setting up a SQL Server 2012 Failover cluster using an Equallogic SAN. After a whole lot of digging about I found two different posts on two different sites which got me about 90% of the way there. However there were some key “gotcha’s” and other information that was missing in both cases and I wanted to document those here in addition to referencing the articles I followed for my setup.
BTW – Just my 2-cents, but setting up clustering is complicated… especially when you throw SQL in the mix. It isn’t bad once you have done it a few times (I tested again, and again, and again in a virtual environment) but there are honestly like 50+ considerations to take into account to ensure everything goes correctly.
I am assuming if you are here you already have a general understanding of failover clustering, know what you are wanting to do and why. This article also doesn’t really cover all aspects of high-availability. I don’t discuss how your SAN(s) should be networked for example. I do touch on a few items though that fall in this area. This isn’t meant to be comprehensive and a lot of it is just for personal reference.
So here are some tips if this is your first go around. These are in NO particular order or grouping (this is very “stream of thought”) so I would suggest reading this from start to finish at least once rather than referencing it as you are going through your setup.