Tag Archives: sql2014

Why I won’t be running SQL Server Standard Edition on Business Critical Servers!

Posted on by

I keep getting questions around what core features are missing in Standard Edition of SQL Server,and why there is a need to install Enterprise edition for business critical systems.Here are my top features which are missing in standard edition –

Scale limits

Memory – Standard only supports 128 GB memory.

Memory is relatively cheap now,and I can’t think of any reason why someone can’t have more memory on business critical servers.

CPU cores Limited to lesser of 4 Sockets or 16 cores.

High Availability

AlwaysOn Failover Cluster Instances – Possible but only 2 nodes supported.

AlwaysOn Availability Groups*.

*AG is the future of HA/DR strategy within SQL Server and because of great demand, Microsoft will be allowing AG in standard edition of SQL 2016, however it will only support 2 nodes. Support of just 2 nodes is not sufficient for a solid HA/DR architecture.

Online page and file restore.

Page restores can come handy if you have to deal with corruption of a very large database and corruption is only for very few pages.

Online indexing.

You don’t really want to lock down the table while doing index maintenance right? Online index operations are key to maintain the uptime of the database.

Online schema change.

Hot Add Memory and CPU.

Database Snapshots.

Scalability and performance

Table and index partitioning.

Database size is increasing like never before and 3-5 TB DBs are very common these days. Maintaining these DBs without a partitioning strategy can be a very bad idea.

Data compression.

One of the most ignored features in SQL Server. Compression can be really helpful to save some space without much overhead.

Resource Governor.

IO Resource Governance.

In-Memory OLTP.

In Memory OLTP might be having lot of restrictions today, however things are going to change soon. I’m pretty sure that there will be multiple use cases when this feature is stable.

Memory optimized columnstore indexes.

Column store indexes dramatically improves DW workload performance and this feature will get more and more enhancements in the future.

Security

Transparent database encryption.

Conclusion – 

Business critical systems needs to be treated as critical as there is a potential risk of losing revenue if these systems are not highly available and uptime requirements are not met.Performance is also a key factor, and SQL Server standard edition is not the right edition for these requirements.

Advertisement

Full Domain Trust and Natively Compiled Stored Procedures – In Memory OLTP

Posted on by

Recently one of my developers brought this specific problem to my attention. He was trying to create a natively compiled stored procedure(With Execute as Owner), and was getting an error –

ERROR:

Could not obtain information about Windows NT group/user ‘Domain\UserName’, error code 0x5.

My first reaction was to check the builds and to ensure that the user is running latest CU.(Not yet SP1 !) . The user was indeed running the latest SQL 2014 build.

I decided to repro this issue at my end, and I was able to create/compile the procedure without any errors.

My next troubleshooting direction was to check user permissions, and understand more about the environment where the user is running into this error.

The user in this particular scenario was using his domain credentials to create the SPROC. The account was something like  HQ\Developer. HQ is the domain and Developer is the user name.

SQL Engine Services was running under account LAB\SQL.LAB is the domain and SQL is the SQL Service Account.

I read more about Natively Complied Stored Procedures, and especially focused well on this MSDN article, Creating Natively Compiled Stored Procedures.

This portion of the article was my focus area(Underlying the main point) –

” Regarding EXECUTE AS and Windows logins, an error can occur because of the impersonation done through EXECUTE AS. If a user account uses Windows Authentication, there must be full trust between the service account used for the SQL Server instance and the domain of the Windows login. If there is not full trust, the following error message is returned when creating a natively compiled stored procedure: Msg 15404, Could not obtain information about Windows NT group/user ‘username’, error code 0x5.”

In my case there was one way trust between HQ and LAB, however LAB didn’t trust HQ. The error message which the user got was obvious.

There are multiple workarounds to solve this problem and its listed in the article too –

  • Use an account from the same domain as the Windows user for the SQL Server service.
  • If SQL Server is using a machine account such as Network Service or Local System, the machine must be trusted by the domain containing the Windows user.
  • Use SQL Server Authentication.

I decided to leverage a SQL account and the user was able to create the procedure with that specific account.

Conclusion – 

In memory OLTP is a very exciting feature and there are a lot of dependencies which you should be aware before deciding to deploy it on Production.

Thanks for reading and keep watching this space for more !

This one enhancement makes AlwaysOn Availability Group better in SQL Server 2014 !

Posted on by

I keep getting questions on what is improved other than the increased number of secondaries(8 now) with AlwaysOn Availability Group(AG) in SQL 2014. In this blog post  I would like to talk about one major enhancement around the increased readable secondaries availability in SQL Server 2014.

In SQL Server 2012 there is a good chance that your read workloads can be killed during network failures. Typically this happens in a multi subnet environment where you have AG nodes in 2 or more data centers.If network failure results a communication failure between primary and secondary replicas, then the secondaries will go into resolving state.In resolving state, the read workloads are basically killed. (Yes, yet another reason why you need to test your AG deployments)

With SQL Server 2014, there is good news. Read workloads are not impacted during network failures, primary replica down or cluster quorum issues.

Yes, it stays up all the time, and will be serving your read workloads. The replica will be in resolving state, however the databases are still available for your read connections/workloads.

Keep in mind : Read-only routing is not yet supported for this, and you cannot use listener for your reporting workloads.Brent Ozar calls out this in his blog post here.

Lets now walk though this enhancement :

I will be using a 2 node AG setup and the data transfer mode(Availability Mode) is asynchronous.

If I query the DMV sys.dm_hadr_availability_replica_states I can see the replica states.

SELECT role_desc,connected_state_desc  FROM sys.dm_hadr_availability_replica_states

role_desc               connected_state_desc
PRIMARY              CONNECTED
SECONDARY        CONNECTED

Lets double-check if the data transfer is set as asynchronous

SELECT  is_primary_replica, synchronization_state_desc FROM sys.dm_hadr_database_replica_states

is_primary_replica       synchronization_state_desc
0                                  SYNCHRONIZING
1                                  SYNCHRONIZED

Everything looks good, and now lets take down the primary replica.

Note : Dont try this at home ! (I mean production)

Okay, As this is a test instance I really don’t care and what I did was shutdown the SQL Services to take the primary down.

If we query the DMV again, we can see

SELECT role_desc,connected_state_desc  FROM sys.dm_hadr_availability_replica_states

role_desc                connected_state_desc
RESOLVING           DISCONNECTED

And note, there is no entry for primary as we I have taken it down.

Even though the secondary replica is in resolving state, read connections just works.

USE DemoAG
GO
SELECT COUNT (*) AS [TotalRowCount] FROM [dbo].[InsertCount]

TotalRowCount
6886

The secondary database is in NOT SYNCHRONIZING, NOT_HEALTHY state, however its ONLINE.

SELECT synchronization_state_desc,synchronization_health_desc,database_state_desc FROM sys.dm_hadr_database_replica_states

synchronization_state_desc         synchronization_health_desc       database_state_desc
NOT SYNCHRONIZING               NOT_HEALTHY                           ONLINE

Here is a view from Management Studio :

AG Resolving

Conclusion:

Yes, this is a very good enhancement which will make sure your readable secondaries are highly available and I’m pretty sure that this is one of those enhancement which will force you to deploy AGs in SQL 2014 rather than opting SQL 2012.

Thanks for reading and keep watching this space for more !

SELECT INTO Improvements in SQL Server 2014 !

Posted on by

SQL Server 2014 is a great release with high performance features like In Memory OLTP, Updatable Column Store Indexes etc.

These features still remains as the top favorite and the most compelling reasons for organizations to do an upgrade, However SQL Server 2014 also includes very small but very relevant and useful enhancements, and this post talks about one of those enhancement:

SELECT INTO now runs Parallel !

MP900386077

When I first heard about this feature my first impression was: What…was it not running parallel before ?

No, SELECT INTO was single threaded all these years and with SQL Server 2014 this will now use multiple threads.

Let’s do a quick walk-through of this enhancement.

In this walk-through we will use a table named MasterData which has around 12 million rows.

--Select Count
 SELECT COUNT(*) AS Row_Count FROM [dbo].[MasterData]

Row_Count
11,999,880

Data from this table will be copied over to a different table named MasterData_Copy using SELECT INTO operation.

--Turn on Actual Execution Plan
 SET STATISTICS TIME ON
 GO
 SELECT * INTO [dbo].[MasterData_Copy] FROM [dbo].[MasterData]

This operation was completed in 6 seconds (Desktop class machine) and if we examine the execution plan we can observe parallelism.

SELECTINTO1

And the distribution of row’s among multiple threads:

SELECTINTO2

Now let’s make this little more interesting. We will change the database compatibility level to SQL2012(110) and see what impact that brings to the SELECT INTO operation.

--Change Database Compatibility
 ALTER DATABASE SELECTINTO_Demo SET COMPATIBILITY_LEVEL = 110
--Turn on Actual Execution Plan
 SET STATISTICS TIME ON
 GO
 SELECT * INTO [dbo].[MasterData_Copy] FROM [dbo].[MasterData]

This operation was completed in 11 seconds and it generated a serial plan.

SELECTINTO3

And here is the row distribution.

SELECTINTO4

<Update 1/13/2015 2 PM> 

I received a comment from Toni which says “It might be a good idea to set MAXDOP, most likely having 16 threads isn’t the best option”.

I absolutely agree with this and a well tested MAXDOP setting will definitely bring better results and this is something which you have to keep in mind when queries goes parallel.

</Update>

Conclusion:

If you are using SELECT INTO a lot in your environment, then I would highly recommend you to compare and see what improvements you can again out of SQL Server 2014.

Monitoring Memory Usage of Memory Optimized Objects – SQL Server 2014

Posted on by

Starting SQL 2014 monitoring memory usage of the memory optimized objects is super important to ensure that your instance don’t run out of memory and cause real bad situations.

One way of ensuring that the memory optimized objects wont utilize a certain amount of memory is by setting up resource governor. This is a great level of control when you have multiple databases running on the same instance and you don’t want memory optimized tables eating up the whole memory.

Is there an easy way to get a quick overview of memory usage of the memory optimized tables in SQL2014?

Yes, you have a real good SSMS report just for this purpose. This one report utilizes DMVs under the hood and provides you with some valuable information.

InmemOLTP Reports

Lets do a quick walk through of what this report gives –

InmemOLTP Reports1

At any point of time I will be interested to see the usage of Table Used Memory counter.

Index Used Memory is based on the bucket count which you mention during non clustered hash index creation.

The query which the report runs under the hood to give you the usage related value is –

SELECT t.object_id, t.name, 
ISNULL((SELECT CONVERT(decimal(18,2),(TMS.memory_used_by_table_kb)/1024.00)), 0.00) 
 AS table_used_memory_in_mb,
ISNULL((SELECT CONVERT(decimal(18,2),(TMS.memory_allocated_for_table_kb - TMS.memory_used_by_table_kb)/1024.00)), 0.00) 
 AS table_unused_memory_in_mb,
ISNULL((SELECT CONVERT(decimal(18,2),(TMS.memory_used_by_indexes_kb)/1024.00)), 0.00) 
 AS index_used_memory_in_mb,
ISNULL((SELECT CONVERT(decimal(18,2),(TMS.memory_allocated_for_indexes_kb - TMS.memory_used_by_indexes_kb)/1024.00)), 0.00) 
 AS index_unused_memory_in_mb
FROM sys.tables t JOIN sys.dm_db_xtp_table_memory_stats TMS 
ON (t.object_id = TMS.object_id)

This query leverages the power of a new DMV dm_db_xtp_table_memory_stats

Conclusion

SSMS reports are great ways to get a quick overview of what is happening, and I expect more and more powerful reports getting incorporated to SQL 2014 during the coming days.

Thanks for reading and keep watching this space for more.

Checkout SQL Server 2014 CTP1 and Windows 2012 R2 preview up in the cloud

Posted on by

You must be really busy spinning up VMs to test SQL Server 2014 CTP1 and Windows Server 2012 R2 preview.

If your on-premise desktop computer/laptop is running low on resources and if you don’t want to take the pain of handling multiple VMs locally,then you have an option to test them out right up in the cloud.

Yes, You can always spin up a Windows Azure VM and do all your testing and shut that down once you are done with your work to save some money.

Sounds like a deal right? Yes, its pretty easy and fast. It takes less than 10 minutes to provision a Windows Azure VM.

Lets look at the options we have in Azure for the test VMs –

1. SQL Server 2014 CTP1 Evaluation Edition on Windows Server 2012

This image will setup SQL2014 CTP1 for you on Windows Server 2012. Once the VM is provisioned and is in running state, you can connect it via RDP to see that SQL Server is ready to test.

VMs

2. Windows Server 2012 R2 Preview

This image will provision a VM for you with Windows Server 2012 R2 preview. You can spin up multiple VMs to do your hardcore testing with Windows.

VMs1

Conclusion

This is indeed a great way to test the latest builds with minimum efforts and you always have the option to shutdown the VMs after use to avoid a heavy bill.

Thanks for reading and keep watching this space for more.

SQL Server 2014 – A new and exciting journey !!! “Into the cloud”

Posted on by

Today Microsoft announced their next major version of SQLServer, SQL Server 2014 at Tech-Ed NA 2013.

This version of SQLServer is indeed a reflection of Microsoft’s vision towards the future computing. A bright and colorful “In to the Cloud” future !

The journey of SQLServer from SQL2000 – SQL2014 is just amazing. On a related note,Its worth reading this post by Quentin Clark which explains this journey with the help of a neat diagram.

So when can I start exploring the new features? Where can I download the same to evaluate it ?

Thinking boy

Here is the answer for you  –

Go to http://www.microsoft.com/en-us/sqlserver/sql-server-2014.aspx  and choose option Get-Notified >

You will be asked to enter few details and you can sit and wait for that “email” which will eventually allow you to download the bits and do some testing.

What’s new in SQL2014 that I am excited about ? Lets do a quick walk through of my favorite features –

Note- This is not an extensive list of features which SQL2014 will provide you.I will write about that altogether in a different post.These are some of the features which I really like about.
  • In Memory OLTP(Code named Hekaton) – 

This is one of my favorite features that I’m really looking forward to work with. This feature will be called as SQLServer In-Memory OLTP Engine for SQL2014.

SQLServer In-Memory OLTP Engine will boost  OLTP workloads to perform remarkably better. The beauty of this feature is you can now pick tables and define it as memory optimized and these tables are fully transactional.

Anytime I will suggest you to download and read the white paper ‘SQL Server In-Memory OLTP Project “Hekaton” Internals Overview for CTP1’ by Kalen Delaney(B/T) to understand this feature really well.

  • 8 Readable Secondaries –

With SQL2014 we will have the ability to add up-to 8 readable secondaries for our read-only workload(Mainly reporting queries).

  • Always on to Windows Azure Virtual machine – 

This is one feature which I’m really looking forward. This will enable us to add a secondary replica directly in Windows Azure Virtual Machine. This is one feature which the CIOs would love, as you pay per use for Windows Azure VMs.

  • Buffer pool extension using SSDs – 

This can improve query performance by allowing the use of non-volatile solid-state drives to reduce SQLServer memory pressure with no risk of data loss.

Brent Ozar(B/T) has a great post explaining this and you can read it here.  He also have provided a great deal of information about other SQL2014 features too. Always a great read.

  • Performance Data Collector – 

I’m also looking forward for the enhancements which Performance Data Collector module will come with. The details are not yet out on what is new, but I definitely hope that there will be some good changes here and there will be deep integration with SSMS for this module.

Conclusion

With SQL2014 I’m pretty sure that your move towards the cloud will be more easy and a nice experience and I’m really looking forward for it.

Thanks for reading and keep watching this space for more(Lot of SQL2014 for sure !)