I’m happy to say that the webcast which I did with Jason Hall on Thursday October 30, 2008 has been archived for after the fact viewing. They say that it is up there for a limited time, so I’m not sure how long it will actually be available for.
You can find the webcast over in Quest Software’s Events section.
A while back I was talking to a Microsoft Support Engineer and he had mentioned that in a high load Service Broker environment such as ours there can be some impressive performance improvement can be achieved by reusing the service broker sessions.
The cost of creating and closing a new conversation for every message is about 4x, while the performance increase when receiving messages is about 10x. Remus Rusanu talks more about the numbers and shows a solution for reusing conversations on his blog posting Reusing Conversations.
I liked Remus’s solution, but an issue that I had with it was that I didn’t want to have a different conversation for each spid. If I used this method I would end up having hundreds of conversations open and I’d need a job to close them. Within our application any number of events can trigger a service broker message to be sent off. We usually have a few hundred threads logging in and out of the database at any one time.
This required that I take Remus’s solution and make it more flexible before moving it into our environment. The solution that I’ve come up with supports a single conversation for each process within our application. And at random intervals those conversations are closed, and new ones are started. Continued »
So yesterday I was asked to setup a few x64 workstations on our ESX server. The Vista x64 machines detected the LSI controller no problem, but the Windows XP machines couldn’t find the controller, and the drivers that VMware included with ESX (3.0.2) were x86 only drivers.
Tomorrow morning at 8am Pacific / 11am Eastern / 4pm London Jason Hall and myself will be talking about how to find out how healthy your SQL Server Enterprise is using both the native tools as well as Quest Software’s Spotlight On SQL Server Enterprise.
Today is the unofficial launch of the new SQLServerPedia.com Wiki. The official launch for the wiki is during PASS, so you have to go to the /wiki folder manually. The SQLServerPedia.com site is run by Quest Software, and the Wiki contains all the information which used to be contained in the KnowledgeXpert product which they used to sell. They are know giving it away for free to the entire SQL Server community.
I could go on all day about it, but forunitely I don’t have to because Brent Ozar already has in his post “The answer to SQL Server training problems“. I highly recommend reading Brent’s post about the site. It gives some great background and tells all about the site. Continued »
I’ve finally found an offline blog post editor which I like. With all the traveling that I’ve been doing recently its very handy to be able to edit them offline so that I don’t have to pay for internet access at airports. It also lets me blog while on an airplane without having them write them in Word or notepad and uploading them later.
Here are the slide decks and sample code from my sessions at this weekends SoCal code camp.
While the first two are both two part sessions, there is only one download for both halves as they run together.
I would say that yes you should. The PASS summit is an excellent learning opportunity that simply shouldn’t be missed if you can’t avoid it.
No matter how much you think you know about SQL Server, there is always more to learn.
I tend to think I know quite a bit about SQL Server and I enjoy going to PASS, because there is so much that I don’t know about SQL Server, this is a great opportunity to learn from some of the best minds in the SQL Server community and Microsoft.
You can still register for the summit on the PASS Community Summit 2008 web page.
I hope to see you there.
I’d like to thank the Orange County SQL Server Users Group for having me come and speak at their meeting last week. I had a great time speaking with everyone, and there were a lot of good questions.
If you didn’t have a chance to download the slide deck you can download it here. If any new questions come up, feel free to post them here, or email them to me directly.
Here’s a handy query which I’ve written which I’ve used to see how much of each table is loaded into the SQL Server Buffer Cache. It shows you how much data is in memory, how much space is used on disk, and the percentage of data in memory compared to the amount of data on disk.
SELECT sys.tables.name TableName,
sum(a.page_id)*8 AS MemorySpaceKB,
SUM(sys.allocation_units.data_pages)*8 AS StorageSpaceKB,
CASE WHEN SUM(sys.allocation_units.data_pages) <> 0 THEN SUM(a.page_id)/CAST(SUM(sys.allocation_units.data_pages) AS NUMERIC(18,2)) END AS 'Percentage Of Object In Memory'
FROM (SELECT database_id, allocation_unit_id, COUNT(page_id) page_id FROM sys.dm_os_buffer_descriptors GROUP BY database_id, allocation_unit_id) a
JOIN sys.allocation_units ON a.allocation_unit_id = sys.allocation_units.allocation_unit_id
JOIN sys.partitions ON (sys.allocation_units.type IN (1,3)
AND sys.allocation_units.container_id = sys.partitions.hobt_id)
OR (sys.allocation_units.type = 2 AND sys.allocation_units.container_id = sys.partitions.partition_id)
JOIN sys.tables ON sys.partitions.object_id = sys.tables.object_id
AND sys.tables.is_ms_shipped = 0
WHERE a.database_id = DB_ID()
GROUP BY sys.tables.name
When you run this code, some of the smaller objects will show more pages in memory than on disk. It appears that this is normal. My guess is that SQL Server is preallocating data pages so that if data is added to these pages there is space allocated to them already in order to save time when adding data to the table.
I’ve removed the system objects from this query, as you can’t do anything to optimize them.