Forum posts to 'Support Forums'

Re: Benchmark Stop Responding

Ami — Thu, 01 Jul 2010 07:47:23 +0000

Hi atanas_g,

Thanks for contacting DBSophic.
Qure's error logs are located in the logs folder in the installation directory.
By default, this is 'C:\Program Files\DBSophic\Qure\Logs".
It could help us a lot if you can zip them and send them to the email address support@dbsophic.com.
I assume you mean the step is 'Applying Recommendations'? Can you also send a screenshot?
I would also try to open Management Studio's "Activity Monitor" (right click on server in object browser) and there you should be able to easily identify Qure's connection from the application name. In the connection's properties window, you'll be able to see what it does.
I would also take a quick look at the SQL Server error logs.

In any case, it might be perfectly normal for the benchmark to take a long time, potentially even a few days.
Its duration depends on many factors, among them are
> DB size
> Workload characteristics
> Available hardware resources
> Number and characteristics of generated recommendations
and many others.
Remember that the benchmark has a lot of work to do.
It replays the entire workload (multiple times) against the analysis database and also tries out different sets of recommendations for fine tuning.
If a lot of indexes or schema changes are being benchmarked, this could take quite some time.

If you don't see any errors and the connection is still 'alive', I would recommend to wait a little longer.

If you can, please send us the above info - we will investigate it and let you know if it is indeed hanging.
Feel free to contact us with any additional questions.

I recommend that you direct your questions from now on to the Email support address support@dbsophic.com as this forum will be shut down temporarily within a few days.

Thanks again,

DBSophic support.

Posted to: Benchmark Stop Responding Show Thread | Post Reply

Benchmark Stop Responding

atanas_g — Thu, 01 Jul 2010 07:09:20 +0000

Hi,
Im using DBSophic Qure demo version 2.0 to try its functionality. My trace file is about 330MB the DB that I'm analizing is 21GB. When I start analysis proces it run to step Recomendation generation an then the gold arrow moves to Benchmark step but the right side of window stays at step Benchmark recomendation. And in this condition the tool stays about 15 hours without any sign of activity. Is this normal. How long do I wait before considering faillure. Is there any logs to check for errors or something.

Posted to: Benchmark Stop Responding Show Thread | Post Reply

Re: OutOfMemory exception

Sandr — Mon, 28 Jun 2010 14:36:10 +0000

Hi, Ami,

As I'd written above, the version which I tried to use was 2.0 demo analysis, which not allowed to use more than 10 files.
Files were not large.
I tried to perform the same analysis 3 times with the same files and every time I received the same error in the same place.
When only 3 files of this set were used, this step was passed successfully.

According to beta version, I'll be very glad to receive it.

Thanks,
Alex

Posted to: OutOfMemory exception Show Thread | Post Reply

Re: OutOfMemory exception

Ami — Mon, 28 Jun 2010 11:58:56 +0000

Hi Alex,

The issue you are experiencing is a .Net GC (Garbage Collector) issue that is sometimes unable to handle very large amounts of data, such as Qure might generate.
Although it is quite rare, the GC is sometimes 'overwhlemed' and the .Net infrastructure itself generates the exception and terminates the process.
See This Link for more information.

We are aware of this issue and have been working hard to workaround it and to minimize its chances of occurrence.
We even plan to release a minor version within the next few weeks that addressees this issue specifically and reduces the chances for this error significantly.
For the mean time, you can do one of the following:
1. Retry to perform the analysis - There is about 50% chances that it won't happen again.
2. Use smaller traces - this reduces the chance of the error significantly as you already witnessed.
3. Within about 2 weeks, I can send you the Beta version of Qure V2.2 that should also workaround it.

Let me know if you wish to wait for the beta version and I will add you to the beta list.

Thanks and sorry for the inconvenience.

Ami

Posted to: OutOfMemory exception Show Thread | Post Reply

OutOfMemory exception

Sandr — Mon, 28 Jun 2010 11:09:34 +0000

I tried to execute demo analysis (10 trace files, 1 DB) on the DB serverr in test environment.

On the step of "Recommendation Generation" on the second step ("Analyzing batch database operation") when reaching the last operation (617 of 617) following exception occured:

===================================

Fatal error occured during analysis. Aborting analysis.

===================================

Exception of type 'System.OutOfMemoryException' was thrown. (DBSophic.Common)

------------------------------
Program Location:

at DBSophic.Common.Synchronization.TaskBase.ThrowCore(Exception ex)
at DBSophic.Common.Synchronization.TaskBase.<>c__DisplayClassa.<Throw>b__8()

When I tried to perform the same analysis on the same server with the same files (but not 10, only 3) this step was successfully finished.
Total size of log files is 100Mb
Size of analysis file is 163Mb.

Any ideas?

Posted to: OutOfMemory exception Show Thread | Post Reply

Re: Qure and prepared statements

Rayman — Sun, 27 Jun 2010 08:52:17 +0000

Great.
Let me know if you need any more assistance.
Ray Maor
Qure R&D manager

Posted to: Qure and prepared statements Show Thread | Post Reply

Re: Qure and prepared statements

addieleman — Tue, 22 Jun 2010 15:11:35 +0000

Sorry to react after such a long time. The prepare event is captured by TSQL, Prepare SQL. The event does not supply any text data. From your answer I conclude it is sufficient to run the trace with the template as defined, so I will do that now and ship the data to our application supplier who will use the data to get recommendations for query optimization.

Posted to: Qure and prepared statements Show Thread | Post Reply

Re: Trace Analyzer limitations

Sandr — Tue, 15 Jun 2010 12:48:45 +0000

Hi, Ami.

I'm working now with Trace Analyzer and .trc files.
When the Analizer will stop (according to limitation), I'm planing to start additional analyze of the rest of the files.
One problem is to understand what was the last analyzed file before the limitation reaching...

I'll post a results of such tricks...

Alex Lutsky

Posted to: Trace Analyzer limitations Show Thread | Post Reply

Re: Trace Analyzer limitations

Ami — Tue, 15 Jun 2010 12:43:27 +0000

Hi Alex,

Such large trace WILL take a long time to analyze, but I bet that trace analyzer will do it much faster than a human ;-)
Let us know how it worked out.

Ami.

TIP: If you store the trace in a SQL Server table, loading by Qure should perform a little better
AND
you will be able to easily delete only the rows that have been analyzed and perform a new analysis on the rest.

Posted to: Trace Analyzer limitations Show Thread | Post Reply

Re: Trace Analyzer limitations

Sandr — Tue, 15 Jun 2010 12:37:12 +0000

Hi, Ami.
I'm close to this.
As I'd write above, we collected traces of half of day from our production server and now we want to understand if there's something wrong (or, in other words, how we can to improve performance).
Total amount of trace files is about 18Gb. If I'll perform analyze of all of them in one thread, estimated .tran file size will be about 11G...

If Trace Analyzer will stop automatically when the limit will be rich, I'll try to perform additional analyze of the rest of the files...

Alex Lutsky

Posted to: Trace Analyzer limitations Show Thread | Post Reply

Re: Trace Analyzer limitations

Ami — Tue, 15 Jun 2010 11:53:44 +0000

Hi Sandr,

The short answer is 'Yes'.
The more accurate answer is 'It depends'
For Qure, It's not as 'clean cut' as 2GB / Analysis and 2GB / Benchmark. it is roughly closer to 3GB / 1GB and depends on quite a few factors as we try to make best use of the 4GB size limit that is enforced by the database storage infrastructure that we use to store the analysis data (SQL CE).

For Trace Analyzer, we can actually store much more data as the analysis phase does not need to store any database data or code.
Also, as opposed to Qure, Trace Analyzer does not need to parse the statements as it does not need to 'understand' them so much less data is stored for each event.

So to sum up:
The 4GB repository size limit is true for both Qure and Trace Analyzer.
You can use trace analyzer to analyze workloads that are far larger than double what you could in Qure.

Trace Analyzer will use all the available space and when it hits the 4GB limit (or close to it), It will stop the analysis and present the results of what was stored so far.

Just out of curiosity - did you actually hit this limit?

Ami Levin

Posted to: Trace Analyzer limitations Show Thread | Post Reply

Trace Analyzer limitations

Sandr — Tue, 15 Jun 2010 11:30:33 +0000

Qure V2 is limited in size of analyze file of 4Gb for store data. This file is separated for 2 parts: 2Gb for analysis and 2GB for benchmark.
What about Trace Analyzer? I suppose that it use the same engine for store data. In this case limitation must be 4GB because it not need to perform benchmarking. Is this correct?

And how can I (if all above is correct) analyze 1679 trace files with summary size 17,527,726,720 b (~16Gb)?
I'd started analyze and estimations are ~ 11Gb of result file.

Posted to: Trace Analyzer limitations Show Thread | Post Reply

Re: Qure and prepared statements

Rayman — Wed, 09 Jun 2010 07:42:02 +0000

Well, the Qure template takes just 2 events: RPC:Completed and SQL:BatchCompleted.
Generally, catching the sp_execute is sufficient for any statement, and the prepared statement creations phase (when the parsing takes place) is less relevant when looking at the workload.
Either way, could you tell me which event do the prepared statements that you are using raise?
And perhaps send me the text as an example so I can check if the auto system filters filter it out.

Ray Maor
Qure R&D Manager

Posted to: Qure and prepared statements Show Thread | Post Reply

Qure and prepared statements

addieleman — Tue, 08 Jun 2010 09:33:25 +0000

The application we're trying to analyze, uses prepared statements. I've just started SQL Server Profiler with the Qure template and it only seems to catch the sp_execute statements, but not the prepared statements. Is that OK or do I have to extend the events of the Qure template?

Posted to: Qure and prepared statements Show Thread | Post Reply

Re: Trace Analyzer not working

Ami — Wed, 19 May 2010 06:37:26 +0000

Hi Hitesh,

1. What error message are you receiving?
It would help if you provide a full quote of the error text.

2. Please note the mandatory requirements detailed on the right text box of the trace analyzer page:
http://www.dbsophic.com/traceanalyzer

3. What event types and data columns are you collecting in your trace?
Try to use Qure's trace template which you can download here:
http://www.dbsophic.com/assets/support/QureTraceTemplates.zip
Simply double click the *.tdf file to import the trace definition into profiler and start a new trace using that template.

Let us know if that solved your problem or if you require further assistance.

Sincerely,

DBSophic support.

Posted to: Trace Analyzer not working Show Thread | Post Reply

Trace Analyzer not working

Hitesh — Tue, 18 May 2010 17:45:33 +0000

I'm starting a new analysis with Trace Analyzer. I'm trying to add the trace file created with SQL Server Profiler. But I'm getting the error message.

Posted to: Trace Analyzer not working Show Thread | Post Reply

Re: Doubt accuracy of the parameter set count reported by DBSpohic

Bernie — Thu, 18 Mar 2010 11:12:03 +0000

I would like to thank the folks at DBSophic who worked with me to solve this problem. They are dedicated and capable people. The level of support I received was excellent. Taking into account this was for a product they give away, over a weekend, it was perhaps the best I support I have ever received. Thanks to all who helped and good luck with your product!

Posted to: Doubt accuracy of the parameter set count reported by DBSpohic Show Thread | Post Reply

Re: Doubt accuracy of the parameter set count reported by DBSpohic

Inbal — Tue, 16 Mar 2010 08:17:31 +0000

This was a known issue in V1.0, please install the latest version (V1.5, available since October 2009) and the issue should be solved.

Posted to: Doubt accuracy of the parameter set count reported by DBSpohic Show Thread | Post Reply

Doubt accuracy of the parameter set count reported by DBSpohic

Bernie — Sat, 13 Mar 2010 18:05:25 +0000

Hi there,

I have used and blogged about your trace analyzer tool (see the link in my sig) and just love it. Recently, I have come to doubt the accuracy of the batch parameter set count value. I have a trace file that has minimal data (for sql statement end event, I collect spid, duration, text data, reads, writes, and cpu). This image www.iperformax.com/downloads/dbs_trace_pic.JPG shows parameter set #8 as having a count of 4. However, when I open up the trace file with the sql server trace utility and search for the text of the batch, it finds just a single occurrence. What do you think? Shouldn't I find 4 instances of the query?

Here is the query text:

SELECT SUM(PLANNED_SIZE1),SUM(PLANNED_SIZE2),SUM(PLANNED_SIZE3),SUM(ACTUAL_SIZE1),SUM(ACTUAL_SIZE2),SUM(ACTUAL_SIZE3),SUM(PLANNED_PICKUP_SIZE1),SUM(PLANNED_PICKUP_SIZE2),SUM(PLANNED_PICKUP_SIZE3),SUM(ACTUAL_PICKUP_SIZE1),SUM(ACTUAL_PICKUP_SIZE2),SUM(ACTUAL_PICKUP_SIZE3),SUM(OVER_SIZE1),SUM(OVER_SIZE2),SUM(OVER_SIZE3),SUM(SHORT_SIZE1),SUM(SHORT_SIZE2),SUM(SHORT_SIZE3),SUM(DAMAGED_SIZE1),SUM(DAMAGED_SIZE2),SUM(DAMAGED_SIZE3) FROM TSDBA.DP_LINE_ITEM_EXTENDED_VIEW WHERE REGION_ID = 'DAL' AND ROUTE_DATE = {TS '2010-03-12 00:00:00'} AND ROUTE_ID = 'DEL_RBD_13' AND INTERNAL_STOP_ID = 150 AND ORDER_NUMBER = '1'

If you are interested, I have uploaded the .trc file to my web site. You can download it here www.iperformax.com/downloads/cpubusy.trc

Thanks very much for your time and a great free tool. I know how hard it can be.. I donate a LOT of my free time supporting the open source load testing tool OpenSTA.

All the best,

Bernie Velivis
Performax Inc (www.iPerformax.com)
Blog: http://iperformax.blogspot.com/
2 Colburn Lane, Hollis, NH 03049, (603) 860-7900

--------------------------------------------------------------------------------

Posted to: Doubt accuracy of the parameter set count reported by DBSpohic Show Thread | Post Reply

"Save as workbook" limitations

Inbal — Mon, 15 Feb 2010 09:31:35 +0000

A new feature in Qure v2.0 - the "save as workbook" option exports most of the analysis information to an excel report.
Please note that this option is only supported in Office 2003 or later.

Posted to: "Save as workbook" limitations Show Thread | Post Reply

Qure V2 multi database analysis limitations

Ami — Tue, 12 Jan 2010 13:52:17 +0000

Starting from version 2.0, Qure enables multi database analysis.

> Trial analysis is limited to a maximum of 2 databases per analysis.
> Full (licensed) analysis is limited to a maximum of 3 databases per analysis.

Posted to: Qure V2 multi database analysis limitations Show Thread | Post Reply

I'm not receiving the email with the download link after pressing the download button, what should I do?

Inbal — Sun, 15 Nov 2009 14:48:49 +0000

Spam filters may be preventing our email from reaching your mailbox.
Please add "webmaster@dbsophic.com" to your whitelist and press the download button again.
If you have logged out and do not know the password to your DBSophic account, please go to the Log in page and select "I've lost my password", and an email will be sent to you with a link which will enable you reset your password.
Then return to the Download page and download Qure again.
If the problem persists, please contact support@dbsophic.com and one of our experts will assist you.

Posted to: I'm not receiving the email with the download link after pressing the download button, what should I do? Show Thread | Post Reply

Re: Problem with installation

Inbal — Mon, 09 Nov 2009 12:25:09 +0000

Thanks for asking, Armine.
This problem can result either from a corrupt installation file, or a problem with the Windows installer version.
First, please download the installation file again by logging on to the DBSophic web site using the credentials you provided during registration, going to the download page and clicking the green “download” button.
If the installation fails again after this second download, please consider upgrading your windows installer version. Follow this link in order to upgrade to Windows Installer version 4.5.
If you still encounter problems with the installation, please contact DBSophic support, either by sending an email to support@dbsophic.com or by posting on this forum.

Posted to: Problem with installation Show Thread | Post Reply

Problem with installation

Armine — Mon, 09 Nov 2009 11:25:24 +0000

Hi,
When I try to install Qure, the windows installer raises an error saying it's not a valid windows installer file. What do I do?
Thanks.

Posted to: Problem with installation Show Thread | Post Reply

Re: Running a benchmark

Inbal — Sun, 04 Oct 2009 15:31:29 +0000

Please note that as of version 1.5 (released October 5th 2009), the benchmark is performed as an integral part of the analysis process

Posted to: Running a benchmark Show Thread | Post Reply

Re: Red error messages in Qure

Ami — Tue, 29 Sep 2009 15:35:01 +0000

Hi Tate and thanks for your inquiry.
The answer is no, it doesn’t. The red-colored lines display error messages returned from the database engine. For example, if the analysis uses a copy of the database from a point in time after the trace was started, it is likely that inserts or updates may violate a primary key and will fail. Therefore, it is recommended that you backup the database right before starting the trace for Qure’s analysis. Please note that the failed batches will have recommendations generated for them, but these recommendations will not be benchmarked.

Posted to: Red error messages in Qure Show Thread | Post Reply

Red error messages in Qure

Tate — Tue, 29 Sep 2009 15:14:54 +0000

While the benchmark process is running, I keep seeing red-colored lines with error messages like ‘primary key violation’. Does that mean there’s something wrong with my analysis?

Posted to: Red error messages in Qure Show Thread | Post Reply

Re: Deadlocks

Ami — Wed, 09 Sep 2009 11:58:31 +0000

During the analysis process, Qure executes the trace activities serially. Therefore, deadlocks or any lock contentions are not reproduced. However, by improving access paths and reducing the required number of reads, Qure’s recommendations have proved to be extremely efficient in reducing deadlocks.

Posted to: Deadlocks Show Thread | Post Reply

Deadlocks

Tommy — Wed, 09 Sep 2009 11:14:11 +0000

Can Qure help me solve recurring deadlocks?

Posted to: Deadlocks Show Thread | Post Reply

Re: Qure templates

Ami — Wed, 26 Aug 2009 15:02:01 +0000

The templates (Qure_For_SQL_2005.tdf and Qure_For_SQL_2008.tdf) are located in the 'TraceTemplates' folder in the Qure installation folder.
Additionally, you can download them directly from http://www.dbsophic.com/assets/support/QureTraceTemplates.zip
For more information about preparing the trace files, please see “Creating an efficient trace”.

Posted to: Qure templates Show Thread | Post Reply

Qure templates

SQLKing — Wed, 26 Aug 2009 13:44:19 +0000

Where do I get the Qure templates (for SQL 2005 and 2008)?

Posted to: Qure templates Show Thread | Post Reply

Re: Running a benchmark

Ami — Mon, 24 Aug 2009 12:44:25 +0000

Yes, which means you cannot use the same database to run a second analysis or benchmark. You’ll need a “fresh” copy of the database for these purposes. Please note that there is a detailed explanation of the benchmark process in the online help, under the “Benchmarking” chapter, in the “Viewing the benchmark process” subtitle.
Please note that in the upcoming version to be released in September the benchmark will be an integral part of the analysis process.

Posted to: Running a benchmark Show Thread | Post Reply

Running a benchmark

John — Mon, 24 Aug 2009 12:26:16 +0000

Will running a benchmark process implement all the recommendations on the copy of the database?

Posted to: Running a benchmark Show Thread | Post Reply

Re: Instructions for creating the Trace file

Ami — Wed, 12 Aug 2009 08:20:40 +0000

The only requirements (other then the requirements listed on the Trace Analyzer page) for the trace file is to include TextData, Duration, Logical Reads, Writes and CPU.

Posted to: Instructions for creating the Trace file Show Thread | Post Reply

Re: Selecting tables

Ami — Wed, 12 Aug 2009 08:16:08 +0000

Please note that the ‘tables list’ is intended for the trace tables and not the database tables.
This means we are looking for tables that were created by SQL Profiler using the “Save as table” option.
To show up in the list, the table should contain at least the following columns: Duration, Reads, Writes, CPU, TextData.
The more typical scenario is to use trace files and not tables. You can also mix the 2 trace formats.

Posted to: Selecting tables Show Thread | Post Reply

Instructions for creating the Trace file

Mike T — Wed, 12 Aug 2009 07:56:33 +0000

Are there any special requirements?

Posted to: Instructions for creating the Trace file Show Thread | Post Reply

Selecting tables

Mike T — Wed, 12 Aug 2009 07:45:08 +0000

When I try to select Trace Table I get the message: “Cannot see tables list”.
What needs to be done in order to be able to view/select the Tables List?

Posted to: Selecting tables Show Thread | Post Reply

Welcome to trace analyzer forum

Ami — Fri, 31 Jul 2009 06:18:45 +0000

Use this forum to post questions, feedback, bugs, feature requests and anything else regarding DBSophic Trace Analyzer!

Posted to: Welcome to trace analyzer forum Show Thread | Post Reply

Welcome to the DBSophic Support Forum

Ami — Sun, 29 Mar 2009 09:49:45 +0000

Hello,

My name is Ami Levin, I'm the CTO and co-founder of DBSophic.
I would like to welcome you to the DBSophic support forum.

In this forum room you can post questions, report bugs, request features, suggest improvements, discuss and share your feedback of Qure.
If you prefer not to use the public forum, see alternative support options on the Support page.

Privacy policy:
We will not release any member information unless under court order to do so.

General forum rules:

> Please post in English only.
> Be polite.
> Do not spam.
> Commercial content, commission, referral or affiliate links are not allowed.
> Do not cross-post. Cross-posting refers to posting new duplicate threads or posts, or the linking to threads or posts already published.
> Do not post any content that may encourage illegal intent.

Sincerely,

Ami Levin.

Posted to: Welcome to the DBSophic Support Forum Show Thread | Post Reply

SQL Server version and compatibility level support

Ami — Thu, 19 Mar 2009 11:30:59 +0000

Qure supports analyzing databases that use SQL Server 2005/2008 and have a compatibility level of 90 or 100.
If your database was migrated from SQL Server 2000 and its compatibility level was left at its previous setting of 80, you can try and change its compatibility level to 90.
Note that changing the compatibility level may fail, subject to the backward compatibility issues as published by Microsoft in the Books On Line article “SQL Server 2005 Database Engine Backward Compatibility”.

Posted to: SQL Server version and compatibility level support Show Thread | Post Reply

Database default collation support

Ami — Thu, 19 Mar 2009 11:29:58 +0000

Qure does not support analyzing databases with default case sensitive collations.
For more information about database collations, see the Books On Line article ”Collation Options and International Support”.

Posted to: Database default collation support Show Thread | Post Reply

T-SQL Parsing limitations

Ami — Thu, 19 Mar 2009 11:29:12 +0000

The Qure parser does not support parsing and analysis of the following types of queries:

Queries that use quoted identifiers
Qure does not support parsing queries that use quoted identifiers for object names such as “Table1” instead of [Table1].
This is an inherited limitation as Qure cannot know each process’s run time QUOTED_IDENTIFIER setting.

Dynamic queries in server side objects
A less common query writing technique involves dynamically creating the query as a string literal, and then using EXEC or sp_executesql to execute the string containing the query.
If this dynamic construction of the query is done at the application tier, Qure captures the final constructed query from the trace and then parses and analyzes it.
If the dynamic construction of the query is done in server side objects (such as stored procedures, triggers or functions), Qure is not be able to follow its construction or execution, and is therefore unable to analyze it.
The benchmark process however, executes and records batches that call the containing object, either directly or indirectly.

Extended procedures and CLR objects
Qure does not analyze OLE automation (extended procedures) and CLR objects in the database.
All referenced objects in the queries that use OLE or CLR objects are analyzed, except for the expressions or clauses containing references to either the OLE or CLR objects.

XML X-queries
Qure does not parse X-Query expressions and does not analyze any query that contains an X-Query expression.

Column name resolution limitations
The following column aliasing syntaxes are not supported:

> Aliases provided in an alias list.
For example, SELECT * FROM (<Sub query>) AS X (Alias1, Alias2…)

> Aliases in DELETE and UPDATE using the proprietary additional “FROM” SQL Server syntax.

> Aliases derived from sub queries that use set operators such as Union, Except and Intersect.
For example, SELECT Column1 FROM (<Sub query1> UNION <Sub query2>) AS X.

> Queries that contain any of the above syntax constructs are partially analyzed.

ODBC Date and time formats
ODBC specific Date & Time formats are not supported. For example, “{d '1998-02-23'}”

Posted to: T-SQL Parsing limitations Show Thread | Post Reply

SQL Server 2008 error logs analysis support

Ami — Thu, 19 Mar 2009 11:26:25 +0000

Collecting and analyzing error log information of SQL Server 2008 is not supported.

Posted to: SQL Server 2008 error logs analysis support Show Thread | Post Reply

Benchmark limitations

Ami — Thu, 19 Mar 2009 11:25:03 +0000

> Server side cursors and prepare/exec statements are partially supported in benchmark.
These include sp_cursoropen, sp_cursorprepare, sp_cursorprepexec, sp_prepare and sp_prepexec.
Qure extracts the query from the string literal of these commands and treats it as if it were a regular query.
The fetch commands are ignored and the benchmark will show the improvement gained only for the query.

> Schema column modifications may fail to apply in benchmark if any objects, other than indexes or statistics, are dependent upon the column.
You should manually drop any referencing objects in order to modify the column and recreate them after modifying the column properties.

Posted to: Benchmark limitations Show Thread | Post Reply

SQL Server client tools version compatibility

Ami — Thu, 19 Mar 2009 11:24:16 +0000

The version of the installed client tools must be the same or of a higher version than the version of the Profiler or server that generated the trace file.
A trace file generated by Profiler 2008 or a server side trace of SQL Server 2008 cannot be used if Qure is installed on a computer with SQL 2005 client tools.

Posted to: SQL Server client tools version compatibility Show Thread | Post Reply

Analysis data file 4GB size limit

Ami — Thu, 19 Mar 2009 11:23:02 +0000

Qure uses an internal repository (Analysis data file) for storing all analysis and benchmark information.
The repository has an embedded limit on its size which cannot exceed 4GB.
As a result, the trace analysis process stops analyzing the trace files if the repository reaches a size of 2 GB, in order to leave sufficient space for the benchmark and other data.
This means that only batches captured up to that time are analyzed.
Qure displays an informational message if this size limit is reached.
If your system contains an exceptionally large number of unique batches, the benchmark process may exhaust the available space of the repository.
In this rare case, the benchmark stops and an informational message is displayed.
Benchmark results for batches benchmarked successfully up to that point are available.

Posted to: Analysis data file 4GB size limit Show Thread | Post Reply

Analysis and benchmark time limits

Ami — Thu, 19 Mar 2009 11:22:09 +0000

Qure uses complex AI algorithms to analyze your database and its activities.
On complex databases with many objects and a large number of unique activities, the analysis and benchmark processes may take considerable time and resources to complete. To avoid the potentially very long duration of these processes (as much as a few days), the following limitation are imposed by Qure:

> The time limit for the trace analysis, which is typically the longest phase of the analysis process, is set to 20 hours.
After 20 hours, the trace analysis is stopped and only batches analyzed up to that point are used to generate recommendations.

> The benchmark uses a database connection with a timeout limit of 600 seconds.
This means that any batch that takes longer than 10 minutes to complete, either before or after the recommendations are applied, will fail the benchmark.

Posted to: Analysis and benchmark time limits Show Thread | Post Reply

Opening a report may take a long time

Ami — Thu, 19 Mar 2009 11:20:47 +0000

Depending on the analysis data file size, the number and complexity of the database objects, batches and recommendations, opening a report may take up to a few minutes.
In some cases, especially after the analysis data file is first copied to the local computer, the Windows operating system may indicate that Qure is “not responding”. It is caused by a known issue in SQLCE, click here for more details.
Ignore this warning and wait for Qure to complete loading the report.

Posted to: Opening a report may take a long time Show Thread | Post Reply

Unhandled exception in recommendations window

Ami — Thu, 19 Mar 2009 11:19:48 +0000

In some cases, selecting a large number of recommendations when the “script preview” pane is active, and attempting to change the status of these recommendations before the script generation completes may cause an unhandled exception message to appear.
Click ’OK’ to close the exception window and Qure should resume normal operation.

Posted to: Unhandled exception in recommendations window Show Thread | Post Reply

Can Qure help me optimize the performance of ANY database?

Ami — Thu, 19 Mar 2009 06:40:47 +0000

Qure can help you optimize the performance of almost any database.
Qure’s recommendations have proved to be extremely effective, yielding significant performance improvements, on “real world” production databases in ~ 8 out of 10 databases.
About 2 out of 10 databases, experienced less dramatic improvements and in very few cases, no or very little improvement was achieved.
Our tests and careful manual analysis of these databases revealed that the databases that gained the least from Qure’s recommendations were either highly optimized to begin with or that the main causes of the slow performance were rooted in issues such as design and coding techniques beyond Qure’s current scope of analysis.
We are constantly working on expanding and perfecting Qure’s analysis algorithms with every new version of Qure to increase the scope of Qure’s analysis.

Posted to: Can Qure help me optimize the performance of ANY database? Show Thread | Post Reply