Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> Re: MS SQL 7.0 HANGS
Please try to use OPTION (MAXDOP 1)
This helps you to use a single processor.
Sometimes parallelism does not give advantages on SQL7...
Think also that when using parallelism the indexes and joins are performed differently allocating a lot of memory and scanning a lot of tables.
The best way to check this query is to look into the execution plan for a "TOP 1" query; then the execution plan for the query without TOP.
"An Unhappy Microsoft Customer" <unhappy_at_microsoft.com> wrote in message
news:930607826.300.90_at_news.remarQ.com...
> Issue--
> We are experiencing apparent server hangs. Microsoft has traced the
> problem to a complex query that is being run that uses parallelism and the
> result set is being processed slowly (one row every 500 ms or so) The
total
> wall time that the query is open could be in excess of 30 minutes.
>
> Facts--
> MS SQL 7.0 allocates various resources, including memory for running
> queries. This amount of memory is not user configurable and is based on
> total memory in the server.
> Complex queries, especially where the query plan indicates
parallelism,
> can use up a substantial amount of these resources.
> If a complex query result is open without leaving sufficient resources
> for other queries MS SQL 7.0 will not be able to process additional
queries.
> The server will appear "hung" to users until the complex query result set
is
> processed and closed.
> Microsoft does not believe this is a flaw in MS SQL 7.0. We strongly
> believe this is a serious bug.
> These complex queries were not a problem under MS SQL 6.5 because of
> it's simpler query engine
> Microsoft claims the problem is "poorly written queries". We believe
no
> user query, no matter how poorly written, should be able to "hang" an
> enterprise class database server (if that's what you consider MS SQL
> Server). We would expect results like this from MS Access, not from MS
SQL
> server.
>
> Discussion--
> Is this a bug?
>
>
>
Received on Thu Jul 01 2004 - 02:38:55 CDT
![]() |
![]() |