Viewing 15 posts - 1 through 15 (of 21 total)
Doesn't the column set returned by the SELECT * include the column rowDiffHash ?
In which case the hash is non-deterministic - in the sense that repeating the UPDATE multiple times...
November 25, 2013 at 7:03 am
Here is my guess as to what is happening. It's only a guess since we don't have full information.
The classic reason for a deadlock is that 2 processes try to...
March 20, 2013 at 4:50 am
We use ODBC and SQL Server for our entire application. There is definitely a problem with intermittent loss of connection resulting in "Communication Link Failure" - we see this...
August 12, 2010 at 3:15 am
I'm a little puzzled by this.
There seem to be two points.
Firstly, we have a long-running SELECT which must not prevent modifications to the tables being queried. In practice, shared locks...
June 28, 2010 at 11:16 am
Jonathan,
thanks for the links. Bart Duncan's explanation of the problem is excellent. I wish I'd seen it before !
You say 'This isn't what I would call a design flaw.'...
July 15, 2008 at 4:39 am
Jonathan,
you are right to some extent. The index I am using - designed to pick out a few hundred rows from hundreds of thousands - is the correct index; usually...
July 14, 2008 at 10:57 am
OK, I have attached a ZIP containing the DDL for all tables and indexes, plus execution plans and statistics for the two SELECT statements. These were measured this afternoon on...
July 14, 2008 at 9:48 am
There are many examples of the SELECT/UPDATE deadlock. Here are a few:
SELECT CJ.JobID,
CJ.NeedsVerification,
JC.Name,
...
July 14, 2008 at 7:47 am
Yes, I think SNAPSHOT ISOLATION would solve the problem - it is, after all, the way Oracle does it. Unfortunately it's not available in SQL Server 2000. When and if...
July 11, 2008 at 11:24 am
Thanks for 2 prompt replies.
My point about having the SELECT in its own transaction was that I wasn't doing a SELECT in a transaction where a previous UPDATE, say, had...
July 11, 2008 at 9:55 am
With the greatest respect - I disagree.
Admittedly I am using SQL Server 2000 so perhaps this is something that has since been resolved. But I work on a complex real-time...
July 11, 2008 at 8:50 am
I've been using ODBC for over 10 years, and never seen a problem. Most of the applications we are currently shipping open 2 or 3 simultaneous connection to SQL Server,...
January 15, 2007 at 10:42 am
Hang on a minute. What are you trying to do here ? You have a query which is essentially "SELECT xxx FROM UMF_SUM_MATCHLOG WHERE LOAD_ID = :1" (ignoring the joins)....
April 10, 2006 at 10:48 am
1. A view is nothing more than a way of packaging a SELECT statement in a convenient re-usable form. Querying a view is no different from running the equivalent SQL...
March 29, 2006 at 2:49 am
Kenneth is right: if both connections locked Table 1 before Table 2, the problem would go away. In fact, there is no other way to fix deadlocks.
If you really can't...
February 9, 2006 at 8:26 am
Viewing 15 posts - 1 through 15 (of 21 total)