On Tue, 22 Jul 2003 16:09:36 +1200, Murray Wood (IList)
<ifxmaillist@quanta.co.nz> wrote:
I will check on the indexes. That's a good place to start.
I did update the stats, but basically what I have done with all my 4gl code
is putting anti-lock while loops around OPENs and Fetches. It is a simple
WHILE statement around the open that will allow up to 500 attempts to
perform the operation. This has always solved my locking issues but I hate
the fact that I have to do that. I hate that the database doesn't
automatically retry. It gets the lock and crashes. Upping the wait time of
a set lock mode never seemed to help me. I need it to stop attempting to
get the record, sleep 1 second and then I try again. I've seen vendor code
do this to.
Anyway I will check to make sure these are hitting the index.
>
> Are the sqls that crash using the correct index to find data from the
> database? (explain plan).
> Update stats?
> Increase the duration of the set lock mode to wait.
>
> Neither of those messages necessarily indicate deadlocks. Why do you
> think
> there are deadlocks?
>
> MW
>
>> -----Original Message-----
>> From: owner-informix-list@iiug.org
>> [mailto:owner-informix-list@iiug.org]On Behalf Of Jason
>> Sent: Tuesday, 22 July 2003 11:54 a.m.
>> To: informix-list@iiug.org
>> Subject: Deadlock question
>>
>>
>> Broad question.
>>
>> Are there any settings in the onconfig, that if not properly
>> set, could
>> lead to an increase in deadlock issues (showing a LOT of
>> -243's and -154's)
>> .
>>
>> I just upgraded to 9.30.FC3 and I am seeing an awful lot of
>> dead locks...
>>
>> Just wondering.
>>
>>
>>
>> --
>> Using M2, Opera's revolutionary e-mail client:
>> http://www.opera.com/m2/
>
> sending to informix-list
>
--
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/