Re: IDS Performance problem

Nothing you mention here specifically indicates a problem. If you know how to use dbaccess, run it and select a database. Use another session to take a look at what your backend process is doing. You will find that it is in exactly the state you describe below. This status will only change when you issue an sql command - in order to monitor it, make sure that the sql you issue takes a while to complete. You will see the thread status change to indicate that it is doing some work. Once the first set of records is returned to dbaccess you will be presented with a menu allowing you to step through the selected record set [assuming that there are enough records returned to require more than one screen]. Again if you check the status of the thread you will see that it is in 'cond wait' and will remain that way until you do something with the dbaccess session. Potentially you have a lot of client sessions that are connecting to the dbserver and spending a lot of time 'doing nothing'. About the only thing wrong with that is you are paying for user licenses that you may not require. Mark ----- Original Message ----- From: "jpp" <jean-philippe.planquart@wanadoo.fr> To: <informix-list@iiug.org> Sent: Thursday, October 02, 2003 17:27 Subject: IDS Performance problem > Hi, > > I have performance problems with on IDS 9.30 on hpux. When i look > current sessions with onstat -g sql somme are always with status cond > wait(netnorm) and wait more than 1000 times. What is meaning for this > status ? > > When i look all waiting sessions on threads with onstat -g con, i have > always more than 100 waiting sessions with cond wait netnorm. Does it > means my instance has not good parameters to get very acceptable > performances. And is it possible to see sql orders relating to these > threads. > > Thanks for your experience > sending to informix-list