Re: Bloody ERP schema

Hi everybody, Well, yesterday we have tested the recommended NETTYPE parameter on a hpux machine as it could be set on this one of course. That is to say: #NETTYPE ipcshm,2,200,CPU # Configure poll thread(s) for nettype NETTYPE ipcshm,5,80,CPU # Configure poll thread(s) for nettype We have change the poll threads from 2 to 5 because we have 5 cpu associated to the database and in order to use all of theses ones and we have decrease the number of connection because we have not so much users using this type of connection. # NETTYPE soctcp,2,200,NET # Configure poll thread(s) for nettype NETTYPE soctcp,2,100,NET # Configure poll thread(s) fornettype We have decrease the number of connection because we have not so much users using this type of connection. I know it is not exactly the same than our production machine nonetheless we have divide the queries time by two which is not negligible. By the way, originally on this machine the CPU was and is always associated with ipcshm and the NET as and is always associated with soctcp. So I still have not test this particularity as mentionned by Mr. Art Kagel. Well, I will not forgot to test all the rest of parameters mentionned above as soon as they can. Nonetheless, they could be really promising. To conclude this post, I still would like to thanks you all for what you have done. Best Regards Oxow

I'll be interested to see if any of this clears up your problem. Please let us know. Personally, I'd be surprised. Config parameters rarely make *that* much difference. I suspect your problems lie elsewhere .... "Oxow" <oxow@yahoo.com> wrote in message news:d169ee4.0305240612.9781a8e@posting.google.com... > Hi everybody, > > Well, yesterday we have tested the recommended NETTYPE parameter on a > hpux machine as it could be set on this one of course. > That is to say: > > #NETTYPE ipcshm,2,200,CPU # Configure poll thread(s) for nettype > NETTYPE ipcshm,5,80,CPU # Configure poll thread(s) for nettype> > We have change the poll threads from 2 to 5 because we have 5 cpu > associated to the database and in order to use all of theses ones and > we have decrease the number of connection because we have not so much > users using this type of connection. > > # NETTYPE soctcp,2,200,NET # Configure poll thread(s) for > nettype > NETTYPE soctcp,2,100,NET # Configure poll thread(s) for> nettype > > We have decrease the number of connection because we have not so much > users using this type of connection. > > I know it is not exactly the same than our production machine > nonetheless we have divide the queries time by two which is not > negligible. > > By the way, originally on this machine the CPU was and is always > associated with ipcshm and the NET as and is always associated with > soctcp. So I still have not test this particularity as mentionned by > Mr. Art Kagel. > > > Well, I will not forgot to test all the rest of parameters mentionned > above as soon as they can. Nonetheless, they could be really > promising. > > To conclude this post, I still would like to thanks you all for what > you have done. > > Best Regards > Oxow