Thanks to those who have offered suggestions. This is not a new problem, but has worsened recently, perhaps because more users are logging off at the same time, although that's just a guess. Yes, I have run howmessy reports, and there is always room for some database maintenance, but I don't see how the condition of the database (number of inefficient pointers, number of secondaries, etc) would affect open/close times. It is clear that the problem is control block contention rather then a problem accessing database contents. I say this because it is triggered by several users logging off at the same time, and because most of the impeded processes are doing DBCLOSE or DBOPEN at the time. I know this by looking at stack traces in Glance while this is happening. The wait reason is "SIR". A SHOW LOCKS in DBUTIL does not indicate a locking issue. Yes, we do have DSEM and PREFETCH enabled.
By the way, this problem is usually exacerbated by the fact that users experiencing a "hung" session will often disconnect and log in again. User education will definitely be part of the solution.
John> Date: Tue, 15 Jan 2008 09:10:23 -0800> From: [log in to unmask]> Subject: Re: [HP3000-L] RE : [HP3000-L] DBCLOSE Causing Impedes> To: [log in to unmask]> > Hi John :)> Christian raises some great points and ideas....> > Since you have GLANCE, do you have any data about the conditions within> the database? Maybe a HowMessy run? My thot is that maybe the database> needs a bit of help internally?> > Art "just a thot :) " Bahrs> > =========================================> Art Bahrs, CISSP> [log in to unmask]>
_________________________________________________________________
Put your friends on the big screen with Windows Vista® + Windows Live™.
http://www.microsoft.com/windows/shop/specialoffers.mspx?ocid=TXT_TAGLM_CPC_MediaCtr_bigscreen_012008
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|