RE: Bufwaits and Foreground Writes

From: "Bernstein, Rick" <rbernste@alarismed.com> > >See imbedded responses and comments. > > > -----Original Message----- > > From: Obnoxio The Clown [mailto:obnoxio@hotmail.com] > > > > From: "Bernstein, Rick" <rbernste@alarismed.com> > > > > > >Despite our tuning efforts, we are encountering a > > significant number of > > >bufwaits and foreground writes > > >on our production SAP system. This system runs IDS > > 7.30.UC7XK on a 4-way > > >RS6000 R50 (AIX). > > >I attended Informix's Performance Tuning class last week in > > Menlo Park. > > >Their suggestions > > >included:. > > >1) Use different ONCONFIG at different times of the day. > > > Our system must be available to online users 24 hours > > a day, 6 days a > > >week. > > > Therefore, this suggestion is not realistic for us. > > >2) Reduce the number of LRUs to 4 * number of CPU VPs. > > > Will this really help? Or will it make matters worse? > > > > Utter bollocks. > > > > >3) Increase the LRU_MIN_DIRTY and LRU_MAX_DIRTY values to > > improve write > > >cache %. > > > > This seems equally dumb -- your checkpoints will probably just become > > longer. > > > (snip) > > >=>> 'onstat -c' command output follows: > (snip) > > > > >NETTYPE ipcshm,2,30,CPU # Override sqlhosts nettype> > parameters > > >NETTYPE soctcp,3,45,NET # Override sqlhosts> > nettype parameters > > > > That's not a lot of users, is it? > > For SAP the number of tcpip connections is based on the application >servers' > work processes, not end-users. Of course. Sorry. > > >NUMCPUVPS 4 # Number of user (cpu) vps> > >083098cl > > > > You *might* get better performance by adding more CPU VPS > > than physical > > CPUs. > > > We are already specifying 4 CPUs on a 4-way machine, despite > recommendations to set it to one less than the number of CPUs. > The processors are slow on this machine. Therefore, I am hesitant > to overload them. It's worth a try. :-) > > >BUFFERS 150000 # Maximum number of shared buffers> > > > This seems a bit low for *any* SAP site. Max out, baby, go > > for 768,000! :-) > > This machine only has 1.5 Gbytes of memory. Ah. OK. > Would not 3 Gbytes (768,000 * 4K) of buffers cause excessive paging? Yep. > I also need to reserve space for a small application server, Informix > virtual storage memory, etc. Any further increases in BUFFERS will > likely cause an unacceptable amount of OS paging. > > We plan to install a new server in a few months. At that time I > will increase the number of buffers. > > > > [SNIP] > > > > >Thank you for any assistance or insight which you can provide. > > > > onstat -l?> > Below is the requested output from onstat -l: > >Informix Dynamic Server Version 7.30.UC7XK -- On-Line -- Up 3 days 00:44:27 >-- 1457264 Kb >Physical Logging >Buffer bufused bufsize numpages numwrits pages/io > P-2 236 256 158253 653 242.35 You can probably increase your physical log buffer (PHYSBUFF) > phybegin physize phypos phyused %used > d00035 57287 4163 6475 11.30 >Logical Logging >Buffer bufused bufsize numrecs numpages numwrits recs/pages pages/io > L-3 4 8 3126345 186825 68707 16.7 2.7 ______________________________________________________ Get Your Private, Free Email at http://www.hotmail.com