Re: (2) Onconfig Tweak Problem

From: "Obnoxio The Clown" <obnoxio@hotmail.com> > >From: wyn441 <wclindenNOwcSPAM@mbf-law.com.invalid> >> >> >>I've been tweaking my onconfig file to enhance performance and >> >>today I see an warning error in my online log. >> >>"Physical Log size 20000 is too small. Physical Log overflows >> >>may occur during peak activity. Recommended minimum Physical >>Log >> >>size is 20 times maximum concurrent user threads." >> >>Ok, using onstat -u I see I have: >> >> 258 active, 384 total, 274 maximum concurrent >> >>Assuming this is where I find my maximum concurrent thread >>info, >> >>that would mean my PHYSFILE entry should be 5480 (or 5k) Ah. This manual thing is wonderful, isn't it? :-) The formula in the manual is: PHYSFILE = (connections * max_log_pages_per_crit_sect * 4 * pagesize) / 1024 And in the same manual, it says that if you're not using R-Tree indexes, max_log_pages_per_crit_sect = 5. So, 274 * 5 * 4 * 2048 = 10960, which is *still* less than 20000, and even assuming allocation for all 384 (total) connections, 384 * 5 * 4 * 2048 = 15360, which is *also* less than 20000. I can therefore only assume that at some stage, even more users were connected, and that is when the message occurred. Or there's a bug. Or there's something missing in my interpretation. :0) >> >>Here is my current settings. >> >># Physical Log Configuration >> >>PHYSDBS rootdbs # Location (dbspace) of physical log >> >>PHYSFILE 20000 # Physical log file size (Kbytes)>> >> >> >>Now, nobody sucks at math more than me, so I could be wrong >> >>here, but isn't 20000(KBytes) larger than 5480 (bytes). >> >> >> >>Tell me where I err? ________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com

And Shouldn't we also keep in mind the number of dirty pages between checkpoints ? Manel Falcó On Sun, 30 Jul 2000 13:48:15 BST, "Obnoxio The Clown" <obnoxio@hotmail.com> wrote: > >From: "Obnoxio The Clown" <obnoxio@hotmail.com> >Ah. This manual thing is wonderful, isn't it? :-) > >The formula in the manual is: > >PHYSFILE = (connections * max_log_pages_per_crit_sect * 4 * pagesize) / 1024 > >And in the same manual, it says that if you're not using R-Tree indexes, >max_log_pages_per_crit_sect = 5. So, 274 * 5 * 4 * 2048 = 10960, which is >*still* less than 20000, and even assuming allocation for all 384 (total) >connections, 384 * 5 * 4 * 2048 = 15360, which is *also* less than 20000. I >can therefore only assume that at some stage, even more users were >connected, and that is when the message occurred. Or there's a bug. Or >there's something missing in my interpretation. :0) >