RE: Logical Log Sizing issues

I like to keep my logical logs at a smaller number, around 20 or so. When I do a `onstat -l` I like to see real quick the status of all the log files by seeing them all on the screen at once. I compensate for the smaller number of logs by making them real big. And I have a cron script that runs every hour to force the next log file so that I can be assured that the most I'll lose from a disk crash is an hour's worth or work. The script also checks that there is no long transaction going on and that the logs are getting backed up and freed. I wouldn't want to run myself out of useable logs. Just my preference. Don. >-----Original Message----- >From: David Williams [mailto:djw@smooth1.demon.co.uk] > > >In article <76gvhe$bji$1@taliesin.netcom.net.uk>, Neil Truby ><ntruby@netcomuk.co.uk> writes >>insufficient disk space. If I were you I would STRONGLY advise you to stop >>wasting further time on pondering the issue, and add a large amount of >>logical logs, say 100MBytes. >> > > Agreed, I normally go for 1000 250K logical logs + physical log of > 200Mb. Remember logical logs get written to disk once full and not > longer needed for rollback, hence smaller logs = less on vulnerable > disk storage. > > Remember online allows for up to 32767 logical logs!! > >>If your cause with your management would be strengthened by the considered >>written (and expensive!) opinion of a reputable consultant, please drop me a >>line! >> > And me! >>Neil Truby > >David Williams

Simpson, Don wrote: > > I like to keep my logical logs at a smaller number, around 20 or so. When I > do a `onstat -l` I like to see real quick the status of all the log files by > seeing them all on the screen at once. I compensate for the smaller number > of logs by making them real big. And I have a cron script that runs every > hour to force the next log file so that I can be assured that the most I'll > lose from a disk crash is an hour's worth or work. The script also checks > that there is no long transaction going on and that the logs are getting > backed up and freed. I wouldn't want to run myself out of useable logs. > > Just my preference. Preference aside, if you are forcing the logs to change every hour then those REALLY BIG log files are only being partially used and you might as well make them smaller. Only the backed up ones will be available to save you from long transactions and then only if you recognize the situation in time to stop that cron job from moving you into the last log file. I say go with more smaller log files and get a higher res monitor so you can run 66 line xterms like I do. Art S. Kagel