RE: Re: Buffer busy waits are 10.96% of non-idle waits

2002-07-13 Thread Cary Millsap
One more detail: "buffer busy waits" is not an indication of competition for a block on disk. It indicates competition for a block IN MEMORY. More memory won't help, faster disks won't help. The cure is to make the competing processes not compete. The problem is usually a result of applications t

Re: Re: Buffer busy waits are 10.96% of non-idle waits

2002-07-12 Thread Greg Moore
> CLASS COUNT TIME > -- -- -- > data block 131525173 225446798 > will not > -- using bind variables instead of literals > -- seperating tables and indexes to diferent tabelspace > > solve my problem ? No and no. You have

Re: Re: Re: Buffer busy waits are 10.96% of non-idle waits

2002-07-11 Thread chaos
oraora oraora£¬ hi, i think use bind variable will give better performance, but it has nothing to do with the buffer busy wait.Using bind variable will solve the repeated parse and thus reduce library latch free wait. And seperate table and index won't help either. There is a greate

Re: Re: Buffer busy waits are 10.96% of non-idle waits

2002-07-11 Thread oraora oraora
Hi Kavitha, querying v$waitstat gives me the o/p below. CLASS COUNT TIME -- -- -- data block 131525173 225446798 sort block 0 0 save undo block 0 0 segment header 4968