Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: Log file sequential read ??
spamdump_at_nospam.noway.nohow (Ed Stevens) wrote in message news:<3dda63b9.57518006_at_ausnews.austin.ibm.com>...
> Platform: Oracle 8.1.7.4 EE on Win2K
>
> A statspack report shows most of our wait event time (39.5 %) for 'log file
> sequential read'. A search of MetaLink turned up lots of hits for Rdb. The few
> other hits didn't really address this event, just had it mentioned in a list of
> events. Pretty much the same for a search of this ng's archives. OraPerf
> simply says 'The logfile header block was read. This could happen for example
> when a new log file is added or when a logfile is dumped.'
>
> So, with this event fingered as my 'heavy hitter', what course of action would
> you recommend?
I've never seen significant wait time on this event. In addition to adding or dumping logfiles, this event obviously occurs when you switch logs (v$system_event.total_waits should increment by the number of log members when you switch logs). If you don't use RAID or any disk management hardware/software, make sure your redo log files in one group are on separate disks. Make sure you don't switch logs too frequently. Create no more than 3 members per group. If you do have hardware mirrors, I suggest you only use one member per group. Some people may not like this but I've never had problems since I started to do this on our highly active Oracle Portal database almost 2 years ago.
Are you saying your redo logs are on RAID5?
Yong Huang Received on Fri Nov 22 2002 - 12:22:43 CST