Temp-Contention
- From: io <io@xxxxxxxxx>
- Date: 22 May 2007 01:12:40 -0700
Hi,
we are running a load job with a parallel degree of 8 on a machine
with 4 processors each with two cores. Running statspack I've found
out, that we see average read time of 775.1ms whereas the average
write time is 3 ms on our temp tablespace. Doing roundabout 70,000
reads this seems to slow down the load process heavily.
Looking at the tablespace, that consists from 20 files I can see, that
the read activities are located on certain files...
Anyone any ideas how to further investigate or to solve this?
Best regards,
Ingo
.
- Follow-Ups:
- Re: Temp-Contention
- From: DA Morgan
- Re: Temp-Contention
- Prev by Date: Enterprise manager nls problem
- Next by Date: Re: two total from select
- Previous by thread: Enterprise manager nls problem
- Next by thread: Re: Temp-Contention
- Index(es):
Relevant Pages
|
Loading