Re: multiple unable to extend - 1 disk sort - what am I missing?

Hi Tom.

Have a closer look at which segment it was trying to extend. What the error message usually means is that it's trying to extend a regular segment (table, index, etc.), but is going about it in a non-intuitive way ...

Under most circumstances, Oracle extends by creating a temporary segment and then switching it to the appropriate type in the dictionary when the extent map in the segment header has been successfully updated. So, it LOOKS like it's trying to extend temp segments at first glance, but it's really trying to extend something else...

HTH, Dave Haas

"Tom" <zuestra_at_hotmail.com> wrote in message news:ee86ats0gi3ohddip6ca4kq30itm5cp0op_at_4ax.com...
> There was about an hour period where the alert log for an 8.1.6 database
> showed unable to extent TEMP by 256 about 25 times. (800MB of temporary
 files)
> I ran a statspack report on a snapshot one hour before that, and 2 hours
 after,
> and it only reported one disk sort. At the time there was a 1,000 log on
 stress
> test going on. (no batch), so I find it odd that one disk sort could a)
 account
> for such a huge sort without any large batch or bad reporting query etc.,
 and b)
> cause multiple unable to extent messages. Wouldn't that one disk sort
 have just
> failed and quit the one time? Any insight as to what could be going on
 here or
> what else to look at in the report..
>
> Thanks in Advance.
>
>
Received on Sun Mar 04 2001 - 23:50:45 CST