Re: Suddenly recompile package become so slow

dominica_at_gmail.com wrote:
> Hi Jeremy,
>
> No, I don't think so.
> This is really strange.
>
> Today, I found out SYSTEM tablespace is only has 4 Meg left.

It might be reasonable to make the ts autoextend. Did you check before you extended it?

>
> And I add another 200M to it , just for test.
> Since this is a development host, not production.
>
> You know how I come to realize about system tbsp.
> I saw there is a lot of wait event to wait for
> SYSTEM tablespace in a
> "db file sequential read" status.

This probably means it is looking for individual blocks and waiting for the disks. Is your system ts locally managed? Are there user objects in the system ts? What version is it? Do you have some extra tracing turned on? Are your other tablespaces locally managed? If dictionary managed, are your FET$ and UET$ tables really big/extended? What kind of disks and I/O do you have? What OS? How long has this db been around? Are there statistics on the system ts objects? Have you used resource advisors to check your shared pool? What version did you say?  Statspack?

>
> But this might NOT be the only problem in that machine.
> That is also a SQL*NET client wait for a long time.
>
> Just trying to debug.
> Since now, every single developers is complaint about the
> "long compilation of all the packages".
>
> I will keep update on this issue here.
> P.S. The interesting thing happen in this particular company
> is very interesting. I work in a lot of places before and
> never seen all these STRANGE problems. Good place to learn
> though.
>
>

jg

-- 
@home.com is bogus.
http://cryingwhileeating.com/