Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> Re: row cache lock

Re: row cache lock

From: Terry Sutton <terrysutton_at_usa.net>
Date: Sat, 26 Feb 2005 11:48:25 -0800
Message-ID: <080b01c51c3c$24ef5f90$6401a8c0@TerrySutton>


If the state is WAITING and seconds_in_wait is > 10000, that session is not consuming CPU, it is waiting to get the CPU.

BTW, some people consider it rude to sign your messages "New DBA" rather than your actual name. Why don't you want to use your name?

--Terry

Thanks Jonathan,

For most of the sessions, the state is WAITING and seconds_in_wait is >10000. And yes these requests were started 6 hours or so back.

For the sessions with state = WAITED KNOWN TIME, i see seconds_in_wait = a large value (>10000), and WAIT_TIME = 301. So does this mean that the process has been consuming CPU for more than 3 hours without waiting on anything?

ps and top at the O/s level seem to confirm that it is running on 100% CPU all the time, but is it possible for a process to keep consuming CPUs for hours without waiting on anything?

And this is happening with many processes.

Naveen

>
> Have another look at these sessions with
> event = 'row cache lock'.
>
> If the state is not 'WAITING' then they are
> not waiting, they are using (or trying to use)
> CPU.
>
> If the status is 'WAITED KNOWN TIME',
> then (to within about 3 seconds) the amount
> of time they have been using CPU is
> seconds_in_wait minus wait_time
> (the latter is in hundredths of seconds, so
> divide by 100 before subtracting).
>
>
> Regards
>
> Jonathan Lewis
>
> http://www.jlcomp.demon.co.uk/faq/ind_faq.html
> The Co-operative Oracle Users' FAQ
>
> http://www.jlcomp.demon.co.uk/seminar.html
> Public Appearances - schedule updated Jan 21st 2005
>
>
>
>
>
>
> ----- Original Message -----
> From: "New DBA" <new_dba_on_the_block_at_yahoo.com>
> To: <oracle-l_at_freelists.org>
> Sent: Saturday, February 26, 2005 1:58 PM
> Subject: row cache lock
>
>
> Hi,
>
> First let me express my sincere thanks to everyone
> in
> this list who has answered my questions in the past
> and help me learn and understand oracle better.
>
> So here is my problem (9.2.0.5 on HP-UX):
>
> While investigating long running requests, I found
> out
> that most of them were waiting on "row cache lock"
> since a long time.
>
> P1 was 21, and I checked v$rowcache and it returned
> 2
> rows:
>
> dc_table_scns
> dc_partition_scns
>
> The current SQL for most of these sessions was
> COMMIT.
>
> Please explain me what do these 2 row caches mean
> and
> how to find out who is currently holding them so
> that
> I can investigate further. I searched metalink but
> couldn't find any information.
>
> Moreover, I saw one odd thing while looking at
> v$session_wait
>
> I saw there were many rows with STATE=WAITED KNOWN
> TIME and seconds_in_wait a very big value (>10000).
> Moreover, seconds_in_wait kept increasing.
>
> Meanwhile, when I looked at the process for these
> sessions at the O/S level, they were consuming 100%
> CPU.
>
> Does this situation mean that last wait was
> seconds_in_wait back and ever since then it has not
> waited on anything?
>
> Please help me understand the above 2 problems.
>
> Regards
> New DBA
>
>
>
> __________________________________
> Do you Yahoo!?
> Yahoo! Mail - Helps protect you from nasty viruses.
> http://promotions.yahoo.com/new_mail
> --
> http://www.freelists.org/webpage/oracle-l
>
>



Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
--
http://www.freelists.org/webpage/oracle-l


--
http://www.freelists.org/webpage/oracle-l
Received on Sat Feb 26 2005 - 14:51:34 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US