Wednesday, December 3, 2014

ORA 700's make me grumpy ( well or at least confused )

Geez Louise I guess I coulda/shoulda known about this before now.

At some point the oracle software ( 11.2 ish ? 11.1 ish ? ) now starts getting worried and kind of grumpy.

You apparently can get ORA 700's under certain circumstances.  It's not a huge problem apparently ( yet ) when the software decides to let you know ... ( but it might become one later maybe ? ).

I saw this one on a new test vm I am setting up ( Database using ASM and also OEM ).

ORA 700 [kskvmstatact: excessive swapping observed]

So anyway I started doing some looking around at the memory config stuff after seeing that ( but that's a longer story ).

Sunday, November 23, 2014

Great Lakes Oracle Conference 2015 call for speakers now open

Calling all geeks speakers big data nerds and application whiz kids the Great Lakes Oracle Conference ( GLOC ) is open for abstracts/presentation proposals ( May 18-20 2015 ).

This is a growing high quality regional conference in Cleveland where we are hoping to pass 350 ( 400? ) in attendance this year.  We have Kerry Osborne and Steven Feuerstein doing keynotes and on Monday 1/2 day workshops before the conference kicks into high gear for the next two days.

Call for presentations information here: GLOC speakers

Please consider joining us in Cleveland and hope to see you there!

John Hurley @grumpyolddba and NEOOUG President

Sunday, November 9, 2014

rmoug 2015 presentation ready to go ... makes me ( not ) grumpy

All ready to roll with RMOUG 2015 Training days presentation "OS Truth, little white lies, and the Oracle Wait Interface".

Best of course to come to RMOUG 2015 Training days ... but this is link to pdf version of the presentation here: John Hurley RMOUG 2015

If you are out there ( very excited personally my first time at this conference ) please kick me and say hello.  Ok maybe skip the kicking part ...

handy ash queries to look at blocked sessions ( how many when for what event ) ...

Licensing may be required ... please check if applicable.

A query like this can be used to check how much blocking and what session it was ... ( so then you can drill into those blockers sessions ).  Probably can be done easily with some kind of rollup query?

Adjust SAMPLE_TIMEs in where clause below.

select ash_data.*, substr(sqlinfo.sql_text,1,70)
(SELECT to_char(ash.sample_time,'MM/DD/YYYY HH24:MI:SS') what_time,  count(*) sessions_blocked, ash.event, ash.blocking_session, ash.blocking_session_serial#, ash.sql_id, ash.sql_opname
WHERE ash.SAMPLE_TIME >= TO_DATE('01-NOV-2014 13:00', 'DD-MON-YYYY HH24:MI')
  and ash.sample_time <= to_date('08-NOV-2014 17:00', 'DD-MON-YYYY HH24:MI')
-- and ash.event like 'enq: TX - row%'
and blocking_session is not null
group by to_char(ash.sample_time,'MM/DD/YYYY HH24:MI:SS'), ash.event, ash.sql_id, ash.sql_opname, ash.blocking_session, ash.blocking_session_serial#
order by 1) ash_data,
v$sqlarea sqlinfo
where ash_data.sql_id = sqlinfo.sql_id
and sessions_blocked >= 1
order by what_time


For example once you have narrowed it down to something interesting looking ( who is blocked / what sql_id / what event ) ... you can use something like this.  I am now looking at any active history information on what the blockers were doing or waiting on.

select * from DBA_HIST_ACTIVE_SESS_HISTORY where ( session_id, session_serial# ) in
SELECT blocking_session, blocking_session_serial# FROM DBA_HIST_ACTIVE_SESS_HISTORY
and event like 'enq: TX - row%'
and sql_id = '0kbzgn17vbfc5' )
and SAMPLE_TIME > TO_DATE('06-NOV-2014 14:30', 'DD-MON-YYYY HH24:MI')
order by sample_time

Friday, November 7, 2014

setting pctfree to 0 for regular indexes ... ( not familiar with this theory ) ... help!?

One of my coworkers has this theory ( habit ) of creating new regular indexes on regular table columns.  Not compressed indexes or anything special.

He is currently believing ( back from somewhere / not sure of the origin ) that "if a table only gets inserts and deletes" that any indexes created on that table should be set with pctfree set to 0.

I do not agree.  In my opinion one should stick with the tried and true default values here.  I also am skeptical of this belief.

It seems to me like a very small amount of potential tangible gain in space usage while causing potential problems down the road.

Even if this theory is somehow true ( don't think so ) ... what happens in the future when something changes in the application and what one was initially told ( only inserts and deletes ) changes.

Anyone seen / heard /debunked this theory?  Any idea of where it comes from?

Thanks John   

Wednesday, October 22, 2014

getting re certified for 12c ( makes me grumpy ) ...

So let's be honest at the beginning here.

I have way too many professional certifications for any reasonable semi honest human being.  Along with two old Solaris admin certifications 7 and 8 I am ( besides an MBA degree ha ha ) OCP in 7.3 / 8.0 / 8i / 9i / 10g.  At one point a ways back I was teaching all the OCP curriculum for a local community college yes real classes with live instructors so it made sense to actually take the tests my students were taking.

But it's been a really long time now since I did the 10g new features class and did the 10g OCP upgrade exam.  I thought about doing the 11g upgrade at one point ( and attended the 11g new features class ) but never finished it.

Oracle has recently announced a change in policy where old certifications will start "aging out" / expiring.  I don't like this change at all but well it's probably time for me to do the 12c new features class and get 12c certified.

So now it's going to be somewhat of a nightmare to jump through all the hoops.  I have paper copies of ( some of anyhow ) the OCP's and text exams.

You used to take the tests at Sylvan Prometric sites but now there seems to be some other players involved in this whole thing.  Plus I don't have access to any of the old work emails.

This is probably going to make me a wee bit grumpy getting this all sorted out but maybe please please someone inside at oracle might bail me out?

Update on 10/24/2014: So false alarm this is getting sorted out much faster than I was imagining thanks to the oracle university support team.