logs archiveIRC Archive / Freenode / #oracle / 2010 / February / 22 / 1
Junior
yello ;)
TheBonsai
y0
idletask
Hello
knandan
Hi..
idletask
(Action) has found a critical flaw in dbvault - and it's documented
knandan
Hi gurus..
I am facing a problem in the mview refresh in Oracle..
I have described my problem here..
http://pastebin.com/m1bf19793
It would be very nice if someone guides me regarding the probme..
It would be very nice if someone guides me regarding the problem..
idletask
I don't use materialized views, so I can't help
And I'm far from being a guru
hali is
knandan
anybody??
Rudemeister
what
sry cant help on that
         

tijnie
knandan, what happens if you manually refresh the mviews?
I believe that only after a refresh the changes will apear in the mviews.
knandan
Hi tijnie : thanks for response..
i will just chk
i think it still doesnt sync itself..however, let me cross chek
tijnie
knandan, please do. A mview is (more or less) a snapshot.
I think you always have to refresh it manually (dbms_refresh)
knandan
tijnie: nothing happens...it still remains as it was..doesnt sync itself with the core DB tables..
there are two things here..
i have made a refresh group..
and add the mviews to the refresh group.
and the time interval is also specified while making the refresh group..
so, it should keep on refreshing the mviews at those periodic intervals..
and also we also specify the time interval while creating mview..
i am no DB expert..and i was given this script by someone to execute...
i am wndering why i have to specify the time interval in the refresh group as well as while creating the mviews..
and also if one of the tables is synching itself , why not the other two i.e. crs_web_target and crs_web_criteria..
as i have described here..
http://pastebin.com/m1bf19793
tijnie
Do you see problems in the alert_log?
what is the value of JOB_QUEUE_PROCESSES
knandan
i will just check..
tijnie: the value of JOB_QUEUE_processes is set to 10
tijnie
mmm that should be enough
No problems found in the alertlog?
knandan
i dont know how to see the alert log..infact i dont think i would have the rights to see the alertlog..
is there some problem in my command to create the mview..?
will it help, if i twek the command slightly ..?
like this?
http://pastebin.com/me8e98a
that wont work..
and the prev command too is fine..
as per this document
http://www.docstoc.com/docs/24966187/MATERIALIZED-VIEWS---Refreshing
tijnie
knandan, maybe you should ask someone with the proper permissions to have a look at the alert_log.
Unfortunately I'm not to familiar with mviews, so I can't help you to much further I'm afraid.
knandan
tijnie: thanks for your efforts...
i made a small change in creation of mview..
CREATE MATERIALIZED VIEW CRS_WEB_MATCH REFRESH FORCE AS SELECT * FROM CRS_WEB_MATCH@C1DEV.WORLD;
and it seems to be working..
but i dont know how efficient it is..
earlier it was :
CREATE MATERIALIZED VIEW CRS_WEB_HEADER REFRESH FAST START WITH SYSDATE NEXT SYSDATE + (1 / 1440) WITH PRIMARY KEY AS SELECT * FROM CRS_WEB_HEADER@C1DEV.WORLD;
so, i have replaced fast with force
Fast - only apply the data changes
Force - try a fast refresh, if that is not possible, do a complete refresh
as given here..
http://www.oracle.com/technology/products/oracle9i/daily/jul05.html
tijnie: your feedback on this..:-)
tijnie
knandan, I'm not to fond of force options
however...if it works...it works
still it makes me wonder why the fast option dit not work
If the fast didn't work it appears if oracle didn't see the changes at refresh time.
That is a bit strange I guess.
knandan
tijnie: but force option too would first try for fast refresh..
if it fails in that then it will go for complete..
tijnie
Anyway...now it works, but if you have some time left, I would investigate still some more on the fast option,
knandan
ok..first i will test this approach(force) and then test my application for this..then I will look dig again into the fast option..
right now, i have to release my application to the QAs ASAP..
Thanks tijnie ...:-)
tijnie
knandan, hahaha good luck :-)
idletask
Question
What does the system privilege "asminister sql tuning set" stand for?
stede
hi ... anyone knows a good way to accumulate stdandardeviations?
meins: is there an extra function in oracle for that?
*means
idletask
Ow
Your set of problems becomes more and more complicated with time :p
stede
like sum(stddev())
idletask
Working with time (heh) data is an entire discipline by itself, and really belongs to mathematics (and, I'd say, to the application side after you have aggregated what you want from the db)
         

stede
do it per function (power by 2 ...*n .. accumulate .../ N ... sqrt 2 isnt very fast)
idletask
Well, play on the strenghts of the db and have the rest computed by the application
stede
yeah .. but if there is the possibillity to compute things in the db i want to do that ;)
yep .. it seems those analytic functions should be better done by the application-side *g*
idletask
So, then, why in the db? :p
Look at it this way: the db is relational first and for all, aggregation of data is not its primary domain even though it _can_ do it fast
So, do simple aggregates at the db level, return the results to the app, and use whatever the app has, which is a lot, to do analytics
OLAP and friends are (money) traps
stede
yep .. thats the way i'm going to do this job ;) but jst see "STDDEV()"-function and want to know if oracle provides more
OLAP?
idletask
OnLine Analytical Processing
Analytics are calculations done on aggregated data, gathered by a relational database to begin with
Aggregations are done faster by the rdbms only because it has a closer understanding of the data (via statistics and all)
But analytics? It's better done elsewhere
(well, that's _my_ point of view)
stede
i agree absolutly! ... just thinking about an application-page with big amount of analytical data was called from N ppl at the same time ... not very good idea for the oracle server ;)
idletask
You have a solution for that, mind
Take an external indexing product to store the processed data
We use exalead FWIW
stede
the db works well for storing for me
i don't started this project, so i n going now for optimizing .. determine which thing should be done be othe db and which other stuff should be done by the application (php - visualisation in a browser)
ans i should type slower but correctly :D
*and!
tijnie
stede, I did not read back entirely, so maybe it's allready suggested : stddev()
Rudemeister
moarning
stede
tijnie: yea.. i suggested it ;) *g* it was a question about things like sum(stddev()) <-- what oracle calculates here .. i assume stddev1+stddev2+...
tijnie
exactly it sums up the standard deviations.
sums up == adds together :-)
stede
yea
idletask
(Action) bashes Oracle for the dbvault audit trail nonsense
hali
sounds more like a feature
to avoid admins bodging it :)
idletask
hali: eh? As a sysdba, I cannot touch the table anyway, and I won't have the dbvault passwords
But even one with the dv_owner role cannot purge the table, you have to bring the instance _down_, deactivate dbvault, purge the table, reactivate dbvault??
This is no feature
This is a bug, plain and simple
The security officer, or whatever the poor bloke's official role name is, _should_ have the possibility to purge that table
Who _cares_ what happened 5 years ago? If it hasn't been analyzed in this interval, tough
It's a bug with the security officer, Oracle shouldn't have to work around such bugs
You compromise the database availability for no good reason that I can see
And you open the door for a very, very easy DoS
sidh
Greetings gentlemen
i used a query for retrieving storage percent of tablespace, it uses dba_data_files, but i 'd like to see the available space in TEMP tablespace, and TEMPS tablespace is not referenced in dba_data_files, so where could i retrieve this information ?
V$TEMP_SPACE_HEADER <= i found it
idletask
Bah, anyway... It's not like the number of casual dbvault violations will be high anyway
It's just that I hate not having control (or the ability to delegate the control) over that
Logrotate doesn't require that you reboot the OS :/
usn_work
Hi, has somebody an idea about this thread? (Its SQL developer command wrapping madness) http://forums.oracle.com/forums/thread.jspa?threadID=1031847&tstart=0
tnx
RJarett
why arent you using the explain plan for name= select * from emp syntax?
idletask
sqldeveloper is perfectly able to extract data for a query by itself afaics
extract "plan data", that is
usn_work
I don't want an explain plan. I want a execution plan
idletask
Hmm, I'm afraid I'm not skilled enough to make a difference
usn_work
http://www.usn-it.de/index.php/2009/02/26/oracle-explain-or-execution-plans-guess-or-reality/
idletask
usn_work: I see... Just a thing though, the examples don't display well here
Probably because I'm using noscript :p
RJarett
the problem is he isnt passing a sql or profile id. so its just pulling the last thing on the stack
usn_work
yes
but it worked for ages
and it's damn useful for quick diagnosis
The question is: What changed?
And why?
idletask
usn_work: bewteen what version and what version of what (sqldev, the rdbms engine)
sqldev only?
usn_work
Only sqldev, it works from RDBMS 10g above
SQLDEV 1.5.* worked, 2.1 does not.
1.4.* worked as well
1.1.* did not
That's all I tried
RJarett
well 2.1 may be querying more when you run one. makes sense. what you see on the stack is like a paginator it seems
usn_work
I guess its the table() function
any old version acts the same, if you don't specify SET SERVEROUTPUT=OFF;
sinesio
why does my EM dbconsole not working in the Dataguard standby node ?
is there something I should know?
i'm running oracle DB 11gR2
EM dbconsole on the primary is working fine, but on standby it stoped working as soon as I configured dataguard
usn_work
Em needs an active database
the standby node is in mount state (usually)
sinesio
so if perform a switchover i should have EM working on the new primary
usn_work
in theory
But consider using GridControl if it HAS to be EM. I'd prefer simply command line and data guard broker
sinesio
I have dgbroker configured
but I still wanted grid control
the problem is that I was not able to configure it for oracle 11gR2
usn_work
grid control runs on a different host, usually
What you have is perhaps Enterprise Manager Database Control
sinesio
I know that grid control supports 11gr2 as a target
idletask
Hmm, I don't use dgbroker, I don't use dataguard, I don't use em...
(Action) is off scale
sinesio
right now I have EM DB control
because I couldnt configure Grid Control
usn_work
Install GC on a nude, native, virgin host
test it, try it, understand it
and then start integration
sinesio
yes, I tried
installed GC 10.2.0.1 then upgraded to 10.2.0.5
but when running the configuration script it failed
I could't finish the configuration
i've seen some known issues... but this is different
regarding the EM dbconsole not running on the standby.... my standby is in open state
not in mount
usn_work
huh? Do you have active dataguard?
josemoreira
hello
how can one change the PROCESSES configuration value?
in oracle express?
idletask
alter system set processes = <thevalue> scope = spfile;
shutdown
startup
sinesio
physical standby
idletask
Alternative way:
create pfile from spfile;
usn_work
sinesio, a physical standby is in moutn state as long as it's no active dataguard
« prev 1 2 3 next »