logs archiveIRC Archive / Freenode / #oracle / 2010 / March / 4 / 1
MaL0_
hi
hi. I have 5 oracle application server with one app in cluster. We have about 80 httpd process on every server, but there's one moment that the number of httpd increases to 250 (MaxClient limit) and all becames to go slowly.
does somebody have any similar issue ? we don't know what's going on, and I'm really going to become mad
usn_work
hi folks
a question about execution plans
one of my databases does not display predicate informations any more. I can't imagine why, the same script does different things on two databases - but this started some days ago, before that time they completely bahaved similar
http://www.freelists.org/post/oracle-l/Execution-plan-Suddenly-no-predicate-information-displayed-anymore
this is the issue in detail
any idea?
brig
usn_work: try change format ->(BASIC, TYPICAL or ALL) in dbms_xplan.display_cursor
usn_work
select * from table(dbms_xplan.display_cursor(null,null,'ALL LAST')); ?
the three suggestions don't add predicate infos, ALL just gives me Query Block Name / Object Alias
But the syntax I use was already correct, since docs and the reference DB does tell us it works
there must be something in the DB that is different
phimic
hi all
hi usn_work
is it possible to use dataguard with a rac and a single oracle db?
brig
usn_work: after SELECT /*+ gather_plan_statistics */ table_name from user_tables; try select * from table(dbms_xplan.display_cursor(null,null,'IOSTATS'))
usn_work
the three suggestions don't add predicate infos, ALL just gives me Query Block Name / Object Alias
But the syntax I use was already correct, since docs and the reference DB does tell us it works
brig
brig
=(
usn_work: what do you think about it?
usn_work
About what?
I did not think about your suggestions, tried them ;)
brig, my problem is that two databases of the same release baheve differently. And it simply can't be the syntax
a) it works on the other box
b) its from the docs, and they tell me that my expectation regarding the predicate info is correct
c) n0othig I can imagine and change does take effect on the refusing system
(regarding the syntax)
statistics_level and timed_os_statistics and timed_osstatistics
statistics_level and timed_os_statistics and timed_statistics are identical between the systems
         

brig
usn_work: for me work http://pastebin.com/QvHM5div
usn_work
so it does for me, but only on the reference database
brig, thanks, but as I said, it's not syntax
MaL0_
please somebody administrate oracle application server ?
I could see lot of traffic between one apache (mod_oc4j) and an oc4j in other server and don't know why
I have mod_oc4j.conf with roundrobin:local
please I have lots of problems on a production environment and don't know if this could affect ...
Supra7
am installing Oracle server for the first time. I am told to go to /tmp/CVU.*/ and run a script called runfixup.sh . It is not there. Any idea why it is not in the directory it is supposed to be in? I am installing this on a RHEL server (v5)
usn_work
Are you using DBCA?
If not, why?
Supra7
usn_work: sorry.. what is DBCA?
hell
Supra7: ls -l /tmp/CV*
Supra7
sure... jut a sec
[root@plin-db2-str01 ~]# cd /tmp/CVU_11.2.0.1.0_oracle/
[root@plin-db2-str01 CVU_11.2.0.1.0_oracle]# ls
exectask exectask.sh fixup fixup.enable fixup.response scratch
hell
strange.
maybe bug in new OUI. You can fix your warnings/errors by hands)
Supra7
ok
I set the ulimit -n to 65536 but the install gui still does not recognize it
any reason for this hell?
hali
Supra7: did you edit /etc/security/limits.conf
oracle soft nofile 65536
oracle hard nofile 65536
Supra7
ok
thanks hali
usn_work
I'd recommend to read the readme included with the install set, and the best practice for RHEL
Supra7
usn_work: yeah reading it
now i get an error during install: File not found /home/oracle/app/oracle/product/11.2.0/dbhome_1/owb/external/oc4j_applications/applications/WFMLRSVCApp.ear
and many many other files not found... hitting continue button allows the install to continue though
balek
hi i have a question i try to see the list of job rman for waiting because i can't start backup rman i know it's queue job rman problem thanks for your help
usn_work
brig, interested in the solution?
brig
usn_work: of course
balek
i found a job rman but i don't know to delete this job rman
to the queue
usn_work : you have not any idea of that please ?
usn_work
one moment please
balek
ok thanks
         

usn_work
brig, http://forums.oracle.com/forums/thread.jspa?threadID=1038432&tstart=0&messageID=4141935#4141935
g2g, bye
rizzo
anyone know if disabling flashback database requires an instance bounce?
tijnie
rizzo, It's a bit of a guess, but I would say yes
zambaboo
sigh
guys my test db wont start up stating that the first datafile needs to be recovered. i have rman backups but when i try to recover until the earliest scn rman tells me that the datafile needs to be recovered to an scn older than what i have. what do i do?
tijnie
A lot of times the last few SCN are in the redologs
You can supply these redologs to the recovery sequence.
(That only helps ofcourse (If you are going to rollforward to the latest possible time.)
zambaboo
i actually dont care about the latest changes
i just need to get the damn thing back up
whatever means possible
RMAN-06025: no backup of log thread 1 seq 88 lowscn 335869448 found to restore
grand..
tijnie
Otherwise, it can't hurt to restore a few more archives then the one with the requested SCN. Recover about (lets' say 10 more).
zambaboo
the box with these backups left for the datacenter yesterday
tijnie
Oh yup I see.
Which archives do you still have?
zambaboo
i have from the 11th till the 24th (feb)
tijnie
And a database backup also from around that period?
zambaboo
oh you asked about the archives.. no, nothing on hand, but the backups i did with rman backed up the archivelogs as well
tijnie
I'm a little confused now. Lets ask a bit more explicit: "What backups do you have (db as well as archives)? Preferably these should be rman backups"
zambaboo
i have a full from feb 11 and incrementals since then
how can i restore from that full backup?
tijnie
You can try time clause "recover database until time <date time>"
use something like 12 feb in your time clause ..... see whether it works.
If it works you can try some date closer to today/
zambaboo
RMAN-06555: datafile 1 must be restored from backup created before 12-FEB-10
any way to tell rman to use that full backupset?
tijnie
thats odd. What happens if you try 11 feb in your recovery clause.
zambaboo
same!
tijnie
hmm : recover datbase from backupset <...>
zambaboo
what. number?
tijnie
I guess you allready found it but still have also a look at : http://download.oracle.com/docs/cd/B28359_01/backup.111/b28273/rcmsynta030.htm#i84135
You should look into your backuplogging to retrieve thebackupset number.
zambaboo
its 21
tijnie
I'm looking in some rman logging.
zambaboo
running restore datafile 1 validate
tijnie
You can try : recover database from tag ' ..... ';
tag-name is also in your logging.
zambaboo
yep thats the next thing to try
although looks like restore datafile 1 should just do it
tijnie
Well...it sounds promising. If your whole backup is of that same point in time, you maybe in for a working database.
zambaboo
hm. restore archivelog all;
tijnie
:-) there you go.
zambaboo
failed :)
interesting, restore datafile 1 aborts with a ORA-01013: user requested cancel of current operation
but if i run it again, it continues, channel ORA_DISK_1: reading from backup piece /opt/oracle/rman/backup/incremental/incr_backup_t710650901_s27_p1
(Action) crosses fingers AND toes.
and thats the full backup it is reading from
tijnie
I cross my fingers with you.
Problem with recoveries (imho) is that they are very importand, but that you almost never can practice. So evey time again you have to "try and error"
zambaboo
ive been learning rman on this test system
it would be embarrassing if i cant get it to work :)
tijnie
Well I learned on production systems.....then there is some pressure I can tell you :-)
zambaboo
ugh
geez i hope this does not take 40 minutes like it did for the whole database :/
ah crap its doing a fulll database restore
oh.. RMAN: full datafile restore
why so long then..
Froolap
will oracle 11g install on fedora 12 out of the box without having to patch the os or the like?
rizzo
patch the OS?
you'll have to install the necessary packages
there's a guide, one sec
Froolap: http://oracle-base.com/articles/11g/OracleDB11gR2InstallationOnFedora12.php
Froolap
Thanks much
tijnie
Froolap, also you have to tweak some system settings. (But they are all in the documentation rizzo gave I guess)
Froolap, Anyway...I'm looking at such an instal rigth now...so it will eventually work :-)
Froolap
cool, I might be able to manage it.
« prev next »