So exports don’t contain indexes, so they are going to be much smaller than the entire database.
I have a site with 15GB of indexes, 12GB of data and the export file is only 6.2GB. if I use compress on the export file, it comes down to 515MB!!!
So, a 27GB enterprise class database when exported and compressed comes down to 515MB!
You can’t tell me that the future of computing power won’t have a database like this running in memory off a computer the size of a thumb drive!
I know that this is old skool – but I’ve not had time to re-learn dpump, so I’m putting on the glomesh top and nike high tops (and some oakley frogskins) and going old skool on my data refresh knowledge.
You might bang on about a few extra things that dpump might do, but imp and exp is pretty rad (word rad is from the same era as imp and exp).
I’m also making a big thing about global search and replace, because I can never remember the command.
exp PD910/PASSWORD@instance PARFILE=expPD910.par
This file contains the following:
TABLES=('F980021', 'F95622', 'F95623', 'F95624', 'F95620', 'F95621', 'F952420', 'F95600', 'F98710', 'F982400', 'F98305V', 'F98306', 'F95625', 'F95626', 'F95627', 'F95628', 'F969861', 'F9698710', 'F9698712', 'F980011', 'F983051', 'F983052', 'F952400', 'F952411', 'F98711', 'F98712', 'F98713', 'F98720', 'F98740', 'F98741', 'F98743', 'F98745', 'F98750', 'F98751', 'F98752', 'F98753', 'F98760', 'F98761', 'F98762', 'F98770', 'F98950', 'F98950BK', 'F98950D', 'F989998', 'F989999') LOG=exp910.log FILE=export910_b4golive.dmp
This creates the dumpfile with the data and structures (FILE=) and also a log file (LOG=)
cool, now I want to import this data into PP910 – yes I’m refreshing PP central objects with PD central objects.
So I drop all of the indexes in PP910 and truncate all of the tables –simple.
Then I IMP the data with the following (note that I don’t want any of the full packages)
TABLES=('F980021', 'F95622', 'F95623', 'F95624', 'F95620', 'F95621', 'F952420', 'F95600', 'F98710', 'F982400', 'F98305V', 'F98306', 'F95625', 'F95626', 'F95627', 'F95628', 'F969861', 'F9698710', 'F9698712', 'F980011', 'F983051', 'F983052', 'F952400', 'F952411', 'F98711', 'F98712', 'F98713', 'F98720', 'F98740', 'F98741', 'F98743', 'F98745', 'F98750', 'F98751', 'F98752', 'F98753', 'F98760', 'F98761', 'F98762', 'F98770', 'F98950', 'F98950BK', 'F98950D', 'F989998', 'F989999') LOG=impPP910.log FILE=export910_b4golive.dmp ROWS=Y FEEDBACK=10000 IGNORE=Y INDEXES=N
Coolio – see how I did not need to change the tablespace owners or anything (yet!!!). This will insert all of the data. But now I have to generate the indexes. Note that I get a . every time 10000 rows are imported to a table – nice and nerdy oracle. You need IGNORE=Y to avoid it failing because the tables already exist (I truncated, did not drop).
So, I use an imdexfile
TABLES=('F980021', 'F95622', 'F95623', 'F95624', 'F95620', 'F95621', 'F952420', 'F95600', 'F98710', 'F982400', 'F98305V', 'F98306', 'F95625', 'F95626', 'F95627', 'F95628', 'F969861', 'F9698710', 'F9698712', 'F980011', 'F983051', 'F983052', 'F952400', 'F952411', 'F98711', 'F98712', 'F98713', 'F98720', 'F98740', 'F98741', 'F98743', 'F98745', 'F98750', 'F98751', 'F98752', 'F98753', 'F98760', 'F98761', 'F98762', 'F98770', 'F98950', 'F98950BK', 'F98950D', 'F989998', 'F989999') LOG=impPP910DDL.log FILE=export910_b4golive.dmp ROWS=N INDEXES=Y INDEXFILE=createIndexes.sql
Note that an index file will prevent things hitting the database, but will write it to a file. The above example with write the “CREATE INDEX” DDL to a file called createIndexes.sql
I then use vi to change the owner and the tablespaces with:
vi global search and replace (a command I always forget)
:%s/PD910/PP910/g
And then I wang in some parallel action:
:%s/NOLOGGING/PARALLEL 4 NOLOGGING/g
Woot, these indexes are going to fly!
I then execute the script and I have my indexes in the correct tablespace in the correct schema.
What I found is that when I use nmstart, the values that I put into web.xml are not being picked up.
background:
So, in web.xml I have added:
<session-config>
<session-timeout>60</session-timeout>
</session-config>
This is great, timeout working when I start from either weblogic console or E1! Wow, things are working so nicely. But people are telling me (replace t with Y yelling)
So, what is needed
D:\Oracle\Middleware\user_projects\domains\E1_Apps\servers\JWB01P_PDCMSAU_94\data\nodemanager
#Server startup properties
#Fri Apr 04 16:30:07 EST 2014
Arguments=-XX\:MaxPermSize\=256m -Djavax.xml.rpc.ServiceFactory\=oracle.j2ee.ws.client.ServiceFactoryImpl -Xms32m -Xmx1536m
SSLArguments=-Dweblogic.security.SSL.ignoreHostnameVerification\=false -Dweblogic.ReverseDNSAllowed\=false
RestartMax=2
RestartDelaySeconds=0
RestartInterval=3600
ClassPath=D\:\\Oracle\\Middleware\\wlserver_10.3\\server\\lib\\weblogic.jar;D\:\\jde_home\\SCFHA\\targets\\JWB01P_WLS\\config\\jdbc\\E1_Apps\\JWB01P_PDCMSAU_94\\jt400.jar
AdminURL=http\://10.38.16.10\:7001
AutoRestart=true
AutoKillIfFailed=false
weblogic.httpd.session.timeoutSecs=3600
So now, when nmstart is being used, my new timeout and new JVM sizes are being used.
What a great OS, shame to see it go. It hosts all my VM’s that I use for connecting to clients – lean and mean.
I might have to start running linux guests.
If your life depended on it, do you really and truly know the answer?
I know this is a pretty dumb question, but… I was asked to schedule a job for 12 midnight and did not really know if I should be selecting AM or PM in the “AM or PM indicator”.
Google has informed me that midnight is AM and midday is PM – thanks Google, no more midday reboots!
Best practice for starting weblogic processes:
You are going to use weblogic.WLST scripting. This is available in interactive or batch mode. If you are having issues, I recommend using interactive mode for debugging.
You can access the console by setting your domainEnv (as per below):
D:\Oracle\Middleware\user_projects\domains\E1_Apps\bin\setDomainEnv.cmd
I like to check my java version
java –version
Then start the command interface
java weblogic.WLST
(you can exit with exit() )
Generate some secure connection files, but remember to firstly change the default username and password for logging into the node manager:
This was very helpful:
Make sure you log into: http://localhost:7001/console/login/LoginForm.jsp (this is the default URL)
Lock and edit
Activate changes, they should work immediately. See that I’ve set them the same as weblogic
Generate userconfigfile & userkeyfile
Then you’ll be able to issue the connect statement from weblogic.WLST
So, generate the userconfigfile and the userkeyfile – so passwords just are not hanging about!
java weblogic.Admin -adminurl t3://adminserverl:port -username <adminusername> -password <adminpassword> -userconfigfile userconfig -userkeyfile userkey –STOREUSERCONFIG
d:\Oracle\Middleware\user_projects\domains\E1_Apps>java weblogic.Admin –adminurl t3://localhost:7003 -username weblogic -password xxxx -userconfigfile user config -userkeyfile userkey –STOREUSERCONFIG
Copy these files to a dir of choice. I then use 2 files, a bat file and a command file to pass into weblogic.WLST
Note that I have my security files and also the cmd and py commands.
startJDE.bat
@rem Script to start
call D:\Oracle\Middleware\user_projects\domains\E1_Apps\bin\setDomainEnv.cmd
java weblogic.WLST D:\StartWeb\StartJDE.py > d:\startweb\StartJDE.log
@exit
StartJDE.py
print 'starting the script .... '
redirect('d:\StartWeb\WLSTLogs.txt')
startNodeManager(verbose='true',NodeManagerHome='D:/Oracle/Middleware/wlserver_10.3/common/nodemanager',ListenPort='5556', ListenAddress='' );
nmConnect(userConfigFile='d:\StartWeb\userconfigNM.secure',userKeyFile='d:\StartWeb\userkeyNM.secure',domainName='E1_Apps', port='5556', nmType='ssl')
nmStart('AdminServer');
nmStart('JWB02P_PDCMSAU_94');
nmStart('JWB02P_PDCMSNZ_95');
nmStart('JWB02P_PDFIN_96');
nmStart('JWB02P_PDFIN_97');
exit()
Then schedule the script to run as a domain account (same as server manager) in the task scheduler. Therefore all of your processes will start as the correct user and SM will be able to communicate with them.
I could not for the life of my synchronise any jas.ini or jdbj.ini files to one of my production web clusters via server manager. Everything else was working fine. This is for WLS 10.3.5 and 9.1.3.3
I got the following in my server_manager agent logs:
WARNING: Stdout: weblogic.Deployer invoked with options: -adminurl t3://AUBDC00-JWB01P.au.ad.westfield.com:7001 -userconfigfile D:\\jde_home\\SCFHA\\targets\\JWB01P_WLS\\config\\owlConfig.secure -userkeyfile D:\\jde_home\\SCFHA\\targets\\JWB01P_WLS\\config\\owlKey.secure -name JWB01P_PDCMSAU_94 -redeploy webclient.war/WEB-INF/classes<02/04/2014 10:55:17 AM EST> <Info> <J2EE Deployment SPI> <BEA-260121> <Initiating redeploy operation for application, JWB01P_PDCMSAU_94 [archive: null], to configured targets.> [Deployer:149163]The domain edit lock is owned by another session in non-exclusive mode - this deployment operation requires exclusive access to the edit lock and hence cannot proceed. If you are using "Automatically Aquire Lock and Activate Changes" in the console, then the lock will expire shortly so retry this operation. Exitcode: 1
02/04/2014 10:55:22 AM com.jdedwards.mgmt.targets.owl.OWLJ2EEServer updateApplicationFiles
WARNING: Stderr:
02/04/2014 10:55:22 AM com.jdedwards.mgmt.targets.owl.OWLJ2EEServer updateApplicationFiles
WARNING: Update application files. Exception: UPDATE FAILED: Redeploy of application 'JWB01P_PDCMSAU_94' using location 'webclient.war/WEB-INF/classes'.
02/04/2014 10:55:22 AM com.jdedwards.mgmt.targets.webserver.WebServer synchronizeConfiguration
WARNING: Could not update remote configuration files for instance JWB01P_PDCMSAU_94
This is usually the result when one of the application servers of a cluster not running.
Make sure all servers are running and issue the sychronization again. Check logs for root cause.
Exception invoking method updateApplicationFiles
02/04/2014 10:55:22 AM com.jdedwards.mgmt.targets.ManagedTarget recordHistoricalEvent
FINE: Recording historical event 'stateChanged' for instance 'JWB01P_PDCMSAU_94'.
I got the following through SM – classic
Could not update remote configuration files for instance JWB01P_PDCMSAU_94 This is usually the result when one of the application servers of a cluster not running. Make sure all servers are running and issue the sychronization again. Check logs for root cause. Exception invoking method updateApplicationFiles
Failure during invoke:invoke(, , synchronizeConfiguration) javax.management.MBeanException : Could not update remote configuration files for instance JWB01P_PDCMSAU_94 This is usually the result when one of the application servers of a cluster not running. Make sure all servers are running and issue the sychronization again. Check logs for root cause. Exception invoking method updateApplicationFiles
This is an example of an error message being too helpful, as it’s making assumptions about the problem, I have no cluster. Also this is NOT helping, because the problem is more obvious!!! How about the message says “you’ve locked the weblogic config you dill, how about you unlock it via your weblogic console so that I can update my files”.
Wow, sometimes blogging reveals some embarrassing situations.
Once released – update worked as it should.
