Shulou(Shulou.com)06/01 Report--

Some time ago, there was a problem, that is, when rebooting a RAC node, it was suddenly reported that the size of the recovery area was insufficient. I checked that there should be no problem, it should have been modified to 500G, and there should be more than 200G of remaining space, so how to disappear at once, and deleted a lot of files inside, and synchronized the control files with the crosscheck in RMAN, it still didn't work. Later, I checked and found that the node forgot the original modified parameters after rebooting and directly became the initial parameters. no wonder there is not enough space for the initial parameters.

After resetting the parameters, the node gets up, but there is also a problem! EM can not enter, after checking the status of EM, decided to restart EM,EMCTL STOP DBCONSOLE no problem, but EMCTL START DBCONSOLE has a problem, it has always been failed.

Looking at the EM log didn't see anything wrong, and decided to rebuild the EM.

Use the command: emca-deconfig dbcontrol db-repos drop-cluster to delete EM, enter the database name, service name, port number, cluster name, SYS password, SYSMAN password and other information one after another, DBCONTROL deleted, but delete the database hanging there motionless, waiting for most of the day also had to stop.

It was awkward at this time, and I don't know if EM was deleted, but when I rebuilt, I was prompted that the database already exists, so it should not be deleted, or at least it has not been deleted.

Look up a command to delete the database on the Internet, and execute it on the two nodes of the RAC, / sysman/admin/emdrep/bin/RepManager-action drop, and both nodes continue to hang there!

Later, I checked and decided that since I didn't delete it clean, I would delete it thoroughly.

First of all, check whether there are any residual related directories in the directory, and check whether there are OC4J_DBConsole__ files under the / oc4j/j2ee/ directory. As a result, there is no residue in node one, but there is a problem of sid as node 1sid in node 2. Ferociously delete (remove)!

Delete emca-deconfig dbcontrol db-repos drop again, but still report an error. It looks like the monitoring did not start, and it feels that the password entered has a problem, and it cannot be rebuilt. It still reports the error that already exists in the database. There is nothing wrong with checking the monitor, but there is still something wrong with EM itself.

Look up the information again and see a paragraph: "reconfigure DBCONSOLE, you need to enter the DBSNMP password, but any password will display an error and need to be modified in advance!" it feels like the source of all the problems!

Change the password of DBSNMP in advance when rebuilding: sql > alter user dbsnmp identified by xxx

Then manually delete some of the objects that created EM in the database:

SQL > DROP public synonym SETEMVIEWUSERCONTEXT

Synonyms have been deleted.

SQL > DROP PUBLIC SYNONYM MGMT_TARGET_BLACKOUTS

Synonyms have been deleted.

SQL > drop role MGMT_USER

The role was deleted.

SQL > DROP USER SYSMAN CASCADE

User has been deleted.

SQL > DROP USER MGMT_VIEW CASCADE

User has been deleted.

There are still some problems when deleting SYSMAN. I made a mistake when deleting the Times, saying that I could not delete the currently connected users. I checked it with the V$session view and found that there were really three sysman connections, which may have been connected during reconstruction. in order to prevent accidents, lock up the sysman users first, kill three processes, and delete them in a while!

This time, I should delete the feeling and use the command emca-config dbcontrol db-repos create-cluster to rebuild EM, but the result is ferocious. I actually said that there was something wrong with my sentence. I checked it with-help, and I didn't feel anything wrong. Is there any problem?

At this time, when you use DBCA to open the configuration database product, the original EM option becomes optional, then there is no problem, since the command line says I have a mistake, I use the graphical interface to fill in the past, wait for the structure!

But the result once again told me not to take it for granted. At the end of creating the database, I popped up an error that job_queue_processes should be greater than or equal to 1. What's going on? check the parameter and find that the parameter is 0. Modify the generation: alteer system set job_queue_processes=5 scope=both sid='*'

After the modification, the command is deleted again for security reasons: emca-deconfig dbcontrol db-repos drop-cluster.

This time, the deletion is very fast. Use DBCA to choose to create EM again. About half an hour later, the EM is created correctly.

After entering the EM interface, I found that the last backup job created through EM has not been executed in the past few days, and the job is gone, so it really needs to be reconfigured!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.