How to Enable Write-Back Flash Cache in 11.2.3.2.0

Write-back Flash Cache= WBFC

By default the flash cache mode in 11.2.3.2.0 is write-through.

 The following mentioned steps would change the mode from write through to write-back.

All the changes would be executed using cellcli.

How to enable write-back flash cache (WBFC):

-      We will have to drop the existing Write-Through flash cache mode before doing any changes.

For example:

CellCLI> drop flashcache

-       Please Shut down the CellSrv by using the below command:

CellCLI> alter cell shutdown services cellsrv

-      Now we are ready to change the flash cache mode to Write-Back using the below command:

CellCLI> alter cell flashcacheMode = writeback

-       Please Start up CellSrv :

CellCLI> alter cell startup services cellsrv

-       The Flash cache needs a recreation for effect to take place. The below command would do the change:

CellCLI> create flashcache all 

How to rollback a patch using patchmgr in EXADATA

  

Can we rollback a patch applied using the utility patchmgr?

Yes we can, however only those patches can be rolled back which has successfully patched the EXADATA cells. 

Cells with incomplete/Failed patching cant be rolled back.

How to do:

1. Check the prerequisites using the below command:

./patchmgr -cells cell_group -rollback_check_prereq [-rolling]

2. Perform the rollback using the following command:

./patchmgr -cells cell_group -rollback [-rolling]

Change lock time in Exadata

In image 11.2.3.1.1, when entered a wrong password, we will have to wait 10 minutes before retry. How to change/remove this configuration?

Solution: 

1. Change the lock time in /etc/pam.d/sshd.

Log in as root and change the lock_time parameter in sshd file:

For example:

[root@mine-admin ~]# cat /etc/pam.d/sshd
#%PAM-1.0
auth       include      system-auth
auth       required     pam_tally2.so deny=50 onerr=fail lock_time=1 <Seconds>
.
.

2.  Please restart ssh for the change to take effect.

For example:

service sshd restart

Note: There might be a need to  change /etc/pam.d/login for KVM logins.

One system and Two system upgrade in EM 12c

One System Upgrade:

•           This approach is simpler.
•           This approach assumes the same host as the prior version of operation.
•           The repository components are upgraded in the same repository database.
•           This approach is suitable, when the number of agents is less.
•           Requires downtime.
•           OMA , OMS and Repository database are upgraded at once.

Two System Upgrade:

•  This approach is useful in more complex monitoring environments, where:
•   A different host from the prior OMS version of operation will be used.
•  The repository components will be upgraded in a different repository database copy.
•  Best suited for large agent counts and anywhere the unmonitored time must be minimized.
•   Minimal Downtime
•   The existing reports, metrics and jobs are migrated as well.