Rumours about what would come after 8i (internet) 10g (grid) were around, but now it is official, there will be an Oracle Enterprise Manager 12c
Read more : http://www.oracle.com/us/products/enterprise-manager/index.html
And in the blogosphere http://orana.info
emgrid 10.2.0.3 is not the second patchset of 10gR2, it is Oracle Enterprise Manager Grid Control Release 3.
Well, it is still called a patch set and must be applied to 10.2.0.1 or 10.2.0.2.
But it is a new release !
You can download the soft and the doc on otn
http://www.oracle.com/technology/software/products/oem/index.html
Out of the new features, you have a Linux pack for your Unbreakable Linux, Management pack for Siebel and Oracle Content Database, and also Linux Server Administration to manage your Linux (RHEL4 and SuSE9).
My customer wants to receive alerts if a database has not been backed up (either unsuccessful backup or no backup) for ages.
As the customer have Oracle Enterprise Manager Grid Control 10gR2, I started using User Defined Metric (UDM) yesterday. Thanks Troy for his comment at OEM Generic Service
I would do 2 checks, backup of datafiles and backup of redo logs.
Here the steps :
- Open a database target (LSC01)
- Click on User-Defined-Metrics
- Create
- Metric Name = Age of datafile backup
- Type = Number
- Output = Single Value
- SQL Query : the age in hour since the oldest checkpoint time of the newest backup
select (sysdate-min(t))*24 from
(
select max(b.CHECKPOINT_TIME) t
from v$backup_datafile b, v$tablespace ts, v$datafile f
where INCLUDED_IN_DATABASE_BACKUP='YES'
and f.file#=b.file#
and f.ts#=ts.ts#
group by f.file#
)
- Credentials : dbsnmp/*****
- Threshold Operator > Warning 24 Critical 48
- Repeat every 1 hour
- OK
Same for redologs, with a name of Age of redolog backup query of
select (sysdate-max(NEXT_TIME))*24 from v$BACKUP_REDOLOG
I am not going to do this for each instance, so I will create a monitoring template
- Setup
- Monitoring Templates
- Create
- Select the target (LSC01)
- Name = Age of last backup
- Metric Threshold : remove all metrics except Age of datafile backup and Age of redolog backup
- Policies : remove all policies from template
- OK
Start applying your new metric to your databases.
- Apply
- Add
- Select all your databases
- Continue
- dbsnmp/*** (if you have the same database password for dbsnmp on all databases, it is easier)
It is now possible to define alerts.
- Preferences
- Notification Rules
- Create
- Apply to specific targets : Add you productive databases group
- Deselect Availability Down
- Metric: Add : Show all: Check User defined metric : Select : Age of datafile backup , Age of redolog backup
- Severity : Critical and Clear
- Policy : None
- Method : Email
After a while, you can monitor the metric with a historical graphic
I want to monitor the output of one script in OEM.
Let’s say my script is that easy :
#!/bin/ksh
echo $RANDOM
So I go to
That’s all, your generic service is ready. Wait a bit, you will see this in your targets list. You can then click on perform in your “random” homepage and chose “custom metric 1″ to have a graph of your values ($RANDOM).
Happy Holidays !
For those who like free games about skiing with really good graphics, check this :
Ski Challenge 2007
go to http://www.tsr.ch/tsr/index.html?siteSect=860020
click Télécharger le jeu (about 41Mb)
I have been looking for this note for ages : Metalink note 295949.1
Now I know how to rename my targets in Grid Control! the trick is to stop the agent
AH/emctl stop agent
to remove the upload and state files/subdirectories
cd AH/sysman/emd
find state upload recv agntstmp.txt lastupld.xml protocol.ini -type f -exec rm {} ';'
rm -r state/*
col TARGET_TYPE for a17
col TARGET_NAME for a60
set lin 100 pages 0 emb on newp none head on autop on
select TARGET_TYPE,TARGET_NAME
from MGMT$TARGET
where TARGET_TYPE='oracle_emd'
order by TARGET_NAME;
var target_name varchar2(40)
prompt Enter the target_name to delete
exec :target_name:='&TARGET_NAME';if (:target_name is not null) then mgmt_admin.cleanup_agent(:target_name); end if
vi AG/sysman/emd/targets.xml ### Get the source of the traditional viAH/emctl start agentI am waiting for eomgc 10gR2 for AIX so I check the download page often. Instead of adding new plateforms, I notice they just removed Solaris !
Oracle Enterprise Manager 10g Grid Control Release 2 (10.2.0.1) for Solaris Operating System (SPARC)
google cache
Last friday I installed RAC on my Suse10. Today, I installed Grid Control.
Of course it is not supported.
The packages I neeeded additionaly are db1 and openmotif.
It did not work 100%. I finally started the iasconsole (emctl start iasconsole), and within the iasconsole, I restarted the failed elements.
I created the repository manually (sysman/admin/emdrep/bin/RepManager)
An unlocked user is a security problem. Currently, all my unlocked users have only CREATE SESSION as system privilege, evtl ALTER SESSION.
Except SYS, SYSTEM and DBSNMP
To minimize this security problem, I implemented the following strategy on my test system.
1) delete password file, set remote_login_passwordfile=NONE, O7_DICTIONARY_ACCESSIBILITY=FALSE
2) alter user SYSTEM account lock;
3a) in 10gR2 :
alter user dbsnmp quota 1T on sysaux;
create role secure_oem_role;
grant advisor, analyze any, analyze any dictionary, create job, create procedure, create session, create table, manage any queue, select any dictionary to secure_oem_role;
grant EXECUTE on “SYS”.”DBMS_AQ” to secure_oem_role;
grant EXECUTE on “SYS”.”DBMS_AQADM” to secure_oem_role;
grant EXECUTE on “SYS”.”DBMS_DRS” to secure_oem_role;
grant EXECUTE on “SYS”.”DBMS_MONITOR” to secure_oem_role;
grant EXECUTE on “SYS”.”DBMS_SERVER_ALERT” to secure_oem_role;
grant EXECUTE on “SYS”.”DBMS_SYSTEM” to secure_oem_role;
grant EXECUTE on “SYS”.”DBMS_WORKLOAD_REPOSITORY” to secure_oem_role;
exec SYS.DBMS_AQADM.GRANT_QUEUE_PRIVILEGE(‘DEQUEUE’, ‘ALERT_QUE’, ‘SECURE_OEM_ROLE’)
revoke EXECUTE ON “SYS”.”DBMS_SERVER_ALERT” from dbsnmp;
revoke EXECUTE ON “SYS”.”DBMS_SYSTEM” from dbsnmp;
revoke UNLIMITED TABLESPACE from dbsnmp;
revoke SELECT ANY DICTIONARY from dbsnmp;
revoke CREATE PROCEDURE from dbsnmp;
revoke CREATE TABLE from dbsnmp;
revoke OEM_MONITOR from dbsnmp;
grant secure_oem_role to dbsnmp;
3b) in other versions, you probably can remove more and grant less, I think only in 10g it is necessary to have “quota”. In my other databases, dbsnmp have 0 segments.
Check what system privileges are potentially dangerous to the system :
select path
from
(
select
grantee,
sys_connect_by_path(privilege, ‘:’)||’:'||grantee path
from (select grantee, privilege, 0 role from dba_sys_privs union all select grantee, granted_role, 1 role from dba_role_privs)
connect by privilege=prior grantee
start with role=0
)
where
grantee in (
select username from dba_users
where lock_date is null
and password != ‘EXTERNAL’
and username != ‘SYS’)
or grantee=’PUBLIC’
/
:ADVISOR:SECURE_OEM_ROLE:DBSNMP
:ANALYZE ANY:SECURE_OEM_ROLE:DBSNMP
:ANALYZE ANY DICTIONARY:SECURE_OEM_ROLE:DBSNMP
:CREATE JOB:SECURE_OEM_ROLE:DBSNMP
:CREATE PROCEDURE:SECURE_OEM_ROLE:DBSNMP
:CREATE SESSION:USER1
:CREATE SESSION:USER2
:CREATE SESSION:SECURE_OEM_ROLE:DBSNMP
:CREATE TABLE:SECURE_OEM_ROLE:DBSNMP
:MANAGE ANY QUEUE:SECURE_OEM_ROLE:DBSNMP
:SELECT ANY DICTIONARY:SECURE_OEM_ROLE:DBSNMP
it sounds better…
I just found in my RSS feed that a new agent has been released. 10.1.0.4 for AIX5L. Well, if you go to the OEM download page, you will see only 10.1.0.2 version for AIX, HPUX, Solaris Sparc. But by clicking on the link, for example
Grid Control (10.1.0.2)
for AIX5L Based Systems, you will find a link OEM 10g Grid Control Management Agent Release 1, from there I finally discovered that 10.1.0.4 agent download.
Well, I already had 10.1.0.4, but it was a 10.1.0.2 installation patched with 10.1.0.4. So I prefer installing directly 10.1.0.4 agent.
Not trivial to found, but hopefully, I subscribed to RSS feeds on Newest Downloads
Oracle Enterprise Manager Downloads: oem download
this is a great announcement!
We just received new PCs. Brand new with XP. Nice? Let’s see!
As usual, I started my Exceed and logged on my AIX server. I tried to start the Grid Control Engine (opmn). Hard luck. Failed to start OC4J instance 
The first problem is, when I installed the Grid, my old workstation DISPLAY name has been registered in opmn.xml!
The second problem, my server has no graphic card, and no X server running.
Actually, I do not need X on this server. But to let the Grid start, I will use Frame Buffer. In AIX, you install X11.vfb base package, then mkitab “xvfb:2:once:/usr/lpp/X11/bin/X -force -vfb :1 >tmp/x.txt 2>&1″ to start it at the next reboot. Start it now as well.
Then edit your opmn.xml file and set the DISPLAY variable to localhost:1
<variable id=”DISPLAY” value=”localhost:1″/>