Oracle Grid Infrastructure: change the interval for the Cluster Verification Utility (ora.cvu)

2017-08-14 Off By Markus Flechtner

Since Oracle 11.2.02. Oracle has integrated the Cluster Verification Utility (CVU) as a cluster resource. It‘s invoked every 6 hours and checks the health of your cluster. If you think, this is too often, you can change the interval. I‘ll show you how.

Let‘s take a look at our cluster first:

grid@ernie:~/ [-MGMTDB] crsctl status res ora.cvu -t
--------------------------------------------------------------------------------
Name           Target  State        Server                   State details
--------------------------------------------------------------------------------
Cluster Resources
--------------------------------------------------------------------------------
ora.cvu
      1        ONLINE  ONLINE       ernie                    STABLE

The runtime settings of ora.cvu are:

grid@ernie:~/ [-MGMTDB] crsctl status res ora.cvu -v
NAME=ora.cvu
TYPE=ora.cvu.type
LAST_SERVER=ernie
STATE=ONLINE on ernie
TARGET=ONLINE
CARDINALITY_ID=1
OXR_SECTION=0
RESTART_COUNT=0
FAILURE_COUNT=0
FAILURE_HISTORY=
ID=ora.cvu 1 1
INCARNATION=1
LAST_RESTART=08/13/2017 14:13:43
LAST_STATE_CHANGE=08/13/2017 14:13:43
STATE_DETAILS=
INTERNAL_STATE=STABLE
TARGET_SERVER=ernie
RESOURCE_GROUP=
INSTANCE_COUNT=1

The command „crsctl stat res ora.cvu -p“ will show us the configuration of ora.cvu:

grid@ernie:~/ [-MGMTDB] crsctl stat res ora.cvu -p
NAME=ora.cvu
TYPE=ora.cvu.type
ACL=owner:grid:rwx,pgrp:oinstall:rwx,other::r--
ACTIONS=
ACTION_SCRIPT=
ACTION_TIMEOUT=60
ACTIVE_PLACEMENT=0
AGENT_FILENAME=%CRS_HOME%/bin/orajagent
AUTO_START=restore
CARDINALITY=1
CHECK_INTERVAL=60
CHECK_RESULTS=7983364
CHECK_TIMEOUT=600
CLEAN_TIMEOUT=60
CRSHOME_SPACE_ALERT_STATE=OFF
CSS_CRITICAL=no
CV_DESTLOC=
DEGREE=1
DELETE_TIMEOUT=60
DESCRIPTION=Oracle CVU resource
ENABLED=1
FAILOVER_DELAY=0
FAILURE_INTERVAL=0
FAILURE_THRESHOLD=0
GEN_NEXT_CHECK_TIME=1502648023
GEN_RUNNING_NODE=ernie
HOSTING_MEMBERS=
IGNORE_TARGET_ON_FAILURE=no
INSTANCE_FAILOVER=1
INTERMEDIATE_TIMEOUT=0
LOAD=1
LOGGING_LEVEL=1
MODIFY_TIMEOUT=60
NEXT_CHECK_TIME=
NLS_LANG=
OFFLINE_CHECK_INTERVAL=0
PLACEMENT=restricted
RELOCATE_BY_DEPENDENCY=1
RELOCATE_KIND=offline
RESOURCE_GROUP=
RESTART_ATTEMPTS=5
RESTART_DELAY=0
RUN_INTERVAL=21600
SCRIPT_TIMEOUT=30
SERVER_CATEGORY=ora.hub.category
SERVER_POOLS=*
START_CONCURRENCY=0
START_DEPENDENCIES=hard(ora.net1.network) pullup(ora.net1.network)
START_TIMEOUT=0
STOP_CONCURRENCY=0
STOP_DEPENDENCIES=hard(intermediate:ora.net1.network)
STOP_TIMEOUT=0
TARGET_DEFAULT=default
TYPE_VERSION=1.1
UPTIME_THRESHOLD=1h
USER_WORKLOAD=no
USE_STICKINESS=0
USR_ORA_ENV=
WORKLOAD_CPU=0
WORKLOAD_CPU_CAP=0
WORKLOAD_MEMORY_MAX=0
WORKLOAD_MEMORY_TARGET=0

The CHECK_INTERVAL is the interval (in seconds) which is used by the clusterware to check the availability of the cvu daemon. The RUN_INTERVAL is the interval in which CVU is started by the clusterware. It is reported in seconds (21600 seconds = 360 minutes = 6 hours).

We want to change the RUN_INTERVAL. As with „ora.*“-resource we must not use the command „crsctl modify resource“ to change the settings. Trying to do so will lead to the following message

1 2	grid@ernie:~/ [-MGMTDB] crsctl modify resource ora.cvu -attr "RUN_INTERVAL=86400" CRS-4995: The command 'Modify resource' is invalid in crsctl. Use srvctl for this command.

We have to use „srvctl modify cvu“:

grid@ernie:~/ [-MGMTDB] srvctl modify cvu -help
 
Modifies the check interval for the CVU resource.
Usage: srvctl modify cvu [-checkinterval ] [-destloc ]
    -checkinterval  Interval in minutes between checks
    -destloc                 Directory for copying and executing CVU files
    -help                          Print usage

Let‘s check the current setting first:

grid@ernie:~/ [-MGMTDB] srvctl config cvu
CVU is configured to run once every 360 minutes
CVU is enabled.
CVU is individually enabled on nodes:
CVU is individually disabled on nodes:
 
grid@ernie:~/ [-MGMTDB]
grid@ernie:~/ [-MGMTDB] srvctl modify cvu -checkinterval 1440
grid@ernie:~/ [-MGMTDB]

Let‘s crosscheck our change:

grid@ernie:~/ [-MGMTDB] srvctl config cvu
CVU is configured to run once every 1440 minutes
CVU is enabled.
CVU is individually enabled on nodes:
CVU is individually disabled on nodes:

So: crsctl reports the interval as „RUN_INTERVAL“ in seconds but if we want to change it with srvctl we have to specify the „CHECKINTERVAL“ in minutes – I think that‘s a little bit confusing.

With Oracle 12.2, you can find the CVU results in $ORACLE_BASE//crsdata/@global/cvu/baseline/cvures

grid@ernie:/u00/app/oracle/crsdata/@global/cvu/baseline/cvures/ [-MGMTDB] ls -l
total 1844
-rw-r--r--. 1 grid oinstall  94487 Aug  7 10:59 cvucheckreport_08072017105516.txt
-rw-r--r--. 1 grid oinstall  94487 Aug  7 16:59 cvucheckreport_08072017165512.txt
-rw-r--r--. 1 grid oinstall 277325 Aug  7 16:59 cvucheckreport_08072017165512.zip
-rw-r--r--. 1 grid oinstall  94487 Aug  7 22:59 cvucheckreport_08072017225513.txt
-rw-r--r--. 1 grid oinstall 277326 Aug  7 22:59 cvucheckreport_08072017225513.zip
-rw-r--r--. 1 grid oinstall 278107 Aug 12 09:27 cvucheckreport_08122017092343.zip
-rw-r--r--. 1 grid oinstall  94487 Aug 13 14:18 cvucheckreport_08132017141412.txt
-rw-r--r--. 1 grid oinstall 278192 Aug 13 14:18 cvucheckreport_08132017141412.zip
-rw-r--r--. 1 grid oinstall  94487 Aug 13 17:43 cvucheckreport_08132017173952.txt
-rw-r--r--. 1 grid oinstall 278822 Aug 13 17:43 cvucheckreport_08132017173952.zip

(my test cluster is not running 24 x 7 so you will not find the 6 hours interval in the file listing above).

In Oracle Enterprise Manager Cloud Control 13c Release 2 you will find the CVU output via „Cluster -> Monitoring -> All Metrics -> Clusterware Status -> Cluster Verification Output“.

Links: