New leafmisscount Option in Flex Cluster 12c

Posted in: Technical Track

In the Oracle RAC previous version we have worked with the option “misscount” of crsctl, which applies for all the servers because there was only one kind of servers, but in the new version of Oracle (12c) we have a new kind of RAC configuration, the “Flex Cluster”. With Flex Cluster we have two kinds of servers “Hub nodes” and “Leaf Nodes”, the Hub Nodes have the same concept of the normal nodes in the previous versions, however the Leaf Nodes are different. I will let Oracle Documentation define these servers:

Hub Nodes are similar to Oracle Grid Infrastructure nodes in an Oracle Clusterware standard Cluster configuration: they are tightly connected, and have direct access to shared storage. In an Oracle Flex Cluster configuration, shared storage can be provisioned to Leaf Nodes independent of the Oracle Grid Infrastructure.

Leaf Nodes are different from standard Oracle Grid Infrastructure nodes, in that they do not require direct access to shared storage, but instead request data through Hub Nodes. Hub Nodes can run in an Oracle Flex Cluster configuration without having any Leaf Nodes as cluster member nodes, but Leaf Nodes must be members of a cluster that includes at least one Hub Node.

Reference: https://docs.oracle.com/database/121/CWADD/bigcluster.htm#CWADD92560

With new kind of servers, we also have a new option to configure the maximum time that a Leaf node can be tolerated with heartbeat issues. This crsctl option is “leafmisscount” and it’s exactly the same than the “misscount” for Hub Nodes, but this is for Leaf Nodes.

Let’s see how does it work:

 My Flex Cluster

Configuration

[root@dg1 ~]# crsctl check cluster -all

**************************************************************

dg1:

CRS-4537: Cluster Ready Services is online

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

**************************************************************

dg2:

CRS-4537: Cluster Ready Services is online

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

**************************************************************

dg3:

CRS-4537: Cluster Ready Services is online

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

**************************************************************

[root@dg1 ~]#

 

[root@dg1 ~]# crsctl get node role status -all

Node ‘dg1’ active role is ‘hub’

Node ‘dg2’ active role is ‘hub’

Node ‘dg3’ active role is ‘leaf’

[root@dg1 ~]#

Checking the current value:

[grid@dg1 ~]$ crsctl get css leafmisscount

CRS-4678: Successful get leafmisscount 30 for Cluster Synchronization Services.

[grid@dg1 ~]$

How does it work with 30 secs:

[grid@dg2 ~]$ date

Fri Nov 14 16:03:54 EST 2014

–Stop connectivity between Leaf Node
and Hub Nodes

[root@dg3 ~]# date

Fri Nov 14 16:04:01 EST 2014

[root@dg3 ~]# crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4535: Cannot communicate with Cluster Ready Services

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

[root@dg3 ~]# date

Fri Nov 14 16:04:29 EST 2014

[root@dg3 ~]# crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4535: Cannot communicate with Cluster Ready Services

CRS-4530: Communications failure contacting Cluster Synchronization Services daemon

CRS-4534: Cannot communicate with Event Manager

[root@dg3 ~]#

Now let’s increase the value to 600 sec:

[grid@dg1 ~]$ crsctl set css leafmisscount 600

CRS-4684: Successful set of parameter leafmisscount to 600 for Cluster Synchronization Services.

[grid@dg1 ~]$

Let’s confirm the value:

[grid@dg2 ~]$ crsctl get css leafmisscount

CRS-4678: Successful get leafmisscount 600 for Cluster Synchronization Services.

[grid@dg2 ~]$

The maximum value we can specify is 600:

[grid@dg1 ~]$ crsctl set css leafmisscount 3600

CRS-1671: The value for parameter leafmisscount is outside the allowed range of 1 to 600

[grid@dg1 ~]$

Let’s see how does it work with 600:

[root@dg3 ~]# date

Fri Nov 14 16:19:54 EST 2014

–Stop connectivity between leaf
node and hub nodes

[root@dg3 ~]# date

Fri Nov 14 16:24:01 EST 2014

[root@dg3 ~]# crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4537: Cluster Ready Services is online

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

[root@dg3 ~]# date

Fri Nov 14 16:25:18 EST 2014

[root@dg3 ~]# crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4537: Cluster Ready Services is online

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

[root@dg3 ~]# date

Fri Nov 14 16:27:30 EST 2014

[root@dg3 ~]# crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4537: Cluster Ready Services is online

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

[root@dg3 ~]# date

Fri Nov 14 16:28:08 EST 2014

[root@dg3 ~]# crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4537: Cluster Ready Services is online

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

[root@dg3 ~]# date

Fri Nov 14 16:29:04 EST 2014

[root@dg3 ~]# crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4537: Cluster Ready Services is online

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

[root@dg3 ~]# date

Fri Nov 14 16:29:50 EST 2014

[root@dg3 ~]# crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4535: Cannot communicate with Cluster Ready Services

CRS-4529: Cluster Synchronization Services is online

CRS-4533: Event Manager is online

[root@dg3 ~]# date

Fri Nov 14 16:30:10 EST 2014

[root@dg3 ~]# crsctl check crs

CRS-4638: Oracle High Availability Services is online

CRS-4535: Cannot communicate with Cluster Ready Services

CRS-4530: Communications failure contacting Cluster Synchronization Services daemon

CRS-4534: Cannot communicate with Event Manager

[root@dg3 ~]#

email

Author

Want to talk with an expert? Schedule a call with our team to get the conversation started.

About the Author

Deiby Gomez is rumored to be the youngest person in the world designated as an Oracle ACE Director, and the first in his country Guatemala at the age of 23. A tenacious and competitive individual, he regularly sets professional goals for himself and uses them to stay motivated. Deiby’s current and former colleagues can always depend on him to help resolve issues quickly—which he credits to his endless studying. When he isn’t working he can be found traveling or playing chess.

1 Comment. Leave new

John R. Hilton
November 24, 2014 2:19 am

New leafmisscount Option in Flex Cluster 12c brings light to the server management engineers, as this consumes much time from the engineers to resolve the issue. you can get more tips on server management from https://iserversupport.com/ blog.

Reply

Leave a Reply

Your email address will not be published. Required fields are marked *