In PRIMECLUSTER Wizard for NAS, it is monitored to the NetApp Filer-Cluster accessible by the ping / ping6 command. When the ping / ping6 command doesn't respond (It is not possible to access it), PRIMECLUSTER Wizard for NAS executes failover of the NetApp Filer-Cluster. In the following, it explains the method of the tuning of the parameter of Netapp resource.
Turning of the control facility for the NetApp Filer-Cluster
It is necessary to consider monitoring between the cluster node and Switching-HUB, and monitoring between the cluster node and the NetApp Filer device to decide the turning of the control facility for the NetApp Filer-Cluster.
The GLS resource monitors it between the cluster node and Switching-HUB. It should be set that failover of the NetApp Filer-Cluster doesn't execute it until the switching of the NIC of GLS is completed.
In a word, it is necessary to meet the requirement of following.
Monitoring Interval of GLS detector (a) + Switching time of the NIC with GLS (b) < Starting time of the NetApp Filer-Cluster failover (X)
(a) + (b) = 10 (sec) + 30(sec) or 60(sec) < X (sec)
Example
* Switching time of the NIC with GLS
case1 | not monitoring between HUB and HUB | about 30 sec |
case2 | monitoring between HUB and HUB | about 60 sec |
See
Refer to "PRIMECLUSTER Global Link Services Installation and Administration Guide."
After 300(recommended value) seconds pass, the RemoteFileSystems resource to monitor between the cluster node and the NetApp Filer device becomes faulted resource if the trouble is detected.
Therefore, the value in which the failure detection time of NetApp Filer-Cluster and the failover time of NetApp Filer-Cluster are added should be within at 300 seconds.
In a word, it becomes like the next expression.
Starting time of the NetApp Filer-Cluster failover (X) <= RemoteFileSystems failure time (300 sec) - time of the NetApp Filer-Cluster failover
Therefore, it becomes the following.
About 40sec. or about 70sec. < X < 300 sec - time of the NetApp Filer-Cluster failover (1)
The NetApp Filer cluster control facility will be initiated as follows:
As shown in the above figure, the time required for completing the NAS cluster control facility (1) is (m).
When executing takeover with the range of (l)
The NAS cluster control facility will be initiated being unable to access NAS devices because of NIC switching, so this will disable takeover processing.
When executing takeover with the range of (n)
RemoteFileSystems resources will time out and enter faulted before the NAS cluster control facility is completed, so this will disable takeover processing.
Turning of the Netapp resource setting for the ping / ping6 command
It is necessary to complete the ping / ping6 processing within the range of (m).
Default value is following:
A | Count of the ping / ping6 command for NetApp Filer-Cluster | 10 count |
B | Timeout time of the ping / ping6 command for NetApp Filer-Cluster | 5 sec |
C | Execution interval of the ping / ping6 command for NetApp Filer-Cluster | 5 sec |
The NetApp Filer-Cluster failover beginning time in default can be calculated from the following expression.
( B + C ) * ( A - 1 ) + B = (5+5) * 9 + 5 = 95 sec.
The NetApp Filer-Cluster failover beginning time in default is 95 sec.
Note
The required time for failover changes by the environment.
