how to modify the VCS HostMonitor Agent on Storage Foundation High Availability (SFHA) 6.2 (stop, start, what to monitor, threshold critical/warning).

[해리]

https://www.veritas.com/support/en_US/article.000024636

Problem

There may be a requirement to stop, start and/or control the HostMonitor feature of Veritas Cluster Server (VCS) which allows reporting on the system CPU, Memory and Swap utilisation and also provide notifications about the same setting the various thresholds.

This article will provide the information about how to modify the VCS HostMonitor Agent on Storage Foundation High Availability (SFHA) 6.2 (stop, start, what to monitor, threshold critical/warning).

Error Message

The Following are messages notifying the usage of various system resources based on the thresholds configured for the HostMonitor agent.

2015/02/10 11:23:42 VCS WARNING V-16-1-50086 Mem usage on node1hostname is 81%

2015/02/10 11:23:42 VCS WARNING V-16-1-50086 CPU usage on node1hostname is 81%

2015/02/10 11:23:42 VCS WARNING V-16-1-50086 Swap usage on node1hostname is 81%

Solution

# haclus -display | egrep '^#Attribute|HostMeters|Statistics'
#Attribute               Value
HostMeters               CPU    Mem     Swap
Statistics               Enabled


By default, the HostMonitor Agent will monitor for CPU, Memory and Swap utilisation.

The change to the HostMonitor Values for the HostMeters cannot be done when the cluster is online. The change needs to be done after stopping the Cluster Services and then editing the VCS configuration file "/etc/VRTSvcs/conf/config/main.cf" manually using an editor.

# hastop -local -force

Edit the "/etc/VRTSvcs/conf/config/main.cf" using an editor like 'vi'

cluster sfha62 (
        UserNames = { admin = chiAhcHtiDiiGqiChf }
        ClusterAddress = "192.168.10.104"
        Administrators = { admin }
        HostMeters = { "CPU", "Swap" }
        )

HostMeters attribute should have at least one entry configured and if you want to completely stop the agent from monitoring, we need to disable the "Statistics" attribute.

cluster sfha62 (
    UserNames = { admin = chiAhcHeiDiiGqiChf }
    ClusterAddress = "192.168.10.104"
    Administrators = { admin }
    Statistics = Disabled
    )


If you only want to change the Threshold Value of CPU, Mem or Swap, then you have to edit these attributes for each system in the cluster configuration :
Note: The Default values are Critical=90, Warning=80, Note=70, Info=60


system node1_name (
    MemThresholdLevel = { Critical=99, Warning=98, Note=95, Info=90 }
    CPUThresholdLevel={Critical=95, Warning=80, Note=75, Info=60}
    SwapThresholdLevel={Critical=95, Warning=80, Note=75, Info=60}
    )
   


To find more information on this subject, please refer to the Symantec™ Cluster Server 6.2 Administrator's Guide : Appendix VCS attributes : Cluster attributes and System attributes