Smarts ESM: No events from ESM for high CPU and Memory from VM hosts; Instrumentation data is missing for VM Hosts
Article ID: 332045
Updated On:
Products
VMware Smart Assurance
Issue/Introduction
Symptoms:
No events from ESM for high CPU and Memory from VM hosts
ESM does not alarm on VM hosts for CPU or memory when the usage exceeds a set threshold. However, it does alarm on CPU and Memory for the same devices in the IP domain.
No events from ESM for high CPU and Memory from VM hosts
ESM does not alarm on VM hosts for CPU or memory when the usage exceeds a set threshold. However, it does alarm on CPU and Memory for the same devices in the IP domain.
Environment
VMware Smart Assurance - SMARTS
Cause
Smarts ESM is working as designed.
Instrumentation is not created to prevent duplicate polling and statistical collection by the two domains:
Smarts Server Manager (Smarts ESM) and Smarts IP domains.
For users who want to see these statistics in their Smarts ESM domain, as well as the Smarts IP domain, a new option was introduced in Smarts ESM version 9.2.2. See the resolution section for details.
Instrumentation is not created to prevent duplicate polling and statistical collection by the two domains:
Smarts Server Manager (Smarts ESM) and Smarts IP domains.
For users who want to see these statistics in their Smarts ESM domain, as well as the Smarts IP domain, a new option was introduced in Smarts ESM version 9.2.2. See the resolution section for details.
Resolution
To enable polling and the collection of these values for VM hosts on Smarts ESM, follow these steps.
1: Ensure you have added the Vcenter credentials by right clicking on the ESM domain name. Ensure the Vcenter host is added on the IP domain. If you do not have SNMP access, you can add it as ICMP only.
2: Set the below parameter in ESM.import to TRUE.
# Set to TRUE to discover Processor, FileSystem, Disk and Memory.
# via SNMP. Not applicable to WMI discovery.
EnableCPUFilesystemMemoryDiscovery = FALSE
Doing this will enable this instrumentation data to be collected:
1: In Vcenter you should see these things:
"Virtual machine memory usage" with Red color (at Vcenter) for generating "IsVMMemoryOverused' event in ESM.
"Virtual machine cpu usage" with red color (at Vcenter) for generating "IsVMCPUOverused" event in ESM.
On the VMwareESX class in ESM to generate the following events:
HostCPUUsageAlarm
InsufficientMemory
2: In ESXi you should see these things:
"Host memory usage" event with "red color" (at Exi) for generating "InsufficientMemory" event in ESM.
"Host cpu usage" event with "red color" (at Exi) for generating "HostCPUUsageAlarm" event in ESM.
NOTE2: You can use these two DMCTL commands to find out what can be discovered through the Vcenter API.
$ ./dmctl.exe -s testESM -b localhost:9886 getEvents Hypervisor |tee
Exported events of class Hypervisor:
Disconnected = <problem>
Down = <problem>
HostCPUUsageAlarm = <problem>
HostHAFailed = <event>
HostLostConnection = <event>
HostLostConnectionToAtLeastOneDatastore = <event>
HostMemoryUsageAlarm = <problem>
HostNotResponding = <event>
HostShutdown = <event>
HostSyncFailed = <event>
HypervisorPowerOff = <event>
InsufficientCPU = <event>
InsufficientMemory = <event>
NotHosted = <event>
NotManagedByVC = <event>
ShutDown = <problem>
And following are the events from the VirtualMachine class:
$ ./dmctl.exe -s testESM -b localhost:9886 getEvents VirtualMachine |tee
Exported events of class VirtualMachine:
DownOrImpaired = <event>
GuestOSNotRunning = <problem>
IncompleteNetworkConfiguration = <event>
NetworkReconfiguration = <event>
PerformanceMightBeImpacted = <event>
PoweredOff = <problem>
Unreachable = <problem>
VMAdded = <event>
VMCPUUsageAlarm = <event>
VMDeleted = <event>
VMHighDiskLatency = <event>
VMIsOrphanedE = <event>
VMIsOrphanedImpacted = <event>
VMLostConnection = <event>
VMMemoryUsageAlarm = <event>
VMMigrated = <event>
VMMigratedErrorE = <event>
VMMigratedImpacted = <event>
VMMovedByDRS = <event>
VMRelocatedErrorE = <event>
VMRelocatedImpacted = <event>
VMRestartedOnAlternativeHost = <event>
VMToolsNotInstalled = <event>
VMToolsNotUpToDate = <event>
NOTE2: More information can be found on page 203 of this guide: https://support.emc.com/docu57508_Smarts_Server_Manager_9.4_User_and_Configuration_Guide.pdf?language=en_US
1: Ensure you have added the Vcenter credentials by right clicking on the ESM domain name. Ensure the Vcenter host is added on the IP domain. If you do not have SNMP access, you can add it as ICMP only.
2: Set the below parameter in ESM.import to TRUE.
# Set to TRUE to discover Processor, FileSystem, Disk and Memory.
# via SNMP. Not applicable to WMI discovery.
EnableCPUFilesystemMemoryDiscovery = FALSE
- Launch this command from this folder : <installed base>ESM/smarts/bin to edit it.
- ./sm_edit conf/esm/ESM.import
Doing this will enable this instrumentation data to be collected:
- Processor.
- FileSystem.
- Disk.
- Memory.
1: In Vcenter you should see these things:
"Virtual machine memory usage" with Red color (at Vcenter) for generating "IsVMMemoryOverused' event in ESM.
"Virtual machine cpu usage" with red color (at Vcenter) for generating "IsVMCPUOverused" event in ESM.
On the VMwareESX class in ESM to generate the following events:
HostCPUUsageAlarm
InsufficientMemory
2: In ESXi you should see these things:
"Host memory usage" event with "red color" (at Exi) for generating "InsufficientMemory" event in ESM.
"Host cpu usage" event with "red color" (at Exi) for generating "HostCPUUsageAlarm" event in ESM.
NOTE2: You can use these two DMCTL commands to find out what can be discovered through the Vcenter API.
$ ./dmctl.exe -s testESM -b localhost:9886 getEvents Hypervisor |tee
Exported events of class Hypervisor:
Disconnected = <problem>
Down = <problem>
HostCPUUsageAlarm = <problem>
HostHAFailed = <event>
HostLostConnection = <event>
HostLostConnectionToAtLeastOneDatastore = <event>
HostMemoryUsageAlarm = <problem>
HostNotResponding = <event>
HostShutdown = <event>
HostSyncFailed = <event>
HypervisorPowerOff = <event>
InsufficientCPU = <event>
InsufficientMemory = <event>
NotHosted = <event>
NotManagedByVC = <event>
ShutDown = <problem>
And following are the events from the VirtualMachine class:
$ ./dmctl.exe -s testESM -b localhost:9886 getEvents VirtualMachine |tee
Exported events of class VirtualMachine:
DownOrImpaired = <event>
GuestOSNotRunning = <problem>
IncompleteNetworkConfiguration = <event>
NetworkReconfiguration = <event>
PerformanceMightBeImpacted = <event>
PoweredOff = <problem>
Unreachable = <problem>
VMAdded = <event>
VMCPUUsageAlarm = <event>
VMDeleted = <event>
VMHighDiskLatency = <event>
VMIsOrphanedE = <event>
VMIsOrphanedImpacted = <event>
VMLostConnection = <event>
VMMemoryUsageAlarm = <event>
VMMigrated = <event>
VMMigratedErrorE = <event>
VMMigratedImpacted = <event>
VMMovedByDRS = <event>
VMRelocatedErrorE = <event>
VMRelocatedImpacted = <event>
VMRestartedOnAlternativeHost = <event>
VMToolsNotInstalled = <event>
VMToolsNotUpToDate = <event>
NOTE2: More information can be found on page 203 of this guide: https://support.emc.com/docu57508_Smarts_Server_Manager_9.4_User_and_Configuration_Guide.pdf?language=en_US
Feedback
Yes
No
Powered by
