Can't see several network counters like net.throughput.usage.vm.average

[samfang]

I refered to this KB, and change collection level for several network counter.

VMware KB: Changing the collection level for Storage DRS and SIOC data counters in vSphere 5.0 Updat...

then level became 1.

Get-PxCounterLevelMapping  | ? {$_.Name -match "net.throughput.usage" }

Name                          AggregateLevel                PerDeviceLevel                Server

----                          --------------                --------------                ------

net.throughput.usage.average  1                             1                             Server1

net.throughput.usage.vm.av... 1                             1                             Server1

net.throughput.usage.nfs.a... 1                             1                             Server1

net.throughput.usage.vmoti... 1                             1                             Server1

net.throughput.usage.ft.av... 1                             1                             Server1

net.throughput.usage.iscsi... 1                             1                             Server1

net.throughput.usage.hbr.a... 1                             1                             Server1

host1 is 5.1.* ESXi Host.

C:\> Get-VMHost host1 | Get-StatType -Interval 1800 | ? {$_ -match "net" }

net.usage.average

net.usage.minimum

net.usage.maximum

net.packetsRx.summation

net.packetsTx.summation

net.received.average

net.transmitted.average

net.throughput.provisioned.average

net.throughput.usable.average

net.throughput.usage.average

net.throughput.contention.summation

net.throughput.packetsPerSec.average

net.throughput.usage.vm.average

net.throughput.usage.nfs.average

net.throughput.usage.vmotion.average

net.throughput.usage.ft.average

net.throughput.usage.iscsi.average

net.throughput.usage.hbr.average

net.droppedRx.summation

net.droppedTx.summation

net.bytesRx.average

net.bytesTx.average

net.broadcastRx.summation

net.broadcastTx.summation

net.multicastRx.summation

net.multicastTx.summation

net.errorsRx.summation

net.errorsTx.summation

net.unknownProtos.summation

host2 is 5.5.* ESXi Host.

C:\> Get-VMHost host2 | Get-StatType -Interval 1800 | ? {$_ -match "net" }

net.usage.average

net.usage.minimum

net.usage.maximum

net.packetsRx.summation

net.packetsTx.summation

net.received.average

net.transmitted.average

net.droppedRx.summation

net.droppedTx.summation

net.bytesRx.average

net.bytesTx.average

net.broadcastRx.summation

net.broadcastTx.summation

net.multicastRx.summation

net.multicastTx.summation

net.errorsRx.summation

net.errorsTx.summation

net.unknownProtos.summation

For the counter like net.throughput.usage.vm.average, I can't see in host2, anyone know reason?

[LucD]

How much time was there between the point where you changed the statistics level for the counters and the moment you did the Get-StatType ?

Since you ask for the Historical Interval 2, there should be at least 1 day between the two.

Perhaps you made the change to the statistics level sooner on host1 ?

[samfang]

host1 and host2 in one cluster1,  I changed performance counter level for cluster1, supposed to have same impact and time period for both host? Btw, I changed yesterday, probably already 24 hours.

Yesterday steps.

1. Run "Connect-VIServer" to connect to cluster1.

2. Change counter level.

3. Get-PxCounterLevelMapping  | ? {$_.Name -match "net.throughput.usage" }

Today steps

1. Get-PxCounterLevelMapping  | ? {$_.Name -match "net.throughput.usage" }

2. Get-VMHost host1 | Get-StatType -Interval 1800 | ? {$_ -match "net" }

3. Get-VMHost host2 | Get-StatType -Interval 1800 | ? {$_ -match "net" }

From spec, seems those counter is on the cluster level. After apply, will impact all hosts under that cluster, right?

net.throughput.usage.vm.average

net.throughput.usage.nfs.average

net.throughput.usage.vmotion.average

net.throughput.usage.ft.average

net.throughput.usage.iscsi.average

net.throughput.usage.hbr.average

[MKguy]

They should show up after the first rollup interval (eg 1800 seconds) already. It's possible that these counters were removed in ESXi 5.5 altogether though.

Check if these counters are available in the realtime performance counters:

Get-StatType -Realtime | ? {$_ -match "net" }

If they don't show up in the realtime interval, then it seems like these host counters were removed entirely. Or they might have been renamed in 5.5, run Get-StatType -Realtime without the filter and check if you see anything.

[samfang]

For realtime, both host can't see those counters. Only show below matched to net.

net.multicastRx.summation

net.usage.average

net.bytesRx.average

net.multicastTx.summation

net.unknownProtos.summation

net.transmitted.average

net.errorsRx.summation

hbr.hbrNetTx.average

net.errorsTx.summation

hbr.hbrNetRx.average

net.packetsTx.summation

net.received.average

net.droppedTx.summation

net.droppedRx.summation

net.packetsRx.summation

net.broadcastRx.summation

net.bytesTx.average

net.broadcastTx.summation

In 5.5, run Get-StatType -Realtime, seems nothing related throughput with network traffic type.

[LucD]

Not sure if the counter should be visible for an ESXi.

Afaik they are aggregated counters that only appear in Historical Interval 1, not in the Realtime interval.

Is the counter returned when you do a Get-StatType for the cluster entity ?

[samfang]

If cluster level, it didn't return.

C:\> Get-Cluster | Get-StatType

cpu.usagemhz.average

cpu.usage.average

mem.usage.average

mem.overhead.average

mem.consumed.average

mem.vmmemctl.average

net.throughput.provisioned.average

net.throughput.usable.average

net.throughput.usage.average

net.throughput.contention.summation

vmop.numPoweron.latest

vmop.numPoweroff.latest

vmop.numSuspend.latest

vmop.numReset.latest

vmop.numRebootGuest.latest

vmop.numStandbyGuest.latest

vmop.numShutdownGuest.latest

vmop.numCreate.latest

vmop.numDestroy.latest

vmop.numRegister.latest

vmop.numUnregister.latest

vmop.numReconfigure.latest

vmop.numClone.latest

vmop.numDeploy.latest

vmop.numChangeHost.latest

vmop.numChangeDS.latest

vmop.numChangeHostDS.latest

vmop.numVMotion.latest

vmop.numSVMotion.latest

clusterServices.effectivecpu.average

clusterServices.effectivemem.average

cpu.totalmhz.average

mem.totalmb.average

clusterServices.failover.latest

[aqiqul_dell]

- found out that the counters went missing in the 5.5x versions :

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=208359...