VMware Cloud Community
justinrutledge
Contributor
Contributor

VMware ESXi on Poweredge 1900

Hello all

I have a Dell Poweredge 1900 with the latest update of VMware ESXi on it. Everything runs exceptionally well, but I have noticed an anomaly. I do not know if this is a normal situation, but I wish to bring it to your attention. While watching the performance charts (updating every 20 seconds, by default) I noticed that my processors seem to spike every 10 minutes, almost on the dot. This anomaly does not seem to affect performance, but it doesn't seem right to me.

The machine specs are as follows:

2 X Intel Core 2 Duo 1.60 Ghz Processor

4 Gb Fully Buffered ECC RAM

1 X 80Gb HDD on the onboard SATA controller (this holds my esxi installation)

4 X 500Gb HDD on the PERC 5/i RAID controller

I have attached a snapshot of the chart as it appears all the time. Any ideas, or more information needed?

Tags (2)
Reply
0 Kudos
8 Replies
Dave_Mishchenko
Immortal
Immortal

Do you have any VMs running?

Reply
0 Kudos
justinrutledge
Contributor
Contributor

Good question! It does not matter whether I have any running or not. The picture attached to this question was taken with NO VMS RUNNING.

Reply
0 Kudos
Dave_Mishchenko
Immortal
Immortal

If you download the RCLI for ESX (get the appliance and not the Windows install), there's a command called resxtop. It'll give you a bit more information about the processes running on the host. If this is a test system you can enable SSH and then just run esxtop. If the RCLI you would run resxtop --server <your_hostname/ip> and then login with root. With SSH you just enter the command esxtop.

I have some intructions for both the RCLI and enabling SSH here - http://www.vm-help.com/. SSH access is not supported. The RCLI is the better option for a production system. If this is a test box, SSH can be a bit easier.

Reply
0 Kudos
Dave_Mishchenko
Immortal
Immortal

What model of SATA controller are you using? Don't think it's an issue - I just maintain a list of whiteboxes that work with ESXi

Reply
0 Kudos
justinrutledge
Contributor
Contributor

I have done what you said. Interestingly enough, the information shown doesn't really seem to correspond with the charts. TOP never shows the utilization over 1% per processor with no VMs running, but the charts stay steady at a spike every 10 minutes or so.

Reply
0 Kudos
Dave_Mishchenko
Immortal
Immortal

You might want to use esxtop in batch mode and then look at the file in Excel just in case the spike is happening too fast (which doesn't seem to be the case in the VI Client). Otherwise, if you have a support contract it might be worth logging a support call to see what they say.

Reply
0 Kudos
jkao
Contributor
Contributor

I'm seeing the same behavior on my VMware ESXi 3.5u2 on a PowerEdge 1950 III.

The spike happens every 10 minutes and using resxtop, one of the sfcbd processes is what appears to be taking up CPU during that time.

A cursory googling pulls up this post: http://communities.vmware.com/message/999785

Although as an ESX newbie, it's not apparent to me from that information, what this process is actually doing.

Reply
0 Kudos
vxxxbazaaz
Contributor
Contributor

Many servers have hardware watchdog timers which power-cycle the machine if the OS hangs for a set period of time, usually 10 minutes by default. Perhaps it is due to the interaction of ESXi with the Dell timer.

Reply
0 Kudos