Recently I have been working with customers on designs for new vSphere 5.1 installs and upgrades. As part of the design, I have been specifying the installation and configuration of the vSphere ESXi Dump Collector service on their Windows vCenter Server. The ESXi dump collector service allows the collection of the diagnostic dump information generated when an ESXi host has a critical fault and generates a “purple diagnostic screen.”
This post is a walk through of installing and configuring the ESXi Dump Collector service on vCenter and configuring an ESXi host to use it.
The Windows Server 2008 R2 VMs I use for vCenter are configured with additional drives for installing applications and storing data. In this example from my virtual lab, I have a “d:\” drive for applications and data.
Install the vSphere ESXi Dump Collector
The installer for the dump collector in included on the vCenter installer ISO image. I mount the ISO image to the Windows 2008 R2 VM where I have installed vCenter server.
Launch “autorun.exe” as an administrator.
After the installer starts, select “English” as the language.
On the Welcome… page, click “Next >.”
On the End User Patent Agreement page, click “Next >.”
On the End User License Agreement page, select “I accept…”; click “Next >.”
On the Destination Folder page, click the “Change…” button beside “vSphere ESXi Dump Collector repository directory:”
On the Change Current Destination Folder page, change the “Folder name:” value to “d:\…”. Click “OK.”
Back on the Destination Folder page, observe that the path has been updated and click “Next >”
On the Setup Type page, select “VMware vCenter Server installation”, then click “Next >.”
On the VMware vCenter Server Information page, enter the appropriate information for connecting to vCenter. Click “Next >” to continue.
If you are using the default self-signed SSL certificate for vCenter, you will receive a message with the SHA1 thumbprint value for the vCenter server’s certificate. Click “Yes” to trust that the certificate for connecting to the vCenter server.
You can verify the thumbprint by looking at the certificate properties on your vCenter server. Notice that the thumbprint from the installer matches the thumbprint on the vCenter server’s certificate.
On the vSphere ESXi Dump Collector Port Settings page, click “Next >” to accept the default value of UDP port 6500.
On the vSphere ESXi Dump Collector Identification page, select the FQDN of the vCenter server and click “Next >.”
On the Ready to Install page, click “Install.”
After the installer has completed, click “Finish” on the Installer Completed page.
You can view the configured settings with the vSphere Client by selecting VMware ESXi Dump Collector from the Administration page.
You can also view the configuration with the vSphere Web Client by selecting the vCenter server, then browsing to the “Manage” tab and selecting “ESXi Dump Collector” under “Settings.”
Configuring an ESXi host to transmit the core dump over the network to the dump collector service.
Now that we have installed the dump collector service, we need to configure the ESXi hosts to send their diagnostic dump files to the vCenter server.
I set this up through the ESXi console. You will notice that I am logged in a “root” as I had not configured the ESXi host to use Active Directory authentication yet. Any user account that has the “administrator” role on the ESXi host can configure these settings.
First, checked the current coredump network configuration:
~ # esxcli system coredump network get
Network Server IP:
Network Server Port: 0
Next, I confirmed the name of the vmkernel connection I planned to use: “vmk0” with the old “esxcfg-vmknic -l” command
Then, I configured the system to send coredumps over the network through the “vmk0” vmkernel port to my vCenter server’s IPv4 address at port 6500:
~ # esxcli system coredump network set –interface-name vmk0 –server-ipv4 10.0.0.51 –server-port 6500
You have to enter the interface name and server IPv4 address. The port is optional if you are using the default of 6500.
Then, I enabled the ESXi host to use the dump collector service:
~ # esxcli system coredump network set –enable true
I verified that the settings were correctly configured:
~ # esxcli system coredump network get
Host VNic: vmk0
Network Server IP: 10.0.0.51
Network Server Port: 6500
I checked to see if the server was running:
~ # esxcli system coredump network check
Verified the configured netdump server is running
Here is a screenshot of the process:
FYI, by default, the diagnostic dump file (core dump) is stored on a local disk partition of the ESXi host. You can find the local partition from the local ESXi console (if it is enabled) with the following command:
# esxcli system coredump partition get
I have highlighted the command in the figure below:
More information about managing the ESXi core dump disk partition is in the online documentation here.
Now that the vCenter server has the dump collector service installed and the ESXi host is configured to use it, I had to try it out!
Using the vsish tool and specific setting that Eric Sloof or NTPRO.NL described in his post “Lets create some Kernel Panic using vsish“, I crashed the ESXi host. As you can see in the screenshots, I was rewarded with a purple screen and success with transmitting the dump over the network to my vCenter server!
Here is the coredump file that was transmitted. Success!
For more information check out these KB articles: