This week I had the task of installing VMware ESXi 7 on a Cisco UCS blade booting from SD Cards.
Once ESXi was installed and added into vCenter, I saw the warning ‘No coredump target has been configured. Host core dumps cannot be saved‘
I had a quick google search and also opened up a support ticket with VMware.
Because the datastores mounted to these new ESXi servers are connected via iSCSI, we do not have the option to save a coredump file to them. Pointing Coredumps to an iSCSI or FCoE backed datastore, is not supported.
The VMware support engineer pointed me to KB article 77009, and asked me to set the boot variable – ‘allowCoreDumpOnUsb=TRUE‘. The default setting is set to false.
To set the boot variable, you have to reboot and press SHIFT+O when ESXi is booting. You have to be quick here as you only have 5 seconds.
Once you press SHIFT+O, you will enter the boot options. Here you append to the end of the boot options – ‘allowCoreDumpOnUsb=TRUE‘. Once you have entered the command, you can press ENTER and the ESXi host will boot.
If you establish an SSH session to the ESXi host and type in – ‘esxcli system coredump file list‘, you will see that the coredump file got created on the VMFS-L filesystem within OSDATA.
The only requirement for the Coredump file to be created is that the volume is larger than 4GB.
Once the Coredump is set correctly, the warning message disappears.
Interesting solution as I too faced this issue. I didn’t know that saving coredump files to an iSCSI disk is not supported. Can you please provide your source for this information. I didn’t want to send that many writes to an SD card as they have a limited number (all be it a very large number) of read/writes before failure. I set the advanced system setting, Syslog.global.logDir to point to an iSCSI mounted volume. This solution has been working for me for going on 6 years now. Thanks for the article.
Hi Dan, you can definately point it to an iscsi target however it is not supported. Please check this KB article: https://kb.vmware.com/s/article/2004299 a quarter of the way down you will see ‘Note: Configuring a remote device using the ESXi host software iSCSI initiator is not supported.’
Another (possibly) suitable alternative is to set these options using advanced settings (either from the host client or vSphere client) You could also insert these into a kickstart file if trying to PXE boot …
Apparently the “supported” method is to configure a network coredump target instead rather than the unsupported iSCSI/NFS method:
The vCenter appliance includes a CoreDump Collector service which you can configure as per:
I chose to just disable the warnings instead, but if I start getting PSOD for some reason I’ll implement this feature.
The environment that does not have Core Dump Configured will receive an Alarm as “Configuration Issues :- No Coredump Target has been Configured Host Core Dumps Cannot be Saved Error”.
In the scenarios where the Core Dump partition is not configured and is not needed in the specific environment, you can suppress the Informational Alarm message, following the below steps,
Select the ESXi Host >
Click Configuration > Advanced Settings
Search for UserVars.SuppressCoredumpWarning
Then locate the string and and enter 1 as the value
The changes takes effect immediately and will suppress the alarm message.
To extract contents from the VMKcore diagnostic partition after a purple screen error, see Collecting diagnostic information from an ESX or ESXi host that experiences a purple diagnostic screen (1004128).
Hi Clay, thanks for sharing
you can set this parameter via advanced system settings in the configuration tab. You have two parameters: VMkernel.Boot.autoPartitionCreateUSBCoreDumpPartition and VMkernel.Boot.allowCoreDumpOnUsb. By default, both are in false, but, in my case only with VMKernel.Boot.allowCoreDumpOnUsb is sufficient for clear the error.
Thanks Fernando for sharing this with everyone, great work