Requesting VMware ESXi logs (KBA1460)
KBA
KBA#1460VMware ESXi Logs
Delphix Support may ask for the VMware ESXi logs of the ESXi host which is hosting the Delphix Engine, to facilitate troubleshooting of issues such as a hung or unresponsive engine.
The logs can be extracted by one of the following methods:
Method 1
Use the vSphere client to export a host’s syslog. In the vSphere Client, right-click on an ESXi host in the inventory pane. Click Export System Logs.
Note: you can optionally click on “Gather Performance Data”. More detail on how to do this is in https://kb.vmware.com/s/article/653.
Method 2
Alternatively, the /usr/bin/vm-support command can be used to collect diagnostic information (ESXi log files), as per https://kb.vmware.com/s/article/1010705, under “Running vm-support in a console session on ESXi/ESX hosts”. This produces a gzipped tarball (.tgz file) locally on the host that can be copied off the host using a file transfer protocol, for example, FTP or SCP.
The following VMware KM note explains what data is collected by vm-support - https://kb.vmware.com/s/article/2147388. It might be prudent to avoid collecting core dump files during the host support bundle collection.
Note: For collection of crashdumps from hung or unresponsive engines, see the next section of this document "Getting A Crashdump from a Hung or Unresponsive Engine" below.
Method 3
VMware vSphere PowerCLI can be used to download a vm-support log bundle from VMware vSphere vCenter Server and/or ESX/ESXi hosts, or to view or search the individual log files stored on the host. (VMware vSphere PowerCLI can also be used to download a vc-support log bundle from VMware vSphere vCenter Server if this is also needed).
This VMware note provides information on how to do this: https://kb.vmware.com/s/article/1027932
This note explains what data is collected: https://kb.vmware.com/s/article/2147388
Getting A Crashdump from a Hung or Unresponsive Engine
You may be asked to generate a crash dump from an unresponsive engine (hung VM) before it is rebooted, in order to troubleshoot the cause. This can be achieved by generating a Non-Maskable Interrupt (NMI) or alternatively, generating a VMware snapshot. Information on how to do these is provided in KBA1129 - https://support.delphix.com/Continuous_Data_Engine_(formerly_Virtualization_Engine)/Platforms/How_to_Generate_a_non-maskable_interrupt_in_VMware_ESX__(KBA1129).