http://cormachogan.com/2015/02/24/vsan-considerations-when-booting-from-usbsd/
This clears it up but does not mention viable solutions.
In my opinion now I need to go back to using use two mirrored physical disks for VSAN and vmkernel logs; VSAN trace files; capturing core dumps (PSOD).
vSphere 6.0 Documentation Center
https://kb.vmware.com/kb/1033696
Thoughts
Jeff
a solution for storing them on USB/SD is being worked on... But it will take time.
a solution for storing them on USB/SD is being worked on... But it will take time.
should take a look.
1. mixing raid mode under same controller is not supported.
you may configure RAID 0 or pass-though for vSAN
if you mirror two disks, if your vsan disk are under same controller, vsan disk should be single raid 0 device. if your controller is certified only for pass though, you can't make mirror it.
2. if VMFS and vsan datastore are under same controller.
you should not running VM on VMFS, use it only for scratch and vsantrace.
in case of if you are running Dell H730, it has more restrictions.
Hi,
Thanks for the catch! I will go back to my original ESXi builds using dual SD and deal with it awaiting for Duncan's fix mentioned above. I just received "Essential Virtual SAN (VSAN)" 2nd edition today. I am also investigating VXRail but I am sure this problem will still exist as well.
Cheers
Jeff
I'm really interested in this. Would love to be an early beta tester.
You can redirect the vSAN traces to a syslog server. When you configure syslog such as LogInsight (recommended) for your ESXi syslogs, the urgent vSAN traces are redirected as well.
I'm curious to know what urgent vSAN logs means? The KB mentions "The use of syslog redirection for the urgent traces is a useful troubleshooting tool, as it can permit the evaluation of trace data even if the traces are lost." To me that sounds like you should use this in addition to saving all the trace logs locally. Also, what proportion issues only require the urgent traces for troubleshooting? Is this something like 75% or 99%? What is the stance of VMware support if they only have urgent traces to work with? Will they only be able to troubleshoot some problems? To me this doesn't sound like a good option if you run production workload on vSAN, and you want to be able to get full support for all types of problems.
It would be great to get more details what urgent traces actually mean, and how it should be used.
Thanks
Per
Duncan,
I thought this was already addressed with vSAN 6.5?
Extending an ESXi diagnostic core dump partition on a vSAN 6.5 node - CormacHogan.com
Extending an ESXi diagnostic coredump partition on a vSAN 6.5 node (2147881) | VMware KB
VMware KB I also ask because I have the vSAN sales team presenting now :smileysilly:
Thanks,
BTG