Been trying like crazy to figure out what is filling the ramdisk. I have tried looking at numerous articles but now I am real confused on how to approach this. A 'vdf -h' shows the ramdisk root has 0% free space. I had deleted all *.gz files from /logfiles and that seemed to free up only about 12MB. I looked today and it's at 100% again, so that is not the REAL problem. One I noticed is even though it is at 100% used it is not throwing those "full" messages right now. It has only been a day since I deleted the gz's. Another thing to mention is we have 4 hosts. host 1 showed the full messages and the 100% usage. The other hosts have not shown the "full" messages.
I need a clear procedure on how to check this. I have visited numerous VMware articles including "investigating disk space usage" and it has not helped me. Thoughts?
Hi @AlanRaczek ,
1.Whatis the make and model of the host ?
2. Please confirm if you configured host logging to a persistent location and not the default ramdisk
3. Please check /var/run/log/vmkernel, hostd.log and vmkwarning.log on the host to see which log directory is being reported in the error
Thanks for the reply.
Systems are Cisco C240 M4 servers running ESXi 6.5U3, vSAN . 4 hosts. See pic's below on log config:
I probably have it set incorrectly. Just trying to get around the fact I can't put logs on vSAN. And there really are no other datastores available. Was trying to put logs on local disks.
See errors below in log:
Thank you vm.
Scratch partition is configured to default /tmp directory. Change it to a local or SAN datastore
Hope that helps
Forgive me for being ignorant, but therein lies my problem, identifying a local disk to use and how I reference it. I honestly get confused on what directories are a "ram disk" and what directories are "local disk".
...alan
Run either esxcfg-scsidevs -m or esxcli storage vmfs extent list