I just want to confirm that "vSAN Recovery" during boot is not a normal thing? If it were not BAD then I would assume they'd use "initializing" or some other nomenclature that didn't suggest something had gone awry. Is this called PLOG recovery?
Just want to make sure that this is indeed a problem before I spend too much more time trying to chase down a cause.
"Recovery" sounds like a bad thing has happened.
So far, I have destroyed the vSAN datastore and recreated it. During that re-creation/configuration, the disks came up as already claimed as I had not tried to re-initialize/format them. After confirming that the "recovery" was still going on with every boot, I decided to go through and 1 by 1 "recreate" each disk group with the automated option for that ("The existing disks will be removed from disk group "52c9a082-352d-168f-039d-2acd87840284", and the disk group will be deleted. vSAN will recreate the disk group using the same disks.")
This is a new 3 node cluster and I have been making a LOT of configuration changes, hardware upgrades etc. So maybe it's just a good idea to wipe it all out and do a fresh install before populating it with actual VM's and workloads. The cluster is being booted from SD cards and scratch disk is in memory. One significant issue I had was screwing up vSAN networking and I thought I had fixed that OK, or that removing the datastore that was in use at that time would be sufficient to correct problems (if any) that remained.
Logs are showing many things to potentially investigate but nothing that I can see as a smoking gun. Health of the cluster in vCenter seems fine. Seems to function fine.
Just troubled by the word "recovery" and that it needs to do this even when the datastore is nearly empty.
Just want to make sure that this is indeed a problem before I spend too much more time trying to chase down a cause.
"Recovery" sounds like a bad thing has happened.
So far, I have destroyed the vSAN datastore and recreated it. During that re-creation/configuration, the disks came up as already claimed as I had not tried to re-initialize/format them. After confirming that the "recovery" was still going on with every boot, I decided to go through and 1 by 1 "recreate" each disk group with the automated option for that ("The existing disks will be removed from disk group "52c9a082-352d-168f-039d-2acd87840284", and the disk group will be deleted. vSAN will recreate the disk group using the same disks.")
This is a new 3 node cluster and I have been making a LOT of configuration changes, hardware upgrades etc. So maybe it's just a good idea to wipe it all out and do a fresh install before populating it with actual VM's and workloads. The cluster is being booted from SD cards and scratch disk is in memory. One significant issue I had was screwing up vSAN networking and I thought I had fixed that OK, or that removing the datastore that was in use at that time would be sufficient to correct problems (if any) that remained.
Logs are showing many things to potentially investigate but nothing that I can see as a smoking gun. Health of the cluster in vCenter seems fine. Seems to function fine.
Just troubled by the word "recovery" and that it needs to do this even when the datastore is nearly empty.
Last edited: