Hello,
I have a HP Microserver G10 with dual NIC and ESXi running VMs on 2 different physical networks (LAN and DMZ). The DMZ is wholly contained within ESX and all those VMs are on a virtual switch, with the NIC linked directly to the firewall/gateway. I have 1 SSD with ESXi installed and it also hosts all my VMs on that same local datastore. I have 2 JBOD HDDs that are mirrored using ZFS inside one of the VMs on the LAN, where all my important files/media/vm backups are stored. I do not have backups for the ESX instance. I also have 2 of 4 open drive bays and some large spare disks to play with.
This morning I woke up to find my DMZ unaccessible from other networks. The VMs themselves were able to ping each other, for example, but not the gateway nor any LAN/WAN addresses.
After putting the host into maintenance mode and shutting it down & rebooting it, it never came back up. The boot hits a fatal error when loading the state.tgz file (see screenshot). After some research, I think the corrupted file affected the DMZ network's NIC settings overnight causing that network link to no longer function.
I guess the SSD is nearing EOL and this corrupted system file is the first sign of its impending demise. I still had the USB installer from 2.5 years ago when I built this system. I put that in and also another 16GB USB stick, and installed a fresh ESX 6.7 on that 16GB stick. I boot into that and I am able to see all my datastores, and I can browse them and see the folders for all my VMs. Then I shut down the host again, until I figure out some action plan.
My questions:
1. Is there any way for me to recover the configs from the corrupted installation and bring them into the new one? I realize I may need to manually reconfigure some of the network settings I've lost.
If not, at worst I'd have to reconfigure ESXi from scratch.
2. If I wanted to migrate the VMs from the SSD to a new drive, how do I do that? Can I just clone it to another drive? Or is there a VMWare method to migrate the files if both datastores are present inside ESX? It's a 500GB SSD with maybe 100GB of free space.
3. If I want to "import" the VMs on the SSD (or the new drive) into the new ESX instance, how do I do that?
4. Should I do #2 on a separate system, before I do #3? I feel this is the most prudent plan.
Yeah, I saw that one. I could try that power cycle / nvram reset. But I really think it's time for a new SSD.
Probably it will be the best option.