JonathanM

For a minority of apps, that may be quicker. For most though, it would slow things down, and possibly create a mess. It could be a nice advanced option to apply per app, I wouldn't want it as a blanket option. Another advanced option possibility could be BTRFS snapshots, but that would probably require special handling and setup.

Sure, remove any disks from the parity slots. The ONLY reason a disk is disabled is that a write failed, which means it's no longer in sync with parity. Remove parity, no more disabled disks.

No clue if it will actually perform better, but there is an iSCSI plugin available.

Did you read the link? What exactly needs more explanation? It would be helpful to know what is unclear in the manual so it can be better explained.

Pass the USB through to a VM.

Have you tried SeaBIOS instead of OVMF?

Are any other people seeing this in their logs? WARNING: SQLite: os_unix.c:36982: (40) openDirectory(/var/urbackup) - errorcode: 14 The only google result that looks relevant references symlinks for var/urbackup no longer working and needing bind mounts, so I'm wondering if I messed something up in my config. To be clear, I can't see any issues besides a WARNING in the log file, everything seems to be working fine.

You are correct, of course, but running /var/log out of space isn't great either, as you know.

The failed attempts will eventually cripple your server by filling up the /var/log mount. You have been warned.

vdisks are single large files, and can only exist on a single drive. They CAN however, be configured to appear larger on the inside, like a tardis. As you fill it, the real file size expands, but CAN'T flow across to other disks, and if you put more data than the disk is big, you will get error messages and crashes. If your VM vdisk is on a drive or pool smaller than 6TB, you will have a bad time if you fill it. Since you say you just created the VM fresh, I'd say nuke it and start over. Whether Steam will let you store files on a network share I don't know, so you will have to figure that out. If not, your only recourse would be to set up a vdisk on a drive or pool large enough to hold everything, or add vdisks and map them to paths internal to your VM.

Restore from backup.

Are you using a specific app password? Your regular google account password will not work.

Looks like your flash drive may be failing. Probably writing the update files pushed it over the edge. Do you have a relatively current flash drive backup?

Can you describe your layout more completely? A VM should have its own unique IP, and ports shouldn't conflict with Unraid.

Attach a monitor and see what's on the local console.

The only thing I can think of would be to make a backup of the config folder on your USB stick, then disable array auto start by editing disk.cfg on your flash drive (config/disk.cfg) and changing startArray="yes" to "no", That way you can look over everything before the array starts to make sure all your disks are correctly detected. (They should be)

Do you have all your drives intact including your USB boot key? Did you have VM's with hardware passthrough? Did you have any hardware RAID controllers?

Neither. Rebuilding an existing drive replaces ALL the bits on the drive, overwriting any existing content including a format.

Not sure we are on the same page. Parity swap actually creates a duplicate of your parity drive as in interim step before rebuilding your data on the old parity drive, so technically for a portion of the procedure you have 2 copies of your parity drive.

Using a different interface besides USB.

See Q4 https://github.com/binhex/documentation/blob/master/docker/faq/vpn.md You need to add the IP you will be connecting from.

Same in Unraid's safe mode?

Identical behaviour on different browsers / incognito modes?

Starting Unraid with only pools is still a possibility. It's been discussed in the past, but I am not privy to how far that discussion has gone. Who knows, it might show up sooner rather than later. I wouldn't count on it until it's officially announced though. Adding single volume and ZFS pools is probably enough major changes for this upcoming cycle. Too many changes at once makes for a very messy release.

Do you read the parity swap link?