I'll try to be detailed...
I was using an older version of Armory on Linux Mint 18.1.  I had not booted the machine in months.  After syncing the blockchain and starting Armory everything looked fine with the watching only wallet I had loaded.  After importing other wallets from paper backups the transaction history for the watching only wallet was reduced to two entries.  The newly imported wallets had 0 balance.  I checked their addresses and "Used" listed no addresses.  This is wrong.  Even if somehow coins were stolen there should be Used addresses, or transaction history, or something.  Rebuild and Rescan Databases hangs for hours.

I remove Armory.  I delete my copy of the blockchain and update bitcoin-core.  I check the version is correct:  https://pastebin.com/9EqTBfsz

I run bitcoin-core via:

nohup bitcoind -datadir= ${btc_local_chain} -daemon >> ${btc_local_chain}/bitcoind.log

where btc_local_chain is defined as:  btc_local_chain="/media/vault/blockchains/bitcoin"

I allow the blockchain to fully sync.  I check the blockheight is correct via:

btc-cli getblockcount
wget -q -O - http://blockchain.info/q/getblockcount

I update gcc and g++ and their symlinks, and confirm version:
https://pastebin.com/L8BDy8ku
https://pastebin.com/Z0DZV7hx

I try installing and uninstalling armory 0.96.5 and 0.96.4 (the GCC 7.2 and 4.9 versions) various times, always getting "Failed to spawn DB!"
I start documenting for this post:

I update mint to 18.3 through Software Manager GUI.  I reboot.

I remove Armory and install armory_0.96.5_amd64_gcc7.2.deb:
https://pastebin.com/qhS8iDRf

I reboot.

I start Armory in offline mode and uncheck "Let Armory run Bitcoin Core..."
I set Bitcoin Home Dir to: /media/vault/blockchains/bitcoin

I check ArmorySettings.txt to confirm the blockchain location is properly set:
https://pastebin.com/BJ0Yt5vz

I create armoryqt.conf and armorydb.conf in ~/.armory/, each contain a single line:
satoshi-datadir="/media/vault/blockchains/bitcoin"

I run: sudo apt-get update
I run: sudo apt-get dist-upgrade
https://pastebin.com/AGmYtbQ4

I reboot.

I run bitcoind as before, and again check the blockheight is correct.  I double check with https://www.blockchain.com/btc/block-height/

I wait for my HDD to stop spinning.

I empty the armorylog.txt file for clarity.

I start Armory from Linux Mint Menu.  Again:  Armory Failed to spawn the DB!
I skip creating a wallet and close Armory.  I check the logs:
armorylog.txt: https://pastebin.com/x4dh1Eqf
dbLog.txt: https://pastebin.com/ajdMszs3

The database invocation looks correct.  It simply fails.
I launch Armory again, again the DB error.  I "create my first wallet" with all defaults and password = password, just to test.
Armory immediately closes without creating the wallet.  I check ps -e | grep armory and it is not running.  There are no .wallet files in the directory.
I launch Armory again.  DB Error.  It wants me to create a wallet.  I click skip.  I do Help->Rebuild and Rescan Databases.
I quit Armory and launch again.  DB Error.  It wants me to create a wallet.
Logfile from the last two launches:  https://pastebin.com/eayz9YNG   [I copied this before closing Armory].

At this point I feel like my only option is to reformat all drives, install the latest mint, download blockchain again, etc. etc. etc.  I'd rather not.  Any other options?

Please help, thank you.

Selling at a loss so I can order 841's:  $1650 a piece, discounts for five or more.  Can match the proper number of AUC3 and Avalon Controller to your order.  Buy enough and I'll go as low as $1300.  Shipping immediately from Saint Louis, MO.  Buyer pays shipping.

Accept PayPal, Wire Transfer, BTC, ETH, or ZEC.

Legally registered and licensed business.
PM me!

James Lea
Infomorphics LLC

Unused and unopened Avalon 741's from the November 2017 batch.  Paypal or BTC or ZCash (escrow).  The person who bought these didn't have space for them and asked me to get rid of them  

Five available.

[I suppose I should just toss the PICO and adapters, but want to see if there's something obvious and stupid I'm doing. And if anyone knows why this would happen at the SATA in of all places. Are these boards especially prone to problems?]

I have been building PC's for 15 years.

With my mining rigs I have been using ASUS and ASRock mobo's.  Always liked the brands (especially ASRock).  Recently Newegg had a bunch of open box Biostar TB250-BTC for cheap so I picked up 7 of them.

So far two have caught on fire shortly after booting with 6 1080 ti's on powered risers.  These 1080 ti's have previously run for a week straight on an ASRock H81.  The fire started next to/underneath the chipset heatsink near the SATA power in.

In both situations I ran the motherboard for some time without any graphics cards plugged in.  Different power supplies.

I am using 160W Pico-PSU (ITX) plugged into the 24Pin.  I have made adapters by hand crimping mini-fit jr's to take the Pico's 5pin outlet to a chain of 3 4pin MOLEX.   The first MOLEX feeds a MOLEX->SATA power adapter (bought) that plugs into the motherboards SATA power in.  The second MOLEX plugs into the inlet near the CPU, the third to the inlet on the side of the board.  The PICO is fed by a 5mm pigtail from Eyeboot that I spliced into a 6pin PCIE -> 8pinEPS at the EPS terminals.  I've made and used these adapters before without issue.  I use 18AWG stranded hookup wire.

I've read the extra power inlet's aren't technically necessary when using powered risers but where I come from we don't risk underpowering hardware.

In the first case the EPS was fed from a 1200watt HP server supply with an X-Adapter.  The risers were also fed directly from this (they have 6pin inlets).  The top two inlets (6 pin and 8pin) were fed from a separate 1200W HP psu.  I figured I'd have a max of 100W per inlet when hashing (PL @ 205W) and wanted to try split supplies.

I figured the MOLEX->SATA adapter was to blame (a ground pin was loose) so I tossed it.  I was concerned that the GFX cards might have tried to draw too much from one PSU on boot so for the next attempt I used a Delta 2400W PSU with two GekkoScience breakouts, and plugged all 19 connectors into it.  I think I used a different PICO PSU but I'm honestly not sure.  I used the same hand made PCI->EPS and 5pin->MOLEX->MOLEX->MOLEX adapters.  I suppose the adapters could have a short, but I don't see one.

Fire in the same spot on the second mobo.

I could test the adapters with an H81/Pico/adapter combo I know is good.  I suppose I should just toss the PICO and adapters, but want to see if there's something obvious and stupid I'm doing.  And if anyone knows why this would happen at the SATA in of all places.  Are these boards especially prone to problems?  I'm thinking about returning all of them, never liked Biostar much anyway.

Maybe I will stop hooking up on-board MOLEX/SATA in's.