Guh. Crashed my server at Linode and spent 24 hours not grokking how to get ”foolproof, just copy & hit start!” backup started. (I had reached the point of d/l-ing an IRC client to get into their chat when I realized I needed to set a new ”configuration”, not boot the old one.)
UPDATE: This isn’t quite right. On further review, the default ‘Wordpress’ config is identical to the new one I created yesterday, but for the kernel setting. The default is set to GRUB2, rather than a specific kernel (of which there are dozens to choose from). It was a GRUB boot error I was having with the backup volume. When I tried doing the configuration myself, I (naively) assumed I should just pick “newest kernel”, so I did and that worked.
AND ALSO? When I first thought I’d succeeded, I hadn’t actually, because I wasn’t putting the swap volume on
/sdb (or whatever). That didn’t manifest as a problem when I ran all my little node apps (although I did think they were slower than normal?), but mysql/mariadb was falling over every time I tried to hit the wordpress install
var/www/html, and/or the Nextcloud app, and/or the Gitea service. When I realized those are all highly db-reliant I knew I had a mariadb issue; after a few hours of fiddling and reading, I was convinced it was a memory-constraint problem but wasn’t sure how to diagnose. Woke up this morning and instantly thought “wait what about the swap disk?” and sure enough. When added to the configuration panel as a second volume, everything’s back to snappy.