After dealing with rouge logs filling up my server's space, let's clean up syslog.
I wanted to clean out what was in there and watch new stuff coming in as I turned off extra logging. I truncated the file with:
truncate -s 0 syslog
I had two domains that were expired and removed from nginx, but I never removed them from certbot.
certbot certificateslists all the certificates certbot is running. Then to delete a specific one:
certbot delete --cert-name example.xom
It already has it's own log, it doesn't need to be in here too! Go to:
/etc/rsyslog.d/20-ufw.confThe last line is commented out, uncomment it. This will stop logging in syslog. Then restart the rsyslog service:
service rsyslog restart
This has been on my to do list for a while but I started getting failing lets encrypt errors and I couldn't manually regen the certificate. Googling the errors shows that you need to upgrade the underlaying versions of things so good enough reason to actually move this.
Steps:
Ran
sudo apt-get update && sudo apt-get dist-upgradeon both droplets, both completed happily.
Thought I had set this up correctly but then gmail rejected my test email 2 days later.
Mailgun settings: Need to not use the mg. prefix
All of the other ones have it so not sure what I did there to make it work but this was a quick fix
DO Network: Put in the relevant Mailgun DNS
Mailgun Receiving Routes: need to set one up to forward the email address just set up to gmail
Optional: Sending from the address at gmail: this is where the SMTP settings come into play
This isn't technically on DO but my last website on nearlyfreespeech.
Looking at vaulters, it's TLS had expired. It had been warning me nightly but I thought the issue was on nearlyfreespeech's end. I first tried switching servers as vaulters was on an deprecated one. This didn't fix it and also killed PHP 5.6. I went to switch back but the oldest server didn't have 5.6.
To fix the TLS, I went into the CLI and just re-ran the Let's Encyrpt certificate generation. That worked and the TLS was fixed within ~30 seconds.
Now to fix the PHP version issue. The main page would load but the login page had lot's of PHP errors. I had never upgraded the site as the upgrader would always error out on me.
I downloaded a new version of wordpress on CLI and put it in a folder new. I figured I could copy over the site into /new and then switch it over to the live site once it was working. Wordpress installed into /new/wordpress so I was copying the files from /wordpress up into /new. Instead I copied into / , overwriting the original files. D:
Little did I know, this is exactly what I needed to do! It fixed the PHP errors and the login page loaded! I logged in and the menu was chaos but otherwise looked okay. I tried to redo the upgrade to grab the missing files in the Wordpress UI but I kept getting the error. I then tried it from the command line and it told me I was already up to date. Knowing some files were missing, I renamed wp-config folder and copied over everything from /new. That didn't fix the menu. I then did the same steps for wp-includes, and this fixed it!
It does look like all the installed plugins got disabled and the theme lost it's custom options though.
This has been going on since at least June. PHP needs to be restarted to get any changes to show up. Database stuff is fine. Wordpress sites have a lot of issues. enstars sever does not have this issue
Some how I had set opcache.validate_timestamps=0 in /etc/php/7.1/fpm/php.ini. I commented that back out to match enstars and actually got an error message on Wordpress!
Next steps:
FINALLY
And it's so much faster!