Build new wordpress container, move to osgeo9

[robe]

osgeo7-wordpress instance is currently at buster and PHP 7.3

Given PHP 7.3 is at EOL and has some security vulnerabilities, this will be my priority to upgrade.

I will upgrade osgeo4-wordpress-dev first which is a snapshot osgeo7-wordress with ​https://dev.www.osgeo.org being about a month or 2 old of current production www.osgeo.org to confirm no issues.

[robe]

I've upgraded wordpress-dev to bullseye (debian 11). Kept the /etc/ldap/ldap.conf and /etc/ssh/sshd_config files, though the sshd_config I was surprised to not see anything really custom, all just commented out stuff and include/sshd_conf.d so I suspect that one i should have allowed overwrite with package maintainers.

The final apt dist-upgrade caused some issues with

Created symlink /etc/systemd/system/sysinit.target.wants/systemd-pstore.service → /lib/systemd/system/systemd-pstore.service.
Failed to get load state of systemd-networkd.socket: No buffer space available

and abruptly shutdown, and had to manually start it back up and showed incomplete install.

Also on continue, after running

dpkg --configure -a

seems to be stuck on:

Failed to reload daemon: Transport endpoint is not connected

forcing a continue results in more errors:

dpkg: dependency problems prevent configuration of systemd-timesyncd:
 systemd-timesyncd depends on systemd (= 247.3-7+deb11u1); however:
  Package systemd is not configured yet.

dpkg: error processing package systemd-timesyncd (--configure):
 dependency problems - leaving unconfigured

At anyrate even after resolving these issues, don't plan to upgrade production until after board vote is over to minimize on disruption.

[robe]

pushing to next milestone since my contract funds have been used.

[robe]

In the middle of this so site is down at moment.

[robe]

Ran into serious problems so rolled back. I think I'll just rebuild this from scratch on osgeo9 using the latest bullseye image I created. Maybe this time I'll go with nginx since wordpress/nginx performance is supposed to be better than apache/wordpress.

[robe]

I have a new container on osgeo9 called wordpress, and have the site working on there but haven't migrated yet. I want to resetup dev to use this container, kick the tires a bit before I make the osgeo9 production one live.

[robe]

I'm going to start the process shortly of moving wordpress (osgeo.org) from osgeo7 to osgeo9.

I've already built a new container on osgeo9 to house it which switches from using apache to using nginx, PHP 7.4 and is running a clean new install of bullseye.

[robe]

Okay I have it now live on osgeo9 now, and put an undermaintenance sign on the osgeo7 one. I'll keep that up for a few days before I stop the container for good. Anyone seeing the under maintenance sign, has a stale dns entry.

[cvvergara]

17. so staging will be in osgeo9 right?

[robe]

Replying to cvvergara:

17. so staging will be in osgeo9 right?

No osgeo9 is production and is called wordpress

osgeo4 (at moment has both a staging and a dev). Staging I'll probably get rid of and reset it up on the dev.

So osgeo4 wordpress-dev, wordpress-staging. Both are copies of what I had on osgeo9 from a week or so ago when you tested them.

[robe]

P.S I still need to move over the cronjobs from osgeo7, but you can start setting up what you are were talking about on the postgresql-15 I installed (which is in all of the envs).

[robe]

okay I moved over the cronjobs. While copying over people's homedrives, I found some unexpected ones filled with stuff. I copied them over, but wonder, if these should even be on this server.

cvvergara -- this one is expected wildintellect - bunch of byobu, what is this for? rouault - swig-2.0.12 (what is it for and should it even be on the wordpress server)

It's possible wildintellect and rouault were copied over from a much earlier time (when osgeo.org was housed on osgeo6).

I didn't check the dates.

[robe]

this was done a while ago. wordpress is now on osgeo9 using nginx instead of apache now.