version 4.4 update: FATAL ERROR!

Hey James, as for
“Excellent, so now we have two reports out of now almost 2.5 million installations, one with 1 site, and the other with 100 sites.”
and similar earlier remarks you made,
I MUCH appreciate you volunteers, just to clarify:

Maybe one in a Million (?) users (of anything!) report their problem. The number is unknown, the relation is a fact.
If you hear of ONE problem of anything, accept the fact that there will be hundreds of thousands similar problems, given the size of wp!

Exception: Someone has a non-standard wp installation.

—
Now here’s what solved it, seemingly: I too, like 2001web finally decided to give it another go after securing a new update (can’t know how much crash the “FATAL ERROR” means he did, right?).
THAT update ran without that bold error message, and then prompted me to “update your database”. I clicked once more, and since then all seems fine. (seems, cause I can’t know, can I?)

This is why I started this thread with “Much appreciated that someone(?) will look at v4.4 update process. Ideally (next time) BEFORE rolling it out to everyone?”

Thanks again James!

Excellent, I’m glad it worked!

Depending on your PHP settings, there is a possibility that new PHP files are not being evaluated, which will result with the errors mentioned, such as:

Call to undefined function wp_oembed_add_host_js()
Call to undefined method WP_Screen::set_screen_reader_content()

This will happen, for instance, if your opcache settings in PHP 5.6+ are set like this:

; When disabled, you must reset the OPcache manually or restart the
; webserver for changes to the filesystem to take effect.
opcache.validate_timestamps=0

As an example, if you are using Debian 8 and this was done to improve the performance of your website, you should restart Apache if you’re using mod_php, or PHP5-FPM if you’re using nginx or FCGI with Apache:
# systemctl restart php5-fpm

After this, if you try to load the /wp-admin again, you will be prompted to upgrade your database.

In this case, of course, the problem wasn’t actually with WordPress.

Thanks, Pacoup!

I can confirm we have these running smoothly now. Issues seem to have been related to php version.

Due to a few clients running older versions of other scripts on the server we had delayed updating. I had techs install cloud linux to our server so that php version could be selected by user.

If we upgrade the wordpress users to latest version of php the upgrades run very smoothly. Just thought I would share.

The php version 5.4 was the release causing WP update to have issues.

Interesting, WordPress supports 5.2.4, so there might have been a misconfiguration with 5.4, similar to what Pacoup describes above.

Just got this error too on my dev machine. I also think this is a cache issue related to php-fpm. My dev machine’s php-fpm instance doesn’t refresh the cached opcode immediately when the files updated (it’s lagged for a bit), and restarting php-fpm (or waiting for a bit until it reload the code by itself) fixed the issue. This doesn’t happen on my production machine (turns out my dev machine somehow has opcache.revalidate_freq option set to 60, no wonder it took a while).

Excellent, thanks for sharing that!

Chiming in to share I have this error pop up from time to time when updating core. It is usually due to w3 total cache and/or cloudflare. Mostly a minor annoyance, when I reload the page all is well.