Category: Wordpress CMS

I first came across this WordPress cum Apache2 Permalink 404 problem back in 2018 and described a fix here. Unfortunately I had forgotten all about it and even though the damn fix was sitting right there in front of me I had to re-endure all the grief and suffering again. However during this grief and suffering I came across another nasty little feature that seem associated with the new Block editor. (previously I was using the classic editor) Namely after changing the permalink I got “Updating failed. The response is not a valid JSON response.” whenever I tried changing a post.

I previously wrote up a fix for the 404 problem when you change permalink style here. While it is much the same as this it is here, it is four years old and pre-dates the Block editor. But it may be worth a look along with this in case it somehow explains thing differently and more usefully for you.

This problem (in total) probably only applies to situations where you are using your own Apache2 server. But it may apply in part if you’ve not got the .htaccess file properly configured on some hosted webspace. I would imaging that if you are using a hosting service then it would be highly unlikely you would need the bit below about a2enmod.

I must add that this problem (2nd time ) almost beat me to death. It appeared to be intractable. But that was only because I had failed to configure Apache properly. (or read my own f**king notes!). I didn’t think (until now) that the two issues in Permalink 404’s & Response is not a valid JSON response” were related. But I think they are. Either that or I fluked a fix to both.

Here’s What Happened

Basically I installed WordPress with no issues. I built a child theme, again with no issues. I could build pages and posts using both the Gutenberg and the Classic editor. BUT I could only use the “plain” default URL settings for both posts and pages. If I changed it under settings→permalinks to something a bit more useful I’d end up with a 404 not found.

In the Gutenberg editor I also noticed that I often (always?) got a red bar with the following error message in it..

“Updating failed. The response is not a valid JSON response.”

So if you are a getting this then maybe it is tied up with this permalink problem. Obviously before I started using the Block editor and before I fixed the permalink problem I didn’t get it. After I fixed the permalink problem in apache2 then the JSON invalid response issue I was getting in the Block Editor disappeared too.

So, What was I getting with the unconfigured Apache?

I changed the permalink setting away from the default. Then in the editor the new URL for the post/page looked correct. But it just did not work. It always returned 404 Not Found.

Here’s what I did to solve this (for me) They were not done in this order but I think that this is the best order to do them in. It may also be that not all these steps are needed.

Here’s the Fix

1. In etc/apache2/apache2.conf: change AllowAll=none to AllowAll=All. I did this for all four settings but I think that only the one related to /var/www is necessary. You change all 4 to get it working back out the ones you don’t need by trial and error. I’m pretty sure you only need /var/www

Doing this enables the .htaccess file (more in a minute). Without doing this .htaccess file is simply not read.

2. Turn on mod_rewrite

sudo a2enmod rewrite

(this was the missing bit that killed me for two days. This MUST be done if it’s your own server)

3. Ensure the .htaccess file exists and holds the following code (or similar. In my previous post I used a slightly different incantation but never mind – the standard wordpress version is fine)

<IfModule mod_rewrite.c>

RewriteEngine On

RewriteRule .* – [E=HTTP_AUTHORIZATION:%{HTTP:Authorization}]

RewriteBase /

RewriteRule ^index\.php$ – [L]

RewriteCond %{REQUEST_FILENAME} !-f

RewriteCond %{REQUEST_FILENAME} !-d

RewriteRule . /index.php [L]

</IfModule>

I gather it is possible to completely dispense with the htaccess file and simply do everything in apache2.conf. But most people only host a site to develop it and then upload it to a host provider where use of a htaccess file is not optional. So you might as well do it in your prototype code too.

Make sure you htaccess file is writable by www-data as well as the owner.

Restart apache2 after the mods. with sudo service apache2 restart

(or whatever other incantation you like)

That’s it.

Hopefully the permlinks should now work and those hideous Permalink 404’s & Response is not a valid JSON response messages will be history too.

Yes. I committed the cardinal sin. I lost my root password to mysql and phpmyadmin. Consequently I could not log in as root in order to change the permissions for a user. Of course I hadn’t noticed I had lost the damn thing because normally I never-ever log in as root to anything.

It was only when I need to fiddle about with other users when I typed in what I thought it was – only to find my memory was crap. Of course, I had written it down somewhere safe. I always do. The problem was (crap memory again) I couldn’t find where.

I always obfuscate saved passwords – just in case some nasty person raids my desk and steals my documents. So it is probably staring me in the face. On the plus side I suppose it shows my obfuscation works pretty well. Hey Ho.

Check Your Mysql Version

Here’s what I did to reset the lost root password to mysql and phpmyadmin. This worked for me. I’m running mysql ver 8.0.30 Check this with:
mysql -V
Mine come back with this.
mysql Ver 8.0.30-0ubuntu0.20.04.2 for Linux on x86_64 ((Ubuntu))
This may not be compatible with all other versions but I expect it should be with this or later versions.

This method (after trying and failing with many others) is based largely around this post on devanswers.co. So major Hat Tip to them. There are a few extra steps here mainly surrounding clean up before actually changing the password. But if you are running mysql 5 you would probably be better off at devanswers as they deal with that earlier version as well as mysql 8. But you may still want to do the mysql killing and clean up as described here.

If, like me you got to this point after many failed attempts to fix this we need to check what sort of dismal state we are in.

We need to kill mysql. Make absolutely sure it really is dead. It’s worth checking that you haven’t got multiple versions running. This may (especially) happen if you’ve been faffing around with other failed attempts to crack in to mysql (like I had).

Cleanup: Kill all Mysqls

sudo ps -A | grep mysql – if this shows you one or more mysql or mysqld processes running then….
sudo pkill mysql
sudo pkill mysqld
sudo ps -A | grep mysql – make sure this time nothing comes back.

Now we have to make sure the following directory (and associated user) exists. It may well not.

If you try and push ahead without having this directory properly set you will get something like this:
chalkie@chalkies-box:~$ mysql -u root mysql
ERROR 2002 (HY000): Can’t connect to local MySQL server through socket ‘/var/run/mysqld/mysqld.sock’ (2)
[1]+ Exit 1 sudo mysqld –skip-grant-tables

If you already have this error message START AGAIN. Go back and check ALL instances of mysql are dead THEN create the directory as below.

ls -l /var/run/mysqld
If it is not there and the owner is not mysqld:mysqld then…
sudo mkdir /var/run/mysqld
sudo chown mysql /var/run/mysqld
check it again with
ls -l /var/run/mysqld

Now restart mysql in safe mode like so..
sudo mysqld_safe –skip-grant-tables&
You’ll get something like this back..
[1] 100757
chalkie@chalkies-box:/var/run$ 2022-08-21T12:57:00.853461Z mysqld_safe Logging to ‘/var/log/mysql/error.log’.
2022-08-21T12:57:00.907704Z mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql

now we can login as root like so…
sudo mysql –user=root mysql

You now should be logged in as root the prompt should now be mysql>

Now you are logged in as Root

From here you can now do all sorts of damage. Or even change the lost password for root as so..

First remove the old password with this…(notice the trailing semi colons and ignore the mysql> prompt)
mysql>UPDATE mysql.user SET authentication_string=null WHERE User=’root’;

mysql>flush-privileges;

mysql>ALTER USER ‘root’@’localhost’ IDENTIFIED WITH mysql_native_password BY ‘a-new-pw-i-promise-never-to-forget’;
mysql>flush-privileges;

And that’s it. Log out and kill the safe mode mysql process

mysql>exit;
sudo killall -u mysql

Now restart it:
sudo /etc/init.d/mysql start
CHECK the new password for root works by logging into mysql and/or phpmyadmin with it!!!

I hope that helps. My other Ubuntu style fixes are listed HERE