mayaliny

https://www.wordfence.com/r/3dfcd74325221057/plugin/registration

redirects to

https://www.wordfence.com/plugin/registration?utm_source=506&utm_campaign=3dfcd74325221057&utm_medium=affiliate

That gives 404

@wfpeter besides the “central” part, the url that is called from plugin activation when it says “get a license” gives 404.. that is blocking us to install licenses the usual way.
I had installed a ton of them so something is broken, i think since last release.

I’m having the same problem.

Also.. if i try to add a new site from wordfence central, it says “license activated” but the license key it’s empty and on the site admin it keeps saying the license nees to be activated.

We fixed our css in order to fix the problem. Thank you!

@tessawatkinsllc i’ve updated the plugin to latest version and the problem is fixed for me! thank you! you did an excellent and fast work!

@tessawatkinsllc new topic is here https://www.ads-software.com/support/topic/version-3-3-0-breaking/

Hey.. this is apparently breaking my install.. with version 3.3.0 url passed from:

https://www.argenteriadabbene.com:443/it/prodotto/ricci-ascot/
to
https://argenteriadabbene.com/it:443/it/prodotto/ricci-ascot/

That /it that comes twice is breaking the link.

Why using network_home_url() ? i’m not on a network at all.

I’m using qtranslateXT for language management and website is served by nginx as a proxy on port 80 and apache on port 443 (ssl).

i temporary solved by modifying the file (link 96) you mentioned with

$url = home_url($_SERVER[‘REQUEST_URI’]);

Bumping the topic since i’m again experiencing memory exhausted error every 10 mins..

current memory_limit is 768 and i’m going to raise it but that’s not a solution..

A final note: when we bumped the mem to 1024M and re-ran the query I was monitoring system resources with htop and observed no increase in memory usage as the query ran.

Thank You in Advance,

M

I failed to mention one detail: the usermeta table for this site is only 96M, so why would any query need ~1024M of mem to run ?

M

@faisalahammad,

Thank You for your response. Mark here on behalf of mayaliny.
Just wanted to share some info about the env we tested against that the site in
question is running on.
Ubuntu 2.0.4.5 LTS, nginx 1.23.2-1 terminating SSL and proxying to an apache 2.4.54-1 backend, mariadb 1:10.6.11 server / client.

We did our testing on a dev server with the same env as production with the only real difference between the two envs being the server spec: prod has 4vcpus and 8GB of RAM and dev has 2vcpus and 4GB of RAM. These are cloud servers. I personally doubt that the server specs had anything to do with the results of our testing.
Anyway I / we approached this differently only in that I made the memory changes to the php-fpm php.ini file rather than altering the wp-config.php file.
The WP site was backed up and restored using Updraft so everything was fresh in the dev env. I didn’t make any other changes to the server env besides that php.ini file.

We started with an increase to 512M. Still mem allocation errors. 768M, same. Then to 1024M and the allocation errors were gone. Not a huge surprise but what struck me as odd though was that after that free memory on the server actually increased while there was no change the buffer/cache usage. Was if that new memory limit was not even used temporarily.

I could be way off base, but what it looked like was that query didn’t actually need that allocated memory but rather was just checking if it could get it _if_ it could.
In this case update_meta_cache (I think), which is apparently the culprit function that is reading / loading the aforementioned list of 14000+ users from the usermeta table then blowing up until we bumped the RAM.

I can and will say I trust mayaliny’s assessment that this is some hinky query code behind this problem or he would not have suggested as much, and feel confident this is not a server-side issue, so any other qualifide or otherwise helpful input or pointers are welcome from all.

Cheers,

Mark

• This reply was modified 2 years, 2 months ago by mayaliny.
• This reply was modified 2 years, 2 months ago by mayaliny.

Hello.. unfortunately we already bumped memory and saw that at least 1GB is needed in order to complete the (useless?) task. Of course i can’t give 1GB of memory to that process only because someone didn’t wrote queries well ??

Anyway increasing memory is just a patch and not a solution, downloading every 10 minutes the whole list of users, loop it and the ask the DB all the metadata from those 14000+ users with an insane query and the loop it again doesn’t seem acceptable to me.

I wonder if some “core guy” can read this and analyze the functions i pointed out.

Also being able to totally disable this “all_users_meta_cache” would be probably a good idea on sites with large amount of users and the fastest thing to solve my problem. Is there a way?

• This reply was modified 2 years, 2 months ago by mayaliny.

Log in with different ip and stop scan. only ip who launched the scan can’t see the site

We found out that when you run the test, your IP is doomed until someone with different ip stops the scan.

Only person who launched the test gets error 500, site is functioning normally for others.

Guess it’s related to WP 5.5 suppressing Jquery Migrate by default..

https://www.searchenginejournal.com/wordpress-5-5-issues/377851/

we need to change plugin since this is not being updated since 5 years, meanwhile i applied the suggested solution, thank you!