BLC version 2.4.9 causing issues serverwide

I run CloudLinux 8 / CloudLinux 9 servers with cPanel, MySQL or MariaDB, and alt-php, some with MySQL Governor and some without.

On any server where BLC was autoupdated on August 5th to version 2.4.9, I’m finding that MySQL/MariaDB resources are being exhausted when a BLC process locks a table and then some other BLC process(es) spin of a ton of ALTER statements. Because the table is locked, the ALTER statements cannot be processed. This builds up connections and outstanding statements in MySQL / MariaDB that cannot be processed. This eventually leads to MySQL / MariaDB exhaustion.

Basically any of the 25 or so sites that updated to 2.4.9 on August 5th can cause this problem to occur, effectively causing a denial of service to MySQL and rendering all the sites on the server unable to connect to MySQL / MariaDB.

I do not have information from “mysqladmin processlist’ or full process list to show you since I had to go through and disable the plugin on every user’s site that updated to ensure that it doesn’t cause a DoS on my servers. My customer’s website availability is #1, and I cannot / will not attempt to provides logs when it happens on production servers. I remediate the problem and then report the problem.

I just want somebody to understand that in a shared hosting environment where there are one or 10s or 100s of sites running 2.4.9 version of the BLC plugin, it is entirely possible for just one site to exhaust MySQL/MariaDB resources. I had to deal with 6 outages caused by BLC 2.4.9 on six separate sites yesterday until I was able to determine specifically what was going wrong and disable those plugins.