• Welcome to Peterborough Linux User Group (Canada) Forum.
 

Thanks Jason, for getting the PLUG website back up and running

Started by ssfc72, May 24, 2024, 06:41:52 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

ssfc72

Mint 20.3 on a Dell 14" Inspiron notebook, HP Pavilion X360, 11" k120ca notebook (Linux Lubuntu), Dell 13" XPS notebook computer (MXLinux)
Cellphone Samsung A50, Koodo pre paid service

Jason

Thanks, Bill. I appreciate that.

The first thing I noticed was that the server load was heavy. And by heavy, I mean the load average was 40x what it should be AT MAX! The CPU was running at 100% and the RAM was near maxing out. The database, which the forum relies on, was the culprit. Why that was, I wasn't entirely sure. I'd restart the database service, things would improve but only briefly. Then it'd be insane again.

Cod3poet helped me walk through some steps and gave me some suggestions. Where I ended up, I should have started. In the SMF logs, one error came up repeatedly, a total of 7700+ instances over 7 hours. Yes, you read that right, over 1000 database read errors per hour! I believe it was causing the issue.  I've attached it so you can view it. It's tiny unless you click on it.

As you can guess, it checks a viewer's IP address to give a total of the unique users online at any given time. You see that total at the bottom of the main SMF page. So every time someone visited, it was choking on this issue.

Fun, huh? I still don't understand the first cause but it works. And that's more than we can say about most things today.



* Zorin OS 17.1 Core and Windows 11 Pro on a Dell Precision 3630 Tower with an
i5-8600 3.1 GHz 6-core processor, dual 22" displays, 16 GB of RAM, 512 GB Nvme and a Geforce 1060 6 GB card
* Motorola Edge (2022) phone with Android 13