CPU, זיכרון ו-Swap ב-100%

סיכום מבוסס בינה מלאכותית

A user reports consistently high CPU, memory, and swap usage on their Digital Ocean droplet hosting a Discourse forum, resulting in HTTP 500 errors. Despite low user activity, the 1 vCPU, 2GB RAM droplet is under heavy load. The user attempted to upgrade Discourse but was unable to complete the process due to system issues. After checking logs, they identified SSL-related SMTP errors and resolved the problem by deleting retry tasks in Sidekiq and upgrading via the shell using ./launcher rebuild app tnolin, itsbhanusharma, gerhard. The system performance improved significantly, though a new error related to Net::ReadTimeout has emerged.

Recently, I’ve been getting a HTTP 500 Error on my Discourse forum. I’m hosted on a Droplet with Digital Ocean and checked to see that my CPU, DISK, and Memory are constantly at 100%.

Restarting the droplet seems to fix this issue temporarily and able to access the forums, but unsure what is causing so much stress on the system, as we don’t get a lot of track on the forum.

I went into the console and found that our Memory Swap is also maxed out and I can’t identify what the problem is. Any suggestions on where to look?

Have your forum started getting a lot of user activity recently? If the CPU is consistently being hammered to 100% it could indicate there are a bunch of background items running. What are the Specs of your droplet? have you considered upgrading the droplet to see if that is what you need?

Hi,

No we haven’t had any surge of users or activity at all. It’s been quite inactive for some time which is why I am puzzled that the droplet is operating at this capacity.

It’s a 1 vCPU, 2GB RAM, 50GB Disk, which has been more than sufficient when it was initially launched and there was much more user activity.

I’d say maybe try upgrading discourse to latest or dig into the logs to see if you find your problem.

Trying to upgrade the docker manager first but the issue with the server is preventing it from completing (I let it run over the weekend and it still didn’t upgrade.

Edit: Looked at the logs and see some issues:

Try upgrading in the shell:

cd /var/discourse
./launcher rebuild app

upgrading the shell and deleting a bunch of retry tasks in Sidekiq (SSL error with SMTP not matching website) has made it drop drastically.

Now get a new error that I will work on.

Jobs::HandledExceptionWrapper: Wrapped Net::ReadTimeout: Net::ReadTimeout with #TCPSocket:(closed)