lovedigit
👽 Chevereto Freak
I noticed an issue on my server recently. It appears to be linked to my chevereto website because it happens only when large images are uploaded by users.
You can replicate this issue on your side if you turn off your swap partition temporarily. Upload an image around 10MB or larger in size. It appears chevereto doesn't always free up the memory that it used to resize the image. In my case when I tested it personally after some complains from the users about the "internal server error" messages on the frontend, I figured out that my VPS with 2GB Ram would easily get swamped when there were few large images in the queue. I tried configuring the php and nginx settings to minimum, but it appears to be the issue with image resize, or clean up after the process. If it doesn't have to do with the memory, then I would like to report random "internal server error" on my chevereto website which show up without any pattern. After further investigation in case I thought it doesn't have to do with the Ram, the backend log suggests that the connection was closed by the client before upstream could send the data. I tried increasing the execution time, keepalive time, fastcgi read timeout and connection time but it is the same.
While trying to replicate the error reported by some users, I noticed that the memory usage would hit to 1.2+ GB for 15MB image and it would process successfully for first few images, but since it didn't free the RAM, the consequent files throws error. I can sure upgrade my server, but it doesn't seem to be worth as most of the time the server would stay idle with less 5% usage. The script should be able to free up the memory at the first place. I am open to all suggestions, don't take it as rant.
You can replicate this issue on your side if you turn off your swap partition temporarily. Upload an image around 10MB or larger in size. It appears chevereto doesn't always free up the memory that it used to resize the image. In my case when I tested it personally after some complains from the users about the "internal server error" messages on the frontend, I figured out that my VPS with 2GB Ram would easily get swamped when there were few large images in the queue. I tried configuring the php and nginx settings to minimum, but it appears to be the issue with image resize, or clean up after the process. If it doesn't have to do with the memory, then I would like to report random "internal server error" on my chevereto website which show up without any pattern. After further investigation in case I thought it doesn't have to do with the Ram, the backend log suggests that the connection was closed by the client before upstream could send the data. I tried increasing the execution time, keepalive time, fastcgi read timeout and connection time but it is the same.
While trying to replicate the error reported by some users, I noticed that the memory usage would hit to 1.2+ GB for 15MB image and it would process successfully for first few images, but since it didn't free the RAM, the consequent files throws error. I can sure upgrade my server, but it doesn't seem to be worth as most of the time the server would stay idle with less 5% usage. The script should be able to free up the memory at the first place. I am open to all suggestions, don't take it as rant.
Last edited: