petecooper Wrote:
-------------------------------------------------------
> Hello.
> I have a single-digit fleet of Ubuntu servers, all running a similar
> configuration:
>
> * Ubuntu 20.04LTS, current kernel via `apt`
> * Nginx 1.23.1 from source, with 3rd party modules
> * PHP 8.0 or 8.1 from source
>
> One of the servers has recently had issues with what appears to be an
> Nginx segfault, which manifests as high CPU load, repeated `apport`
> processes spawning, and errors written to `apport.log`. The other
> servers have a similar config, but are not affected by this issue.
I am providing some follow up for awareness / completeness. According to my logs, the server was upgraded from Nginx 1.23.0 on 2022-07-18, and then upgraded to Nginx 1.23.1 on 2022-08-10:
2022-03-16T16:15:33+00:00 Nginx 1.21.6 compiled
2022-07-18T10:45:47+00:00 Nginx 1.23.0 compiled
2022-08-10T07:04:58+00:00 Nginx 1.23.1 compiled
2022-08-13T08:48:32+00:00 Nginx 1.23.1 compiled
This issue first appeared on 2022-08-11 (a reboot resolved it), then 2022-08-12 at the same time of day, and again for 2022-08-13 & 2022-08-14. My provider graphs show the issue occurs at around 07:00 each day, which - perhaps coincidentally - is the same time that the Nginx 1.23.1 build completed.
I am rolling back to Nginx 1.23.0 to see if that addresses or affects the issue, I will report back with findings after 07:00 tomorrow.
Thank you for your continued attention and expertise, it is greatly appreciated.