Strange Unit failure

[Andrew Clayton]

On Tue, 20 Jun 2023 09:31:17 +0000
Peter TKATCHENKO <p.tkatchenko at bimp.fr> wrote:

> Hello,

Hi,

[...]

> External access to the application is proxied by HAProxy with sticky
> sessions configured. Direct access to NGINX servers is possible too
> (with another URLs).

You mention proxies. Are you using the "proxy" setting in your config?

> Yesterday we began to (randomly) get 503 errors on different pages of
> the application. The same page could be OK for one user, but failed
> with 503 error for other user. The source of these errors was Unit. We

Are the error cases using WebSockets?

> could isolate the problems and solved them restarting Unit service on
> 2 servers (of 3). We did not find any useful information in Unit logs.
> Nothing interesting in NGINX logs neither.
> 
> I suppose that some Unit workers were in strange semi-failured state,
> but the master process did not kill them and continued to send them
> the requests.
> 
> Any ideas?

Could you confirm if you are or aren't seeing actual crashes in the unit
processes?

Do you see process exited on signal 11 (or maybe 6) messages in the unit
log for example or core dumps being generated?

> Best regards,
> 
> Peter

Cheers,
Andrew