Hi, While doing a load test and killing (-9) backend server it appears that ap_proxy_retry_worker() marks workers for retry and it shouldn't.
I think it is because we have everywhere:
+++
worker->s->status |= PROXY_WORKER_IN_ERROR;
worker->s->error_time = apr_time_now();
+++
And we should have:
+++
worker->s->error_time = apr_time_now();
worker->s->status |= PROXY_WORKER_IN_ERROR;
+++
so ap_proxy_retry_worker will retry correctly.
Comments?
Cheers
Jean-Frederic
