Report the service is ready in post_config, which could be during * startup or after a reload. The server could still hit a fatal * startup error after this point during ap_run_mpm(), so this is * perhaps too early, but by post_config listen() has been called on * the TCP ports so new connections will not be rejected. There will * always be a possible async failure event simultaneous to the
| 71 | * always be a possible async failure event simultaneous to the |
| 72 | * service reporting "ready", so this should be good enough. */ |
| 73 | static int systemd_post_config(apr_pool_t *pconf, apr_pool_t *plog, |
| 74 | apr_pool_t *ptemp, server_rec *main_server) |
| 75 | { |
| 76 | if (ap_state_query(AP_SQ_MAIN_STATE) == AP_SQ_MS_CREATE_PRE_CONFIG) |
| 77 | return OK; |
| 78 | |
| 79 | #ifdef HAVE_SELINUX |
| 80 | log_selinux_context(); |
| 81 | #endif |
| 82 | |
| 83 | sd_notify(0, "READY=1\n" |
| 84 | "STATUS=Configuration loaded.\n"); |
| 85 | return OK; |
| 86 | } |
| 87 | |
| 88 | static int systemd_pre_mpm(apr_pool_t *p, ap_scoreboard_e sb_type) |
| 89 | { |
nothing calls this directly
no test coverage detected