From: Magnus Hagander Date: Thu, 7 May 2015 13:04:13 +0000 (+0200) Subject: Properly send SCM status updates when shutting down service on Windows X-Git-Tag: REL9_4_2~30 X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=43ed06816f8d9c591ca767a19d555044648cb8a2;p=postgresql Properly send SCM status updates when shutting down service on Windows The Service Control Manager should be notified regularly during a shutdown that takes a long time. Previously we would increaes the counter, but forgot to actually send the notification to the system. The loop counter was also incorrectly initalized in the event that the startup of the system took long enough for it to increase, which could cause the shutdown process not to wait as long as expected. Krystian Bigaj, reviewed by Michael Paquier --- diff --git a/src/bin/pg_ctl/pg_ctl.c b/src/bin/pg_ctl/pg_ctl.c index 91e99e4c20..f767f2e697 100644 --- a/src/bin/pg_ctl/pg_ctl.c +++ b/src/bin/pg_ctl/pg_ctl.c @@ -1590,15 +1590,27 @@ pgwin32_ServiceMain(DWORD argc, LPTSTR *argv) switch (ret) { case WAIT_OBJECT_0: /* shutdown event */ - kill(postmasterPID, SIGINT); + { + /* + * status.dwCheckPoint can be incremented by + * test_postmaster_connection(true), so it might not + * start from 0. + */ + int maxShutdownCheckPoint = status.dwCheckPoint + 12;; - /* - * Increment the checkpoint and try again Abort after 12 - * checkpoints as the postmaster has probably hung - */ - while (WaitForSingleObject(postmasterProcess, 5000) == WAIT_TIMEOUT && status.dwCheckPoint < 12) - status.dwCheckPoint++; - break; + kill(postmasterPID, SIGINT); + + /* + * Increment the checkpoint and try again. Abort after 12 + * checkpoints as the postmaster has probably hung. + */ + while (WaitForSingleObject(postmasterProcess, 5000) == WAIT_TIMEOUT && status.dwCheckPoint < maxShutdownCheckPoint) + { + status.dwCheckPoint++; + SetServiceStatus(hStatus, (LPSERVICE_STATUS) &status); + } + break; + } case (WAIT_OBJECT_0 + 1): /* postmaster went down */ break;