Make error handling in parallel pg_upgrade less bogus.

reap_child() basically ignored the possibility of either an error in
waitpid() itself or a child process failure on signal.  We don't really
need to do more than report and crash hard, but proceeding as though
nothing is wrong is definitely Not Acceptable.  The error report for
nonzero child exit status was pretty off-point, as well.

Noted while fooling around with child-process failure detection
logic elsewhere.  It's been like this a long time, so back-patch to
all supported branches.
This commit is contained in:
Tom Lane 2018-12-16 14:51:48 -05:00
parent cd24b4eaec
commit f7642cf316

View File

@ -294,7 +294,7 @@ reap_child(bool wait_for_child)
{ {
#ifndef WIN32 #ifndef WIN32
int work_status; int work_status;
int ret; pid_t child;
#else #else
int thread_num; int thread_num;
DWORD res; DWORD res;
@ -304,14 +304,13 @@ reap_child(bool wait_for_child)
return false; return false;
#ifndef WIN32 #ifndef WIN32
ret = waitpid(-1, &work_status, wait_for_child ? 0 : WNOHANG); child = waitpid(-1, &work_status, wait_for_child ? 0 : WNOHANG);
if (child == (pid_t) -1)
/* no children or, for WNOHANG, no dead children */ pg_fatal("waitpid() failed: %s\n", strerror(errno));
if (ret <= 0 || !WIFEXITED(work_status)) if (child == 0)
return false; return false; /* no children, or no dead children */
if (work_status != 0)
if (WEXITSTATUS(work_status) != 0) pg_fatal("child process exited abnormally: status %d\n", work_status);
pg_fatal("child worker exited abnormally: %s\n", strerror(errno));
#else #else
/* wait for one to finish */ /* wait for one to finish */
thread_num = WaitForMultipleObjects(parallel_jobs, thread_handles, thread_num = WaitForMultipleObjects(parallel_jobs, thread_handles,