Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commitbf26179

Browse files
committed
Fix various bugs in postmaster SIGKILL processing
Clamp the minimum sleep time during immediate shutdown or crash to aminimum of zero, not a maximum of one second. The previous code couldresult in a negative sleep time, leading to failure in select() calls.Also, on crash recovery, reset AbortStartTime as soon as SIGKILL is sentor abort processing has commenced instead of waiting until the startupprocess completes. Reset AbortStartTime as soon as SIGKILL is sent,too, to avoid doing that repeatedly.Per trouble report from Jeff Janes onCAMkU=1xd3=wFqZwwuXPWe4BQs3h1seYo8LV9JtSjW5RodoPxMg@mail.gmail.comAuthor: MauMau
1 parent2d6c0f1 commitbf26179

File tree

1 file changed

+10
-5
lines changed

1 file changed

+10
-5
lines changed

‎src/backend/postmaster/postmaster.c

Lines changed: 10 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -1422,9 +1422,9 @@ DetermineSleepTime(struct timeval * timeout)
14221422
{
14231423
if (AbortStartTime>0)
14241424
{
1425-
/*remainingtime, but at least 1 second */
1426-
timeout->tv_sec=Min(SIGKILL_CHILDREN_AFTER_SECS-
1427-
(time(NULL)-AbortStartTime),1);
1425+
/* time left to abort; clamp to 0 in case it already expired */
1426+
timeout->tv_sec=Max(SIGKILL_CHILDREN_AFTER_SECS-
1427+
(time(NULL)-AbortStartTime),0);
14281428
timeout->tv_usec=0;
14291429
}
14301430
else
@@ -1676,10 +1676,13 @@ ServerLoop(void)
16761676
* Note we also do this during recovery from a process crash.
16771677
*/
16781678
if ((Shutdown >=ImmediateShutdown|| (FatalError&& !SendStop))&&
1679+
AbortStartTime>0&&
16791680
now-AbortStartTime >=SIGKILL_CHILDREN_AFTER_SECS)
16801681
{
16811682
/* We were gentle with them before. Not anymore */
16821683
TerminateChildren(SIGKILL);
1684+
/* reset flag so we don't SIGKILL again */
1685+
AbortStartTime=0;
16831686

16841687
/*
16851688
* Additionally, unless we're recovering from a process crash, it's
@@ -2584,7 +2587,7 @@ reaper(SIGNAL_ARGS)
25842587
* Startup succeeded, commence normal operations
25852588
*/
25862589
FatalError= false;
2587-
AbortStartTime=0;
2590+
Assert(AbortStartTime==0);
25882591
ReachedNormalRunning= true;
25892592
pmState=PM_RUN;
25902593

@@ -3544,6 +3547,8 @@ PostmasterStateMachine(void)
35443547
StartupPID=StartupDataBase();
35453548
Assert(StartupPID!=0);
35463549
pmState=PM_STARTUP;
3550+
/* crash recovery started, reset SIGKILL flag */
3551+
AbortStartTime=0;
35473552
}
35483553
}
35493554

@@ -4737,7 +4742,7 @@ sigusr1_handler(SIGNAL_ARGS)
47374742
{
47384743
/* WAL redo has started. We're out of reinitialization. */
47394744
FatalError= false;
4740-
AbortStartTime=0;
4745+
Assert(AbortStartTime==0);
47414746

47424747
/*
47434748
* Crank up the background tasks. It doesn't matter if this fails,

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp