1berry/MariaDB-server
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

MDEV-30836 MTR hangs after tests have completed

Browse Source 
The problem is in manager/worker communication when worker sends
WARNINGS and then TESTRESULT. If manager yet didn't read WARNINGS
response both responses get into the same buffer, can_read() will
indicate we have data only once and we must read all the data from the
socket at once. Otherwise TESTRESULT response is lost and manager
waits it forever.

The fix now instead of single line reads the socket in a loop. But if
there is only one response in the buffer the second read will be
blocked waiting until new data arrives. That can be overcame by
blocking(0) which sets the handle into non-blocking mode. If there is
no data second read just returns undef.

The problem is non-blocking mode is not supported by all perl flavors
on Windows. Strawberry and ActiveState do not support it. Cygwin and
MSYS2 do support. There is some ioctl() hack that was known to "work"
but it doesn't do what is expected (it does not return data when there
is data). So for Windows if it is not Cygwin we disable the fix.
This commit is contained in:
Aleksey Midenkov 2023-08-07 18:49:47 +03:00 committed by Daniel Black
parent 848b3af816
commit a49b9314c1
1 changed files with 34 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

54
mysql-test/mysql-test-run.pl
View File

@ -643,7 +643,7 @@ sub parse_protocol($$) {
{
mtr_verbose2("${line}: saying BYE to ". $names{$sock});
print $sock "BYE\n";
return 1;
return 2;
}
return ["Failure", 1, $completed, $extra_warnings];
}
@ -692,7 +692,7 @@ sub parse_protocol($$) {
$result->write_test($sock, 'TESTCASE');
}
push(@$completed, $result);
return 1;
return 2;
}
}
@ -706,7 +706,7 @@ sub parse_protocol($$) {
delete($result->{result});
$result->{repeat}= $repeat+1;
$result->write_test($sock, 'TESTCASE');
return 1;
return 2;
}
# Remove from list of running
@ -779,7 +779,7 @@ sub run ($$$) {
}
main::mark_time_idle();
my $i= 0;
foreach my $sock (@ready) {
sock_loop: foreach my $sock (@ready) {
++$i;
if ($sock == $server) {
# New client connected
@ -790,23 +790,37 @@ sub run ($$$) {
print $child "HELLO\n";
}
else {
my $line= <$sock>;
if (!defined $line) {
# Client disconnected
--$childs;
mtr_verbose2((exists $names{$sock} ? $names{$sock} : "Worker"). " closed socket (left ${childs} childs)");
$s->remove($sock);
$sock->close;
next;
}
chomp($line);
mtr_verbose2("Connection ${i}". (exists $names{$sock} ? " from $names{$sock}" : "") .": $line");
my $j= 0;
$sock->blocking(0);
while (my $line= <$sock>) {
++$j;
chomp($line);
mtr_verbose2("Connection ${i}.${j}". (exists $names{$sock} ? " from $names{$sock}" : "") .": $line");
my $res= parse_protocol($sock, $line);
if (ref $res eq 'ARRAY') {
return @$res;
} elsif ($res == 1) {
next;
$sock->blocking(1);
my $res= parse_protocol($sock, $line);
$sock->blocking(0);
if (ref $res eq 'ARRAY') {
return @$res;
} elsif ($res == 1) {
next;
} elsif ($res == 2) {
next sock_loop;
}
if (IS_WINDOWS and !IS_CYGWIN) {
# Strawberry and ActiveState don't support blocking(0), the next iteration will be blocked!
# If there is next response now in the buffer and it is TESTRESULT we are affected by MDEV-30836 and the manager will hang.
last;
}
}
$sock->blocking(1);
if ($j == 0) {
# Client disconnected
--$childs;
mtr_verbose2((exists $names{$sock} ? $names{$sock} : "Worker"). " closed socket (left ${childs} childs)");
$s->remove($sock);
$sock->close;
next;
}
# Find next test to schedule