MySQL Bugs: #45771: AIX and i5/OS Perl bug: check_socket_path

Bug #45771	AIX and i5/OS Perl bug: check_socket_path_length in MTR fails
Submitted:	26 Jun 2009 0:17	Modified:	19 Dec 2009 0:12
Reporter:	Timothy Smith	Email Updates:
Status:	Closed	Impact on me:	None
Category:	Tools: MTR / mysql-test-run	Severity:	S7 (Test Cases)
Version:	any new MTR	OS:	IBM AIX (AIX and i5/OS)
Assigned to:	Bjørn Munch	CPU Architecture:	Any

Description:
The i5/OS mtr runs are failing with this symptom:

...
vardir: /home/mysqldev/tmp-200905141024-5.1.35-11475/i5os-power-32bit/test/mysql-test/var
Removing old var directory...
Creating var directory '/home/mysqldev/tmp-200905141024-5.1.35-11475/i5os-power-32bit/test/mysql-test/var'...
mysql-test-run: *** ERROR: Socket path '/tmp/ho5MPUy6Y9' too long, it would be  truncated and thus not possible to use for connection to  MySQL Server. Set a shorter with --tmpdir=<path> option

The root cause of this can be reproduced with the following simple test case:

sock=/tmp/test-sock.$$
perl -MIO::Socket::UNIX -e "warn IO::Socket::UNIX->new(Local => '$sock',
Listen => 1)->hostpath();"
rm -f $sock

The perl command will result in:

Bad arg length for Socket::unpack_sockaddr_un, length is 25, should be 106
at /QOpenSys/perl/lib/5.8.0/aix/Socket.pm line 380.

It should print out:

/tmp/test-sock.3576 at -e line 1.

So, this is a bug in the default Perl on AIX and i5/OS.  Upgrading Perl didn't seem to help either, although there is some Perl version which doesn't exhibit this problem (see /usr/local/perl/bin/perl on aix.mysql.com, and the aix53 release build test runs for, e.g., 5.1.36).

How to repeat:
Try to run a current MTR (e.g., from mysql-5.1) on an AIX or i5/OS host.

Or, see small perl snippet above which shows the base problem (in Perl's IO::Socket::Unix module).

Also, notice the failures on i5os at

http://clustra.norway.sun.com/~bteam/status/mysql-5.1.36/builds.html

Suggested fix:
Since fixing this appears to require a custom Perl build, I think it would be best to add a workaround for this in mysql-test/lib/My/Platform.pm.  Something like this (untested) patch should be good:

=== modified file 'mysql-test/lib/My/Platform.pm'
--- mysql-test/lib/My/Platform.pm       2009-02-25 09:32:13 +0000
+++ mysql-test/lib/My/Platform.pm       2009-06-26 00:10:46 +0000
@@ -126,8 +126,21 @@ sub check_socket_path_length {
     die "Could not create UNIX domain socket: $!"
       unless defined $sock;
 
+    my $hostpath = eval {$sock->hostpath()};
+    if ($@) {
+      die unless $@ =~ /^Bad arg length for Socket::unpack_sockaddr_un/;
+
+      # Bug on AIX and i5/OS Perl IO::Socket::UNIX which dies with something
+      # like:
+      #   Bad arg length for Socket::unpack_sockaddr_un, length is 25,
+      #   should be 106 at /path/to/perl/lib/5.8.0/aix/Socket.pm line 380.
+      #
+      # Just fake it that everything is fine
+      $hostpath = $testfile;
+    }
+
     die "UNIX domain socket path was truncated"
-      unless ($testfile eq $sock->hostpath());
+      unless ($testfile eq $hostpath);
 
     $truncated= 0; # Yes, it worked!

Also, it would be really nice if unexpected errors didn't get swallowed up here.  Something like this would work for that:

=== modified file 'mysql-test/lib/My/Platform.pm'
--- mysql-test/lib/My/Platform.pm       2009-02-25 09:32:13 +0000
+++ mysql-test/lib/My/Platform.pm       2009-06-26 00:14:27 +0000
@@ -126,13 +126,16 @@ sub check_socket_path_length {
     die "Could not create UNIX domain socket: $!"
       unless defined $sock;
 
-    die "UNIX domain socket path was truncated"
+    die "TRUNCATED"
       unless ($testfile eq $sock->hostpath());
 
     $truncated= 0; # Yes, it worked!
 
   };
 
+  die "Unexpected failure when checking socket path lenght: $@"
+    if $@ and $@ !~ /^TRUNCATED/;
+
   $sock= undef;  # Close socket
   rmtree($tmpdir); # Remove the tempdir and any socket file created
   return $truncated;

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/79976

2811 Bjorn Munch	2009-08-04
      Bug #45771 AIX and i5/OS Perl bug: check_socket_path_length in MTR fails
      Bug is actually in Perl
      Fixed by trapping and ignoring error from IO::Socket::UNIX

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/80121

2811 Bjorn Munch	2009-08-05
      Bug #45771 AIX and i5/OS Perl bug: check_socket_path_length in MTR fails
      Bug is actually in Perl
      Fixed by trapping and ignoring error from IO::Socket::UNIX

Pushed to azalea-mtr and 5.1-mtr

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/80144

2812 Bjorn Munch	2009-08-05
      Bug #45771 AIX and i5/OS Perl bug: check_socket_path_length in MTR fails
      Bug in Perl
      Scrap attempt to do this smartly on AIX, just drop the test and assume it's OK
      This commit undoes the previous push and adds a line to ignore on AIX

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/80206

2812 Bjorn Munch	2009-08-05
      Bug #45771 AIX and i5/OS Perl bug: check_socket_path_length in MTR fails
      Bug in Perl
      Scrap attempt to do this smartly on AIX, just drop the test and assume it's OK
      This commit undoes the previous push and adds a line to ignore on AIX

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/80242

2812 Bjorn Munch	2009-08-06
      Bug #45771 AIX and i5/OS Perl bug: check_socket_path_length in MTR fails
      Bug in Perl
      Scrap attempt to do this smartly on AIX, just drop the test and assume it's OK
      This commit undoes the previous push and adds a line to ignore on AIX

Re-pushed to azalea-mtr and 5.1-mtr

Pushed into 5.1.40 (revid:bjorn.munch@sun.com-20090903170735-679d64zp580wl39j) (version source revid:bjorn.munch@sun.com-20090903170735-679d64zp580wl39j) (merge vers: 5.1.40) (pib:11)

Test suite change. No changelog entry needed.

Setting report to NDI pending push into 5.4.x.

Pushed into 6.0.14-alpha (revid:alik@sun.com-20090929093622-1mooerbh12e97zux) (version source revid:alik@sun.com-20090922182109-vs5ign07cwht12z6) (merge vers: 6.0.14-alpha) (pib:11)

Pushed into 5.4.5-beta (revid:alik@sun.com-20090925094254-tjl9eajkzwzgthoe) (version source revid:alik@sun.com-20090922142453-x275o41whpyezh4q) (merge vers: 5.4.4-beta) (pib:11)

Test suite change. No changelog entry needed.

Pushed into 5.1.41-ndb-7.1.0 (revid:jonas@mysql.com-20091218102229-64tk47xonu3dv6r6) (version source revid:jonas@mysql.com-20091218095730-26gwjidfsdw45dto) (merge vers: 5.1.41-ndb-7.1.0) (pib:15)

Pushed into 5.1.41-ndb-6.2.19 (revid:jonas@mysql.com-20091218100224-vtzr0fahhsuhjsmt) (version source revid:jonas@mysql.com-20091217101452-qwzyaig50w74xmye) (merge vers: 5.1.41-ndb-6.2.19) (pib:15)

Pushed into 5.1.41-ndb-6.3.31 (revid:jonas@mysql.com-20091218100616-75d9tek96o6ob6k0) (version source revid:jonas@mysql.com-20091217154335-290no45qdins5bwo) (merge vers: 5.1.41-ndb-6.3.31) (pib:15)

Pushed into 5.1.41-ndb-7.0.11 (revid:jonas@mysql.com-20091218101303-ga32mrnr15jsa606) (version source revid:jonas@mysql.com-20091218064304-ezreonykd9f4kelk) (merge vers: 5.1.41-ndb-7.0.11) (pib:15)