poudriere and tinderbox hanging on shells/bash build

[spork]

On a few edit: 8.3 boxes I'm seeing both tinderbox and poudriere bulk builds hanging when they reach shells/bash. I'm seeing this both with a current (as of yesterday) ports tree and an older snapshot that we normally build from.

The hang always shows a process called psize.aux taking up 100% of cpu time. Both package-building utilities eventually kill this process off, deciding it is hung.

Google is showing me nothing about this. Any hints as to what else to look at?

As best I can tell this psize.aux process is a test to determine something regarding pipes. I suspect this has something to do with building inside the temporary jails that both package building utilities use.

The tail end of poudriere's log of the build shows the process that failed and notes that it was killed because it was a runaway:

cc  -DHAVE_CONFIG_H -DSHELL   -I. -I..  -I.. -I../include -I../lib -I.   -I/usr/local/include -g -o psize.aux ./psize.c
/bin/sh ./psize.sh > pipesize.h
====>> Killing runaway build
===>  Cleaning for bash-4.2.45_1
====>> Leftover processes:
USER   PID %CPU %MEM   VSZ   RSS  TT  STAT STARTED      TIME COMMAND
root 25466  0.0  0.0  3204  1936   1  I+J   9:32AM   0:00.06 make -C /usr/ports/shells/bash build
root 25511  0.0  0.0  8352  1264   1  I+J   9:32AM   0:00.00 - /bin/sh -ec (cd /wrkdirs/usr/ports/shells/bash/work/bash-4.2; if ! /usr/bin/env TMPDIR="/tmp" TMPDIR="/tmp" SHELL=/bin/sh NO_LINT=YES PREFIX=/usr/local  LOCALBASE=/usr/local  LIBDIR="/usr/lib"  CC="cc" CFLAGS="-O2 -pipe -fno-strict-aliasing"  CPP="cpp" CPPFLAGS="  -I/usr/local/include"  LDFLAGS=" -L/usr/local/lib"  CXX="c++" CXXFLAGS="-O2 -pipe -fno-strict-aliasing"  MANPREFIX="/usr/local" BSD_INSTALL_PROGRAM="install  -s -o root -g wheel -m 555"  BSD_INSTALL_LIB="install  -s -o root -g wheel -m 444"  BSD_INSTALL_SCRIPT="install  -o root -g wheel -m 555"  BSD_INSTALL_DATA="install  -o root -g wheel -m 444"  BSD_INSTALL_MAN="install  -o root -g wheel -m 444" /usr/bin/make -f Makefile   all; then  if [ -n "" ] ; then  echo "===> Compilation failed unexpectedly.";  (echo "") | /usr/bin/fmt 75 79 ;  fi;  false;  fi)
root 25512  0.0  0.0  3204  1420   1  I+J   9:32AM   0:00.03 `-- /usr/bin/make -f Makefile all
root 29903  0.0  0.0  8352  1240   1  I+J   9:33AM   0:00.00   `-- /bin/sh -ec (cd ./builtins && /usr/bin/make  DEBUG= libbuiltins.a ) || exit 1
root 29904  0.0  0.0  3204   980   1  I+J   9:33AM   0:00.02     `-- /usr/bin/make DEBUG= libbuiltins.a
root 30280  0.0  0.0  8352  1252   1  I+J   9:33AM   0:00.00       `-- /bin/sh ./psize.sh
root 30282 100.0  0.0  2772   588   1  R+J   9:33AM 119:31.54         `-- ./psize.aux
Killed
build of /usr/ports/shells/bash ended at Fri Feb 14 06:33:21 EST 2014

This seems to be an 8.3-related issue at this point. Can't reproduce on 8.4.

Since I saw no similar problems reported, I'm going to have to blame something in our environment. Maybe zfs-related, maybe some sysctl or something that we set everywhere that is not set on the other boxes I tested with, I don't know...

Still curious as to just what the psize test does and what might make it fail.