In the course of conducting a new installation, I have found that portage reproducibly deadlocks during an emerge -e @world run. I started with the stage3-amd64-20121107.tar.bz2 tarball, switched to python-2.7 and unmasked gcc-4.6 (as I always do). After installing gcc-4.6, I wanted to rebuild the other packages, so I ran: emerge -e @world --exclude sys-devel/gcc Thus far, I have reproduced the issue with the following combinations. --jobs 2 and MAKEOPTS="-j12" --jobs 4 and MAKEOPTS="-j6" Here's an example from a recent run: Calculating dependencies... done! >>> Verifying ebuild manifests >>> Starting parallel fetch >>> Emerging (1 of 122) sys-libs/ncurses-5.9-r2 >>> Emerging (2 of 122) sys-libs/zlib-1.2.7 >>> Emerging (3 of 122) app-arch/xz-utils-5.0.4 >>> Emerging (4 of 122) sys-apps/sed-4.2.1-r1 >>> Emerging (5 of 122) sys-devel/gnuconfig-20120818 >>> Emerging (6 of 122) virtual/libiconv-0 >>> Emerging (7 of 122) dev-libs/expat-2.1.0-r2 >>> Emerging (8 of 122) app-arch/bzip2-1.0.6-r3 >>> Emerging (9 of 122) sys-devel/make-3.82-r3 >>> Emerging (10 of 122) sys-apps/findutils-4.4.2-r1 >>> Emerging (11 of 122) sys-apps/diffutils-3.2 >>> Emerging (12 of 122) sys-devel/patch-2.6.1 >>> Emerging (13 of 122) sys-devel/gcc-config-1.7.3 >>> Emerging (14 of 122) sys-libs/timezone-data-2012f >>> Emerging (15 of 122) app-misc/editor-wrapper-4 >>> Emerging (16 of 122) dev-libs/popt-1.16-r1 >>> Emerging (17 of 122) sys-devel/binutils-config-3-r3 >>> Emerging (18 of 122) dev-libs/libffi-3.0.11 >>> Emerging (19 of 122) app-misc/mime-types-8 >>> Emerging (20 of 122) virtual/libintl-0 >>> Emerging (21 of 122) sys-apps/kbd-1.15.3 >>> Emerging (22 of 122) app-arch/tar-1.26 >>> Emerging (23 of 122) app-arch/gzip-1.5 >>> Emerging (24 of 122) sys-apps/which-2.20 >>> Emerging (25 of 122) sys-devel/autoconf-wrapper-12 >>> Installing (25 of 122) sys-devel/autoconf-wrapper-12 >>> Emerging (26 of 122) app-portage/portage-utils-0.10 >>> Emerging (27 of 122) sys-devel/automake-wrapper-7 >>> Installing (27 of 122) sys-devel/automake-wrapper-7 >>> Emerging (28 of 122) app-arch/lbzip2-2.2 >>> Installing (26 of 122) app-portage/portage-utils-0.10 >>> Installing (28 of 122) app-arch/lbzip2-2.2 >>> Installing (2 of 122) sys-libs/zlib-1.2.7 >>> Installing (5 of 122) sys-devel/gnuconfig-20120818 >>> Installing (6 of 122) virtual/libiconv-0 >>> Installing (3 of 122) app-arch/xz-utils-5.0.4 >>> Installing (4 of 122) sys-apps/sed-4.2.1-r1 >>> Installing (8 of 122) app-arch/bzip2-1.0.6-r3 >>> Installing (7 of 122) dev-libs/expat-2.1.0-r2 >>> Installing (9 of 122) sys-devel/make-3.82-r3 >>> Installing (12 of 122) sys-devel/patch-2.6.1 >>> Installing (13 of 122) sys-devel/gcc-config-1.7.3 >>> Installing (14 of 122) sys-libs/timezone-data-2012f >>> Installing (15 of 122) app-misc/editor-wrapper-4 >>> Installing (10 of 122) sys-apps/findutils-4.4.2-r1 >>> Installing (16 of 122) dev-libs/popt-1.16-r1 >>> Installing (17 of 122) sys-devel/binutils-config-3-r3 >>> Installing (19 of 122) app-misc/mime-types-8 >>> Installing (20 of 122) virtual/libintl-0 >>> Installing (11 of 122) sys-apps/diffutils-3.2 >>> Installing (18 of 122) dev-libs/libffi-3.0.11 >>> Installing (1 of 122) sys-libs/ncurses-5.9-r2 >>> Installing (21 of 122) sys-apps/kbd-1.15.3 >>> Installing (24 of 122) sys-apps/which-2.20 >>> Installing (23 of 122) app-arch/gzip-1.5 >>> Installing (22 of 122) app-arch/tar-1.26 >>> Emerging (29 of 122) sys-apps/texinfo-4.13 >>> Emerging (30 of 122) sys-libs/readline-6.2_p1 >>> Emerging (31 of 122) sys-apps/coreutils-8.16 >>> Emerging (32 of 122) app-misc/pax-utils-0.4 >>> Emerging (33 of 122) sys-apps/file-5.11 >>> Emerging (34 of 122) sys-libs/cracklib-2.8.19 >>> Emerging (35 of 122) virtual/libffi-3.0.11 >>> Emerging (36 of 122) net-misc/rsync-3.0.9-r2 >>> Emerging (37 of 122) sys-apps/module-init-tools-3.16-r2 >>> Emerging (38 of 122) sys-process/procps-3.2.8_p11 >>> Emerging (39 of 122) sys-devel/m4-1.4.16 >>> Installing (39 of 122) sys-devel/m4-1.4.16 >>> Emerging (40 of 122) dev-libs/gmp-5.0.2_p1 >>> Emerging (41 of 122) sys-devel/bison-2.4.3 >>> Installing (41 of 122) sys-devel/bison-2.4.3 >>> Emerging (42 of 122) virtual/yacc-0 >>> Installing (42 of 122) virtual/yacc-0 >>> Installing (32 of 122) app-misc/pax-utils-0.4 >>> Installing (30 of 122) sys-libs/readline-6.2_p1 >>> Installing (33 of 122) sys-apps/file-5.11 >>> Installing (35 of 122) virtual/libffi-3.0.11 >>> Installing (29 of 122) sys-apps/texinfo-4.13 >>> Installing (34 of 122) sys-libs/cracklib-2.8.19 >>> Installing (37 of 122) sys-apps/module-init-tools-3.16-r2 >>> Installing (38 of 122) sys-process/procps-3.2.8_p11 >>> Installing (36 of 122) net-misc/rsync-3.0.9-r2 >>> Installing (31 of 122) sys-apps/coreutils-8.16 >>> Installing (40 of 122) dev-libs/gmp-5.0.2_p1 >>> Emerging (43 of 122) app-admin/eselect-1.3.1 >>> Emerging (44 of 122) sys-apps/debianutils-4.3.2 >>> Emerging (45 of 122) sys-apps/sandbox-2.5 >>> Emerging (46 of 122) virtual/modutils-0 >>> Emerging (47 of 122) sys-apps/gawk-4.0.1 >>> Emerging (48 of 122) dev-libs/mpfr-3.0.1_p4 >>> Emerging (49 of 122) app-shells/bash-4.2_p37 >>> Emerging (50 of 122) app-admin/eselect-vi-1.1.7-r1 At this point, the process tree looks like: screen zsh bash emerge python2.7 And the python2.7 process looks like this: /usr/bin/python2.7 /var/tmp/portage/._portage_reinstall_.czcWaN/bin/lock-helper.py /var/tmp/portage/app-admin/eselect-vi-1.1.7-r1 If I Ctrl+C then --resume, it hangs again. Thus far, that's been on eselect-vi. I shall attach make.conf and lsof output, as requested in #gentoo-portage. Incidentally, the last time I ran through this procedure was in setting up a system with almost identical specs. At that time, I think portage-2.1.10.44 was current and I did not encounter any locking issues.
Created attachment 330976 [details] Last used make.conf
Created attachment 330978 [details] lsof | grep /var/tmp/portage
Just to add that I didn't start with FEATURES="parallel-install". That was merely something I recently tried to see if if influences the behaviour of portage with respect to this issue. Apparently, it doesn't.
Are you sure that pid 21225 is not just an orphan process that needs to be killed? It seems likely, because the emerge pid 24497 that's trying to lock the same file has a higher pid.
Well spotted. You're absolutely right. Sorry for the false alarm.
No problem.
