As discussed on the gentoo-dev mailing list, some ebuilds choose to create a /var/run/${PN} directory at installation time instead of at runtime. Our init.d scripts should check whether a subdirectory is present and create it if need be, instead of using keepdir in the ebuilds. This bug report is a tracker for the specific flaws in affected packages - DO NOT COMMENT on this bug report, and file blocking bug reports for each affected package.
Created attachment 242797 [details] fgrep -H "keepdir /var/run" */*/*.ebuild eclass/*.eclass
Comment on attachment 242797 [details] fgrep -H "keepdir /var/run" */*/*.ebuild eclass/*.eclass Not included here are packages that install subdirectories in /var/run without using keepdir, such as sys-libs/pam.
I believe this one will be a little better: perl -ne 'while(/\\$/){chomp;$_.=<>;}; /keepdir/ and m#/var/run# and print "$ARGV: $_"' */*/*.ebuild eclass/*.eclass(In reply to comment #1)
Created attachment 290391 [details] var-run-PN List of packages that installs something in /var/run
*** Bug 414439 has been marked as a duplicate of this bug. ***
(In reply to comment #5) > *** Bug 414439 has been marked as a duplicate of this bug. *** Arg. Sorry about above. Mistakenly assumed from a log file this bug was for media-video/motion package, instead of a Tracker. Move along ...
How come /var/run was moved to a temporary filesystem without fixing all these bugs *beforehand*? For now, when a stable system gets updated with stable openrc-0.11.5, the latter moves /var/run to tmpfs and the affected daemons suddenly stop working after reboot. A partial solution for the problem would be making the affected ebuilds block the new openrc. (Though this wouldn't save users of the new stage3 that reportedly also have /var/run on tmpfs.) The actual solution would be to correct those ebuilds a year ago. :)
(In reply to comment #7) > How come /var/run was moved to a temporary filesystem without fixing all > these bugs *beforehand*? For now, when a stable system gets updated with > stable openrc-0.11.5, the latter moves /var/run to tmpfs and the affected > daemons suddenly stop working after reboot. +1 for this rant. Not even single warning was given.
(In reply to comment #0) > This bug report is a tracker for the specific flaws in affected packages - > DO NOT COMMENT on this bug report, and file blocking bug reports for each > affected package. ...
i add myself to the rant list... lot of daemons are broken on several servers here now, most of them stable I have to manually edit/fix init.d files. What a mess...
If I find the time for I gonna start fixing all remaining bugs next weekend.
Created attachment 333358 [details] packages for which no bug has been filed yet
*** Bug 523772 has been marked as a duplicate of this bug. ***
Depends on: 538848
commit 2bfe6e3f45c9c3a26b44bdceaff67839504a19de (HEAD -> master, origin/master, origin/HEAD) Author: Bertrand Jacquin <bertrand@jacquin.bzh> AuthorDate: Wed Oct 5 22:10:17 2016 +0100 Commit: Patrice Clement <monsieurp@gentoo.org> CommitDate: Sat Oct 15 16:52:33 2016 +0200 net-proxy/haproxy: major improvements. * Move to /run. * Remove commented USE_LINUX_SPLICE since declaration is implicit. * Remove commented USE_LINUX_TPROXY since declaration is implicit. * Keep lua for live ebuild. Gentoo-Bug: https://bugs.gentoo.org/332633 Package-Manager: portage-2.3.0 Closes: https://github.com/gentoo/gentoo/pull/2486 Signed-off-by: Patrice Clement <monsieurp@gentoo.org> net-proxy/haproxy/files/haproxy.confd | 4 ++-- net-proxy/haproxy/files/haproxy.initd-r2 | 4 ++-- net-proxy/haproxy/files/haproxy.initd-r3 | 4 ++-- net-proxy/haproxy/haproxy-1.6.9.ebuild | 11 +---------- net-proxy/haproxy/haproxy-9999.ebuild | 4 ---- 5 files changed, 7 insertions(+), 20 deletions(-) Merged, thanks!
(In reply to Patrice Clement from comment #15) > commit 2bfe6e3f45c9c3a26b44bdceaff67839504a19de (HEAD -> master, > origin/master, origin/HEAD) Reopening the tracker.
Removing blockers that are have nothing to do with installing things in /var/run at install time.
Adjusting the summary to reduce confusion. This bug was created to track improper usage of keepdir. Other scenarios should be tracked separately.