[Mageia-dev] /var/run after /usr/move

Tue Dec 11 17:50:32 CET 2012

Hiya,
+
+'Twas brillig, and Olivier Thauvin at 11/12/12 16:26 did gyre and gimble:
+> The "/usr move" included the move of /var/run to /run, however this has
+> not been done into packages, like quagga:
+> 
+> [...]
+> %dir %attr(0750,root,root) /var/run/quagga
+> [...]
+> 
+> Shouldn' this be changed ?
+
+Yes, I've mentioned this on the list a few times and also have worked
+through a few of the main, high profile packages already.
+
+I've also pushed a new rpmlint-mageia-policy update which bans files in
+/var/run, /var/lock and /run. It also bans udev rule files in
+/etc/udev.d and tmpfiles in /etc/ (the /etc/ tree is for admin config,
+and packaged files should be in /usr/lib/udev/rules.d and
+/usr/lib/tmpfiles.d respectively).
+
+It would be good to get this updated set of rules pushed to the
+infrastructure before the mass rebuild and beta2.
+
+Help here on updating packages is most welcome. The files/dirs in these
+folders basically need to be translated to tmpfiles configs instead.
+
+I've written a wiki page to document about how to handle tmpfiles:
+https://wiki.mageia.org/en/System_Service_policy
+
+This will give you a nice list:
+urpmf "^/var/(run|lock)/" | sort -u
+
+Unfortunetly in the last week or so, some new ones are appearing:
+
+ngircd
+tomcat
+
+so the sooner the lint policy is added the better IMO :)
+
+I do also still have one problem (reported on this list) that needs
+fixing that's the postfix+cyrus-sasl saslauth socket issue. Previously
+it was hardlinked, such that chrooted postfix worked OK. This is no
+longer possible due to it crossing filesystems, so we'll have to do some
+bind-mounting black magic I think :s
+
+> BTW: I recently updated a fresh mga2 to mga-cualdron and separated /var
+> is still not handle automatically when moving /usr.
+
+How recently? I pushed a new mageia-prepare-upgrade package to mga2
+core/updates_testing on Sunday which should handle /var mounting in the
+initramfs for processing the usrmove. Certainly it worked fine for me in
+my tests (ext4 / + swap + LVM /var and LVM /usr + encrypted LVM /home
+(all on the same Volume Group) and it worked fine with that somewhat
+overly complex example. Minor problem with the vga= line from the kernel
+not being propagated to the newly installed bootloader entry but other
+than that it worked fine, not tested with lilo tho' - only grub,
+although all manipulation is via our tools so should be OK in theory)
+
+I mentioned this already on the "ANN: Upgrading from Mageia 2 via urpmi"
+thread which has seen several new posts over the last few days, so there
+may be some additional info in there.
+
+Hope this helps.
+
+Col
+
+
+-- 
+
+Colin Guthrie
+colin(at)mageia.org
+http://colin.guthr.ie/
+
+Day Job:
+  Tribalogic Limited http://www.tribalogic.net/
+Open Source:
+  Mageia Contributor http://www.mageia.org/
+  PulseAudio Hacker http://www.pulseaudio.org/
+  Trac Hacker http://trac.edgewall.org/
+