Mail archive

[alpine-user] Re: workaround for openrc crashes (Was: [alpine-devel] openrc crashes)

From: Steffen Nurpmeso <>
Date: Fri, 28 Dec 2018 15:17:11 +0100

Posting this to the right list this time, i hope.

Steffen Nurpmeso wrote in <>:
 |I see crashes with openrc. I saw them with the old [edge] i had
 |(i did not update for two week because of download limit excess),
 |and i see them again now:
 | #?0[root_at_essex steffen]# openrc
 | * openrc: failed to create stopping dir `/run/openrc/rc.stopping': \
 | File exists
 |This directory will not be cleaned up. I can manually remove it,
 |but we get there with later crashes, again.
 | #?1[root_at_essex steffen]# rc-status -r
 | wlan
 | #?0[root_at_essex steffen]# rc-status -a
 | Runlevel: nonetwork
 | Runlevel: shutdown
 | killprocs [ stopped ]
 | savecache [ stopped ]
 | mount-ro [ stopped ]
 | Runlevel: wlan-up
 | Segmentation fault
 |This as below
 | #?139[root_at_essex steffen]# rc-status -r
 | wlan
 | #?0[root_at_essex steffen]#
 |This i find in dmesg:
 | [80024.596434] traps: openrc[16381] general protection ip:7fe20d072650 \
 | sp:7ffc6865bd28 error:0 in[7fe20d06b000+46000]
 | [81824.602270] traps: openrc[16966] general protection ip:7f39a1e65650 \
 | sp:7ffc63edb598 error:0 in[7f39a1e5e000+46000]
 | [82724.611610] traps: openrc[17268] general protection ip:7fe069116650 \
 | sp:7fff3b183bf8 error:0 in[7fe06910f000+46000]
 | [82768.761448] traps: rc-status[17295] general protection ip:7f63aea1e65\
 | 0 sp:7ffc3d9a6e98 error:0 in[7f63aea17000+46000]
 |The last reflects the on above.
 |The problem is that at one point openrc refuses to work: it
 |becomes impossible to change the runlevel, for example, it will
 |simply crash earlier, and you are stuck, and only a reboot helps
 |out. (I did that once only until now, though.)
 | #?0[root_at_essex steffen]# openrc wlan-up
 | Segmentation fault
 | [83347.269208] traps: openrc[17596] general protection ip:7f3dcc803650 \
 | sp:7ffd54106698 error:0 in[7f3dcc7fc000+46000]
 |Any ideas on anything regarding this?

So i know now how to solve this without rebooting.
My laptop situation as follows:

- I mostly deal with three runlevels, "default", "wlan", and
  "wlan-up". "wlan" builds upon "net" (empty but dependency on
  "default") and loads kernel driver, wpa_supplicant and then
  calls my firewall script which establishes rules for the new
  interface, which it "ifup"s first. This ifup starts udhcpc
  i think. "wlan-[is-]up" only adds/removes openntp onto that.

- Upon LID close and such we go to "default".

+ To overcome the no-go situation (which still faults at +46000 by
  the way) i have to manually kill [udhcpc and] wpa_supplicant,
  then openrc becomes usable again. "openrc default" then causes
  mess due to the manual intervention, but thereafter system is
  usable again. Hooray!

- I do not understand why dnsmasq is often, but not always
  restarted if i go down from "wlan" to "default". dnsmasq, my
  basic firewall setup, (local) "networking" all is in "default".
  I have already modified init.d/wpa_supplicant and commented out
  "#before dns dhcpcd net" (_after_ the November crashes) to
  overcome this, but it does not seem to matter.


|Der Kragenbaer, The moon bear,
|der holt sich munter he cheerfully and one by one
|einen nach dem anderen runter wa.ks himself off
|(By Robert Gernhardt)

Received on Fri Dec 28 2018 - 15:17:11 UTC