Bug 2804 - Using systemd /proc is available too late for dkms
Summary: Using systemd /proc is available too late for dkms
Status: RESOLVED FIXED
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: Cauldron
Hardware: All Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Mageia Bug Squad
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks: 2120
  Show dependency tree / graph
 
Reported: 2011-09-21 21:03 CEST by Dick Gevers
Modified: 2011-10-28 16:43 CEST (History)
4 users (show)

See Also:
Source RPM: systemd
CVE:
Status comment:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Description Dick Gevers 2011-09-21 21:03:10 CEST 
Description of problem:

I booted earlier tonight.

According to uptime that was at 18:51:33

dmesg shows this error:

rc.sysinit[95]: /usr/sbin/dkms_autoinstaller: line 56: /dev/stdout: No such device or address

while:
stat /dev/stdout
  File: `/dev/stdout' -> `/proc/self/fd/1'
  Size: 15              Blocks: 0          IO Block: 4096   symbolic link
Device: 5h/5d   Inode: 68          Links: 1
Access: (0777/lrwxrwxrwx)  Uid: (    0/    root)   Gid: (    0/    root)
Access: 2011-09-21 18:26:12.852684012 +0000
Modify: 2011-09-21 18:26:08.833872010 +0000
Change: 2011-09-21 18:26:08.833872010 +0000


stat /proc/self/fd/1
  File: `/proc/self/fd/1' -> `/dev/pts/0'
  Size: 64              Blocks: 0          IO Block: 1024   symbolic link
Device: 3h/3d   Inode: 40722       Links: 1
Access: (0700/lrwx------)  Uid: (    0/    root)   Gid: (    0/    root)
Access: 2011-09-21 18:51:58.882687201 +0000
Modify: 2011-09-21 18:51:58.882687201 +0000
Change: 2011-09-21 18:51:58.882687201 +0000

So /dev/stdout is available 15 seconds after bootup and dkms_autoinstaller called from rc.sysinit cannot run (yet).

Before systemd there was no such error.

So I guess the order of doing stuff needs looking at.
Comment 1 John Balcaen 2011-09-22 01:01:23 CEST 
Just in case did you try without speedboot ?

CC: (none) => balcaen.john

Comment 2 Dick Gevers 2011-09-22 23:34:07 CEST 
Yes, it was already disabled, else rsyslog does not start properly for me (see: https://bugs.mageia.org/show_bug.cgi?id=2760 )
Manuel Hiebel 2011-10-07 14:40:16 CEST

Blocks: (none) => 2120

Comment 3 D Morgan 2011-10-22 23:32:27 CEST 
can you test with systemd + initscripts from testing to tell if this is still valid ?

CC: (none) => dmorganec

Comment 4 Dick Gevers 2011-10-23 12:55:58 CEST 
Sure, if you can tell me how: the bug manifests itself only after upgrading or installing a new dkms module. I only use dkms-fglrx and am waiting for a new version: see #2875. If there is a module that has no effect on anything else, tell me and I'll try it.
Comment 5 D Morgan 2011-10-25 15:02:16 CEST 
try to install an other one ( like dkms-virtualbox ), here with the systemd of updates-testing it "seems" to install ok.
Comment 6 Dick Gevers 2011-10-25 18:08:15 CEST 
Well yes, but urpmi dkms-virtualbox not only adds the module, but already installs it. So during boot up it does not go through the routine of installing the module. So Im afraid this doesn't prove anything.
D Morgan 2011-10-27 14:20:39 CEST

Source RPM: initscripts-9.21-12.mga1 => systemd

Comment 7 Thomas Backlund 2011-10-27 14:27:01 CEST 
You can test it by installing a new kernel.
Then dkms should trigger on boot.

CC: (none) => tmb

Olav Vitters 2011-10-27 15:07:56 CEST

CC: (none) => olav

Comment 8 Dick Gevers 2011-10-28 16:43:47 CEST 
Okay, Installed tmb kernel and dkms builds okay. Closing.

Status: NEW => RESOLVED
Resolution: (none) => FIXED
Note You need to log in before you can comment on or make changes to this bug.