Bug 7671 - Excessive delay when "Looking at packages already installed" (microcode)
Summary: Excessive delay when "Looking at packages already installed" (microcode)
Status: RESOLVED FIXED
Alias: None
Product: Mageia
Classification: Unclassified
Component: Installer (show other bugs)
Version: Cauldron
Hardware: x86_64 Linux
Priority: release_blocker normal
Target Milestone: ---
Assignee: Thierry Vignaud
QA Contact:
URL:
Whiteboard: 3alpha2
Keywords:
: 7699 (view as bug list)
Depends on:
Blocks:
 
Reported: 2012-10-02 07:35 CEST by Juergen Harms
Modified: 2012-10-05 18:39 CEST (History)
5 users (show)

See Also:
Source RPM:
CVE:
Status comment:


Attachments
output of lspcidrake on my system (Dell Optiplex 990) (3.11 KB, text/plain)
2012-10-02 07:44 CEST, Juergen Harms
Details
stage1 log of this install - in case that helps (6.08 KB, text/x-log)
2012-10-02 18:09 CEST, Juergen Harms
Details

Description Juergen Harms 2012-10-02 07:35:29 CEST
Description of problem:

When the installer (new install from DVD, custom partition, target partition newly formatted) comes to "Looking at packages already installed), there is a more than 4-minutes period where nothing happens - no wait text in the display, no disk activity and CPU activity shown by the LEDS on the "box".

Switching on the display of the logs (CTRL-ALT-F3), the last message before this waiting period is

*running /sbin/insmod /tmp/microcode.ko

(the message before this message are a rapid sequence of "unknown package <name>"). When after the 4 minutes of inactivity new log message arrive, I switch back to the graphics display (CTRL-ALT-F7), and I get the first stage 2 configuration screen practically without delay)


Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1. see above
2.
3.
Juergen Harms 2012-10-02 07:38:38 CEST

Whiteboard: (none) => 3alpha2

Comment 1 Juergen Harms 2012-10-02 07:44:22 CEST
Created attachment 2901 [details]
output of lspcidrake on my system (Dell Optiplex 990)
Comment 2 Manuel Hiebel 2012-10-02 17:18:29 CEST
I can confirm that too

Assignee: bugsquad => thierry.vignaud

Comment 3 Juergen Harms 2012-10-02 18:09:43 CEST
Created attachment 2903 [details]
stage1 log of this install - in case that helps
Comment 4 Frank Griffin 2012-10-02 18:21:38 CEST
I've seen delays, but more on the order of 30 seconds.  I've always assumed that this was the result of a CPU-intensive scan of the hdlist versus the rpm database.  The delay probably varies with the processor speed.

CC: (none) => ftg

Comment 5 Manuel Hiebel 2012-10-03 09:05:49 CEST
>I've always assumed that this was the result of a CPU-intensive scan of the >hdlist versus the rpm database.

Me too but in fact, since some weeks, it stay _before_ the PackageCallChoice (scan of the rpms) 
and ps was giving nothing useful iirc
Comment 6 claire robinson 2012-10-03 12:24:13 CEST
Confirmed here too (i586 dvd)
claire robinson 2012-10-03 16:08:24 CEST

Priority: Normal => release_blocker

claire robinson 2012-10-03 16:08:43 CEST

CC: (none) => ennael1

Comment 7 Dave Hodgins 2012-10-04 00:09:13 CEST
*** Bug 7699 has been marked as a duplicate of this bug. ***

CC: (none) => davidwhodgins

Comment 8 Dave Hodgins 2012-10-04 00:14:48 CEST
See attachment 2912 [details]
Output of monitor-edid -v --perl

I'm getting a delay of at least 5 minutes with either the i586 or the x86-64
iso images.  I didn't notice a long delay with the dual-cd.

I suspect it's in the parsing/handling of the monitor-edid output, or
possibly due to trying to load the microcode.

The syslog shows the delay ...
<6>[   44.820091] EXT4-fs (sda9): mounted filesystem with ordered data mode. Opts: (null)
<6>[  102.300407] Adding 33554428k swap on /dev/sda2.  Priority:-1 extents:1 across:33554428k SS
<6>[  103.543772] EXT4-fs (dm-0): mounted filesystem with ordered data mode. Opts: acl
<6>[  103.977117] EXT4-fs (sda9): mounted filesystem with ordered data mode. Opts: acl
<6>[  114.563882] microcode: CPU0: patch_level=0x06000626
<3>[  175.325121] microcode: failed to load file amd-ucode/microcode_amd_fam15h.bin
<6>[  175.325128] microcode: CPU1: patch_level=0x06000626
<3>[  236.004271] microcode: failed to load file amd-ucode/microcode_amd_fam15h.bin
<6>[  236.004278] microcode: CPU2: patch_level=0x06000626

Looks like it tries to load the microcode several times before finally giving
up.
Comment 9 claire robinson 2012-10-05 13:23:14 CEST
Confirmed valid 2nd build (Oct 4th)
claire robinson 2012-10-05 13:27:06 CEST

CC: (none) => eeeemail

Manuel Hiebel 2012-10-05 17:22:05 CEST

Summary: Excessive delay when "Looking at packages already installed" => Excessive delay when "Looking at packages already installed" (microcode)

Manuel Hiebel 2012-10-05 17:22:43 CEST

CC: (none) => manuel.mageia

Comment 10 Thierry Vignaud 2012-10-05 18:37:46 CEST
Fixed in git
Comment 11 Thierry Vignaud 2012-10-05 18:39:22 CEST
Closing

Status: NEW => RESOLVED
Resolution: (none) => FIXED


Note You need to log in before you can comment on or make changes to this bug.