Red Hat Bugzilla – Bug 437449
NetworkManager dead unexpectedly
Last modified: 2009-02-05 10:37:39 EST
Description of problem:
I tested Intel iwl4965 wireless driver against RHEL5.2-Beta on Intel SantaRosa
platform. I used NetworkManager to try connect AP with different
configurations. Within 2 hours testing, NetworkManager dead twice.
Version-Release number of selected component (if applicable):
NetworkManager version is 0.6.4-7.el5
It happened 2 times, but cannot be reproduced by regular steps.
Steps to Reproduce:
1. Insert iwl4965 driver by "#modprobe iwl4965".
2. Start NetworkManager by "#service NetworkManager start"
3. Using NetworkManager to connect APs. About 20 times connections and
NetworkManager dead unexpectedly.
Check NetworkManager status,
#service NetworkManager status
NetworkManager dead but pid file exists
NetworkManager is alive and works well.
/var/log/messages is attached.
Created attachment 298035 [details]
Can you get the scanned output from 'iwlist eth0 scan' for the specific AP
you're trying to connect to when NM crashes? It appears that there's an
unhandled case for creating an internal security object based on the scanned APs
properties. I'd need more detail about the configuration of the AP to come up
with a fix. Thanks!
I didn't reproduce this bug to have NM crash during the testing against
RHEL5.2-sp1. I will try to keep the scan log if I meet this bug again.
Ok, flip the bug state back to NEW/ASSIGNED when you've been able to grab the
scan info for the AP.
Created attachment 303117 [details]
output of "iwlist wlan0 scan"
In my testing for Kedron driver on RHEL5.2-SP4, NetworkManager dead onece and
hanging up once. When NetworkManager hanging up, wireless networking is still
good (Ping and data transfer work), but the screen froze and I need to change
to another term to type commands.
The attached file is output of "iwlist wlan0 scan". The AP I connecting when
the error happened is "otc-rh-wtest-01".
Does this still happen with RHEL 5.3?
(In reply to comment #6)
> Does this still happen with RHEL 5.3?
Dongpo did the wireless testing for RHEL5.3. I would ask him if this issue was reproduced with RHEL5.3.
Dongpo, did you see this issue during RHEL5.3 testing?
I have never seen such issue at RHEL 5.3.
Ok, thanks for the update. Please re-open or file a new bug against RHEL 5.3 if the issue comes up again.