Bug 7847 - Controller startup crashers router
: Controller startup crashers router
Status: RESOLVED WORKSFORME
Product: SB Controller
Classification: Unclassified
Component: Product Quality
: unspecified
: PC Debian Linux
: P4 normal (vote)
: Investigating
Assigned To: Ross Levine
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2008-04-16 11:27 UTC by Marc Auslander
Modified: 2009-01-20 18:57 UTC (History)
1 user (show)

See Also:
Category: ---


Attachments
fragment of controller messages soon after boot (7.18 KB, application/octet-stream)
2008-05-18 05:34 UTC, Marc Auslander
Details
jive log during or after router crash (118.12 KB, text/plain)
2008-06-17 17:38 UTC, Marc Auslander
Details
Controller log when wireless crashes (124.36 KB, text/plain)
2008-06-19 12:26 UTC, Jesper Musaeus
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Marc Auslander 2008-04-16 11:27:42 UTC
I've seen this three times - but its not reproduceable.  Twice with 2224, once with 2232.

Turn off controller (hold down home till goodbye).  Turn back on.  Pretty soon after the boot starts, my router hangs completely.  No wireless, no wired.  Have to power cycle to restart.  Controller completes boot up and shows red wireless icon.

Router is Linksys WRT54G-2.  AFAIK, there are no logs available from the router.

I can't find any logs on the controller either.

My setup is standard - controller and receiver are DHCP clients.

Network is 128 bit WEP protected.

Squeezecenter is latest 7.1.
Comment 1 Ross Levine 2008-04-16 12:26:00 UTC
What version WRT54G are you using specifically, and what firmware? 
Comment 2 Marc Auslander 2008-04-16 13:24:38 UTC
WRT54G Ver 2  Firmware 4.21.1 which is latest available from Linksys.

In spirit of full disclosure, there is a second access point on my network - an Old IBM wireless router which is running in access point only mode.  But in the cases described, I believe even the wireless connection is to the Linksys.
Comment 3 Ross Levine 2008-04-16 13:51:52 UTC
Can you share more information about the old IBM access point? Firmware and such? You understand my concerns here, there are many reasons why routers crash, certainly we could have something to do with it, anything is possible. 

Can you reproduce this with the IBM AP out of the equation? Have you ever had this router crash occur before having controller, or anything similar? 
Comment 4 Marc Auslander 2008-04-16 18:55:12 UTC
The other device is an IBM High Rate Wireless Lan Gateway, model 1.01, software version  	R1.2.1 Jan. 03, 2002

It is configured with DHCP turned off, and with no connection to its lan port.  Its wireless is running on the same SSID and Channel as the Linksys, which is, AFAIK, the correct way to run a multiple access point network, and by experiment the only way in which the SB units will roam.  (They need to roam because apparently the Linksys forces them to associate with the other occasionally and they need to associate back to the linksys for best signal.)

I can't reproduce the failure right now.  I don't know how to make it happen.  If I could, I would try without the IBM.  Because the IBM is the switch to connect to the squeezecenter server, there is no convenient way to run without it, although I can turn if off and cycle the Controller.  I will turn off the wireless so the controller can't connect to it at all.

I understand you may not be able to figure this out.  You may just have to see if others see the problem and then try to gather more info.

I doubt the squeezecenter has anything to do with the router crash - I'm guessing DHCP.
Comment 5 Ross Levine 2008-04-17 11:47:31 UTC
I'm not sure where you read same SSID same channel, ideally it is better in terms of interference if you use same SSID and unique channels for each AP, I suggest 1, 6, 11. 

Also it may be worth updating the firmware:
http://www-307.ibm.com/pc/support/site.wss/MIGR-40582.html

I'm skeptical in this case I'll leave this bug open for a little while to see if others have anything to add. I didn't notice anything similar on the forums, we'll give this a little time to see if you come up with repro steps or if anyone else has anything to add. 
Comment 6 Marc Auslander 2008-04-17 12:10:54 UTC
There version numbering is a little inconsistent - my IBM is at the latest level.

I certainly agree you can't fix something that won't fail - so I'll just keep watching and reopen if I can give you something to work on.

Today I did a software upgrade and reboot with no problem.

As for running two AP's on the same channel - AFAIK the SB3 won't roam from one to the other if they are on different channels.  Of course, this is based on not seeing something happen that I would expect, rather than real facts.
Comment 7 Marc Auslander 2008-05-02 07:32:58 UTC
Happened again during the reboot after a forced upgrade to 2416.  Upgrade by scp to the controller.

I have disabled the wireless on the second access point so everything is talking to the linksys WRT54G-V2.

Anything I can enable to catch useful info on this rather rare failure?

I've seen chatter in other bugs and/or forum about excessive wireless traffic.  Could that be the cause of this?
Comment 8 Ross Levine 2008-05-02 14:51:30 UTC
Marc have you ever had crashing issues with this router before the Squeezebox Controller? I did some research and there are a number of crashing issues related to the WRT54G v2.0 and 2.2, a number of people using 3rd party firmware suggest better performance. I wasn't able to find crashing issues that matched the symptoms you've described exactly. 

This is a very common router I'm sure if there is an issue between SBC and WRT54G v2.0/2.2 we will hear about it. Marc you could try enabling player.jive and player.source. 
Comment 9 Marc Auslander 2008-05-02 15:17:53 UTC
I'm running the latest Linksys firmware.

I have not seen router crashes before these.

They are very rare - and seen associated with reboot, and usually with an upgrade.

I've set player.source and player.jive (in squeezeserver) to debug.

Is there anything to turn on in the controller?

I'm not expecting you to fix this based on what I have - but want to get as much info as possible for you.
Comment 10 Ross Levine 2008-05-02 15:24:02 UTC
I understand Marc and I very much appreciate your help! If you'd like you can put an SD card into controller with a directory /var and it will log too. 

Does the router have any logging?
Comment 11 Marc Auslander 2008-05-02 16:06:38 UTC
I'm not sure I understand what you mean by /var.

I've installed a card and made a var directory in its root, which is thus /mnt/mmc/var

is that what you mean?

In any case, the logging entry in adv settings is not active - center button push does nothing.
Comment 12 Ross Levine 2008-05-02 16:23:57 UTC
Yes, var directory on the SD card. You should notice some log data in there, right?
Comment 13 Marc Auslander 2008-05-03 11:05:44 UTC
It appears meessages has moved itself from /var/log to /mnt/mmc/var/log.  Should I see anything else?
Comment 14 Ross Levine 2008-05-05 17:33:39 UTC
Yes, within that directory you should see some logs. From SSH try:

#cat /mnt/mmc/var/log/messages
Comment 15 Marc Auslander 2008-05-05 19:56:05 UTC
I saw that messages had moved from /var/log to /mnt/mmc/var/log - but I don't see any other logs.  Is that normal?  
Comment 16 Ross Levine 2008-05-08 11:52:38 UTC
Marc, messages is the log, did you try the cat command?
Comment 17 Marc Auslander 2008-05-08 12:08:06 UTC
sorry to be unclear.

I have no problem getting as messages - but I'm used to linux systems with a number of logs and expected them to appear here.  Apparently not the case - its all in messages I assume.


If and when something bad happens again I'll capture messages.

For now this is on hold.

Thanks.
Comment 18 Marc Auslander 2008-05-18 05:34:59 UTC
Created attachment 3356 [details]
fragment of controller messages soon after boot
Comment 19 Marc Auslander 2008-05-18 05:36:29 UTC
I noticed something funny in messages right after boot.  Happens every boot, I think.

line

May 18 08:11:29 (none) local0.err udhcpc[314]: bogus packet, option fields too long 

repeated 67 times.

Above (sorry - I'm not a bugzilla wizard :-) is the interesting part of messages.
Comment 20 Marc Auslander 2008-06-17 17:38:47 UTC
Created attachment 3448 [details]
jive log during or after router crash
Comment 21 Jesper Musaeus 2008-06-18 06:15:51 UTC
Hi, I think I've just seen the same problem. 

I've had massive connection problems with my Zyxel router (bug 7153) and I thougth it was fixed by using static ip for my SC (Vista laptop) and WEP instead of WPA-PSK. It helped a lot, but I stille have occasional connection problems where the router crashes (and I have to powercycle the router). 

After reading of others having problems with the controller I tried leaving it in the cradle (still power on). And for more than 36 hours of constant streaming there was no problems, then after 36 hours I took the controller out of the cradle to play with the Information Browser (still listening to music) and after just one minute the router crashed.
Comment 22 Marc Auslander 2008-06-18 08:34:34 UTC
The crash I posted above occured after turning off DHCP in the router - DHCP now done by my linux machine.  So DHCP is NOT part of the story.
Comment 23 Ross Levine 2008-06-18 16:47:35 UTC
Punting to 7.2, this needs more time. Marc when the router crashes, your network is 100% down, right? I'm surprised nobody else is seeing this. I'm trying leaving SBC cradled for 36 hours as you mention in comment #21, is that reproducible for you? I imagine I should wait another 36 hours and then ask. :)
Comment 24 Marc Auslander 2008-06-18 16:57:15 UTC
Be careful - somoeone else (I'm Marc) has joined in.  He's the one with the cradled comments.  Mine only crashes on reboot of the router, and seems more likely after a firmware upgrade.  In the last crash, the sequence was upgrade (mostly in the cradle), reboot success, install applet (out of cradle), reboot and crash.

The crash takes the router down totally - it won't talk wired or wireless - power cycle to get it back.

My router is a WRT54G V2.  The other poster has a different router.
Comment 25 Ross Levine 2008-06-18 17:32:01 UTC
Sorry about the confusion. If you could elaborate one thing for me Marc, your last comment confuses me a little. Are you saying you reboot the router, upgrade the firmware on Squeezebox Controller, and the router then crashes (sometimes)?
Comment 26 Marc Auslander 2008-06-18 18:55:23 UTC
The reboots are reboots of the controller - hold home key till it powers down - or in case of firmware or applet install, automatic reboot of controller.

I would assume, but can't prove, that the triggering interactions occur during the boot sequence when the controller tries to get on the wireless network.

The only router restart in the story is the power cycle to clear the router after it hangs.
Comment 27 Jesper Musaeus 2008-06-19 01:02:58 UTC
Hi Ross,

I'm the one with the crash after 36 hours :-)

Let me first say that the reason for the 36 hours of constant streaming was just to see if I have some sorts of interferrance-problems in my environment. I wanted to be sure that it was possible to maintain a constant signal for a long period of time. No problem there.

So You don't need to leave it for 36 I guess.

With regards to the crash I will have to test further to see if I can 
1. reproduce at will and 
2. wether it is possible to connect wired to the router after the crash or if it is a complete crash (I think it is).

By the way, yesterday I turned off wireless power management on the controller, so if I'm not seeing the problem any more I will let you know.

Is there a log on the controller that I can access next time I see this problem?, if yes, could You give me a step by step description on how to access this (bare in mind that I'm not used to work with linux).

/Jesper
Comment 28 Ross Levine 2008-06-19 11:44:07 UTC
Jesper thank you for chiming in. Regarding logging SBC, see comments 10-17 on this bug. I think what you're experiencing might be just bug 7153, I think your power management test will determine this. Let us know how it goes. :)
Comment 29 Jesper Musaeus 2008-06-19 12:26:20 UTC
Created attachment 3461 [details]
Controller log when wireless crashes

Started in good condition at around 19.29. 

Streaming music, played around with the controller trying to provoke router problems - at first nothing so I left the controller in the cradle for a while. 

At 20:29 I picked it up again and played with the informationbrowser reading a few rss-news and then after a few seconds my wireless connection breaks down. 

I'm able to connect a laptop through wire, but all wireless is gone until I powercycle the router.
Comment 30 Ross Levine 2008-11-12 15:57:14 UTC
Jesper do you still see this with SC 7.2.1?
Comment 31 Marc Auslander 2008-11-12 16:34:53 UTC
I have reorganized my home network a bit and no longer see this - at least not for a while.  But it often happened at upgrade time, and that hasn't happened for a while either - I'm running 7.2.1.
Comment 32 Jesper Musaeus 2008-11-19 08:17:00 UTC
(In reply to comment #30)
> Jesper do you still see this with SC 7.2.1?
> 

Hi Ross,

Sorry for the late answer. no I haven't seen it for a long time. But, as Marc, I have changed a few things in my setup. With out doubt the biggest improvement came from changing the router to a new upgraded model with newer firmware.

Before: ZyXEL 660HW-61 V3.40(PE.11)| 05.22.2006

Now: ZyXEL 660HW-D1 V3.40(ANE.5)| 07.22.2008

I also swithed from WEP to WPA2-PSK (I was using WEP because reauthentication failed).

Theres no doubt in my mind that Vista sp1 and XP sp3 didn't cooperate well with the old router model.

Please also consider this information as input from me in regards to bug 7153, where I also participated.


Cheers
Jesper
Comment 33 Ross Levine 2009-01-20 18:57:23 UTC
I believe this to be ZyXEL / Controller communication issues, and I'm going to close this bug because bug 7153 exists for that purpose. If anyone sees a router crash when Controller starts please feel free to re-open this bug with details. If you're experience a ZyXEL / Controller related connection issue, either during setup or afterward, please feel free to contribute to bug 7153 with relevant details.