Skip to contentSkip to conversation

2.4.20 htb3 oops

LARTC Göran Runfeldt 2 participants 3 messages Mar 4, 2003 – Mar 11, 2003

lore
Share & tools
lore
#0 lore raw

I have arranged a serial console with logging to a terminal client, so next time it happens I will have the output from the panic.

For the record:
One of the machines crashed again this night at 11.30 p.m. (local time). I noticed that the MRTG graph looks a bit odd:
http://hem.wasadata.net/goran/mrtg.png

The total limit is set to12Mbit, as you probably can see if you check the ruleset in my first post. Even about 2-4 hours before the crash the graph shows two network traffic "spikes". The first one tops at about 22Mbps of outgoing traffic on the interface and the second one tops 28Mbps of incoming traffic.
Could this have anything to do with the crash?

Many Thanks,
Göran

----- Original Message -----
From: Abraham van der Merwe <[email protected]>
To: Göran Runfeldt <[email protected]>
Cc: <[email protected]>
Sent: Monday, March 03, 2003 5:29 PM
Subject: Re: [LARTC] 2.4.20 htb3 oops

Hi G?ran!

Unfortunately that is not going to help much since you can't figure out from the information below where in the code it crashes. Next time, please copy the entire kernel panic (including stack trace) and run it through ksymoops or look up the symbols in vmlinux.

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/

#1 lore raw

Hep

Quoted 8 lines · Göran Runfeldt

I have arranged a serial console with logging to a terminal client, so next time it happens I will have the output from the panic.

For the record:
One of the machines crashed again this night at 11.30 p.m. (local time). I noticed that the MRTG graph looks a bit odd:
http://hem.wasadata.net/goran/mrtg.png

Quoted 6 lines

The total limit is set to12Mbit, as you probably can see if you check the ruleset in my first post. Even about 2-4 hours before the crash the graph shows two network traffic "spikes". The first one tops at about 22Mbps of outgoing traffic on the interface and the second one tops 28Mbps of incoming traffic.
Could this have anything to do with the crash?

Any solution to the problems describe above? Im currently looking into building a new kernel 2.4.20 with HTB compiled as module in a production enviroment (2,5mbits average, 6-10mbits/peak). I wont use it if its broken though?

Signature · 14 lines
Venlig hilsen/Kind regards
Thomas Kirk
ARKENA
tlf/phone +04570233456
thomas(at)arkena(dot)com
Http://www.arkena.com


Oh, give me a home,
Where the buffalo roam,
And I'll show you a house with a really messy kitchen.
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
#2 lore raw
In reply to Göran Runfeldt
Quoted 8 lines

Any solution to the problems describe above? Im currently looking into building a new kernel 2.4.20 with HTB compiled as module in a production enviroment (2,5mbits average, 6-10mbits/peak). I wont use it if its broken though?

Unfortunately we have not yet come to a solution when it comes to our problems.

We have had the machines running for more than five days now, without any HTB qdisc/classes loaded. This confirms that this is a problem with HTB. I have activated the rules again in hope of getting a useful crash dump.

I do not know if anyone else has or will have this problem. All I know is that we have the problem over here.

Göran
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/