[Voyage-linux] Better network stability

[Robert]

Hi Michael,

Thanks for the info. We use the hardware watchdog that is built into the Alix with a fair level of success, however I don't like to have the watchdog to monitor network traffic as I don't want our system stuck in a constant reboot cycle every time the internet drops out for 20-30 minutes.

I don't think our issues are hardware related, they're in a very cool environment so it's definitely not likely to be a heat issue, my guess this is just a Linux networking config issue.

Thanks for the input, I really appreciate it.

Rob

-----Original Message-----
From: Michael Cohen [mailto:scudette at gmail.com]
Sent: 26 February 2010 10:46
To: Robert
Cc: voyage-linux at list.voyage.hk
Subject: Re: [Voyage-linux] Better network stability

Hi Robert,
  We found stability issues with the Net5501s in hot environments and
possibly even with the Alixs - we had some units lock up hard and some
fail to boot altogether.  We enabled the watchdog but the problem is
that even after reboot the alix seems to lock up in the bios stage and
there is a window between the machine booting and the kernel
initialising the watchdog that the machine will deadlock with no hope
of resuming.

The watchdog chip seems to need to be enabled by the kernel before it
will fire a reset. This is inadequate for our application since if the
box is overheating it locks up in the bios stage and then never
restarts.

The solution we came up with is a hardware watchdog - similar to

 http://www.linuxfocus.org/English/July2002/article239.shtml

Except we ended up using tx on the RS232 so we can still use it as a
console - we just emit data on it periodically .

Another useful tool is monit which allows you to monitor an external
device like a modem and using the watchdog which can relay all devices
together, we can reboot the modem as well.

Michael.

2010/2/26 Robert <Robert at supporttime.com>:
> Hello Guys,
>
>
>
> I’m looking for some advice on how to make the networking a little more
> fail-safe on our systems.
>
>
>
> I have systems which are out in the field and they seem to intermittently
> lose their internet connection and stop talking home until they’re rebooted.
>
>
>
> I’ve not really been able to replicate the scenario to know exactly what’s
> causing it but it seems to be due to the router-modem rebooting, our voyage
> box then doesn’t lease a new IP from the modem or something like that,
> perhaps an IP conflict? I don’t really know.
>
>
>
> Either way, I’m looking to make the networking a little more robust so that
> if the router-modem looses it’s net connection, or reboots our system will
> pick it’s connection back up again.
>
>
>
> Can anyone advise on how to configure a slightly more robust setup to ensure
> the connection is picked up again as soon as it become available?
>
>
>
> Cheers, let me know if you need any more information.
>
>
>
> Rob
>
> ________________________________
> Support Time Managed Services - monitoring and development for entire
> business technology
>
> Tel: 0800 999 4333
> www.supporttime.com
>
> Support Time Limited
> 1000 Lakeside North Harbour
> Portsmouth PO6 3EZ United Kingdom
>
> _______________________________________________
> Voyage-linux mailing list
> Voyage-linux at list.voyage.hk
> http://list.voyage.hk/mailman/listinfo/voyage-linux
>
>

Support Time Managed Services - monitoring and development for entire business technology

Tel: 0800 999 4333
www.supporttime.com

Support Time Limited
1000 Lakeside North Harbour
Portsmouth PO6 3EZ United Kingdom