[eepro100] EEpro100, Red Hat 7.1, wait_for_cmd_done timeout errors

[Jim Hribnak]

This is totally bizarre..  This just happened (while I was out of the office
today) .  The other funny thing is all these 30 sites were on our main
"Shared" hosting server which never experienced this problem. move them and
the problem started just a month ago.

Are there any other drivers that anyone has tried? ./  Can these NICs be
flashed to upgrade their firmware?

Here is how it started (as per /var/log/messages) it seemed to correct
itself before 7am this morning.

Jan 10 06:56:40 corusent kernel: Command 80 was not immediately accepted,
123 ticks!
Jan 10 10:12:28 corusent kernel: Command 80 was not immediately accepted,
10001 ticks!
Jan 10 10:12:28 corusent last message repeated 8 times


Then we are hosed....

Jan 10 10:12:36 corusent kernel: eth1: Transmit timed out: status 0090  0080
at 34200109/34200121 comm
ands 000c0000 000c0000 000c0000.
Jan 10 10:12:36 corusent kernel: Command 80 was not immediately accepted,
10001 ticks!
Jan 10 10:12:36 corusent kernel: eth1: Restarting the chip

Are the drivers that Dell provides the same ones that we are using now?
(only older perhaps?)  Anyone tried the e100 drivers?

This is the oddest problem especially when all of us have the same nics in
other machines but they are not seeing the same problem.

Something has to be similar with our machines (the ones crashing)

lets post system information and see if we can come to a solution.

Jim


----- Original Message -----
From: "John Madden" <jmadden@ivytech.edu>
To: <hribnak@nucleus.com>
Cc: <eepro100@scyld.com>
Sent: Friday, January 10, 2003 6:54 AM
Subject: Re: [eepro100] EEpro100, Red Hat 7.1, wait_for_cmd_done timeout
errors


> > Not sure what to do.. the perplexing problem is I have another server
> > running RH 6.2 and the same NIC and the SAME 1999 driver by Don, and
> > this has WAY more IP addresses on it the the one I am having problems
> > with and no problems (server been installed and active for 3 or 4 years)
> > This newer server is only since Jun/Jul 2002 and only has about 30 IPs
> > bound to it.
>
> Yeah, of the ~30 machines I run with these nic's (all dell's, too), only
> ONE has ever had this issue, and what-d-ya-know, it's the high-profile www
> box.   Several machines with identical configs, several with slightly
> different configs, none of them have ever experienced this bug.
>
> John
>
>
>
>
>
> --
> John Madden
> UNIX Systems Engineer
> Ivy Tech State College
> jmadden@ivytech.edu
>
>
>
> _______________________________________________
> eepro100 mailing list
> eepro100@scyld.com
> http://www.scyld.com/mailman/listinfo/eepro100