[tulip-bug] LNE-100tx and IDE trouble

Pete Holzmann Tech@ICTA.Net
Sat, 27 Jan 2001 21:24:58 -0700


>On 27 Jan 2001, you said...
>
>>You have debugging turned on to the highest level.  This will put a
>>large load on the machine as the messages are logged to the disk.
>>
>>Try turning debugging off.

"debug=0" doesn't seem to be enough. I'll remove it completely.

OK, that makes ttcp happy, thanks! Next test is with my cool new backup system: 
Ecrix VXA tape drive on high end (Tekram) scsi board (that's what I REALLY want 
to have running well...) Works very cleanly except when pulling in data across 
the network (client machines have an application that pushes data to the linux 
server...)

Below are some of the /var/log/message errors I get with this setup (clearly, I 
am getting a LOT of these :-(...

I tried setting max_interrupt_work to a huge number (1000) and to 30. Neither 
helped.... Any other ideas?

thx,
pete

Jan 27 21:17:02 rose kernel: eth1: Too much work during an interrupt, csr5=0xfc6
9c0d0.
Jan 27 21:17:02 rose kernel: eth1: Restarted Rx at 1087632 / 1087632.
Jan 27 21:17:03 rose kernel: eth1: Too much work during an interrupt, csr5=0xfc6
74050.
Jan 27 21:17:04 rose kernel: eth1: Too much work during an interrupt, csr5=0xfc6
74050.
Jan 27 21:17:04 rose kernel: eth1: Too much work during an interrupt, csr5=0xfc6
9c0d0.
Jan 27 21:17:04 rose kernel: eth1: Restarted Rx at 1090657 / 1090657.
Jan 27 21:17:05 rose kernel: eth1: Too much work during an interrupt, csr5=0xfc6
9c0d0.
Jan 27 21:17:05 rose kernel: eth1: Restarted Rx at 1091803 / 1091803.
Jan 27 21:17:07 rose kernel: eth1: Too much work during an interrupt, csr5=0xfc6
9c0d0.
Jan 27 21:17:07 rose kernel: eth1: Restarted Rx at 1093887 / 1093887.

---
Pete Holzmann             Pete@ICTA.net   12175 Howells Road
President/CEO                 Colorado Springs, CO 80908 USA
ICTA(r)                            Tel/Fax: [1](719)495-8789
"Networking for Global Breakthrough"(r)  http://www.icta.net
    International Christian Technologists' Association