eth0 goes down after a few days up

CRego3D · 03-26-2004, 05:27 PM

I have the weirdest thing happening.

Out of the blue, after the server is up for many days, these errors start loggin in

Quote:

Mar 26 16:18:09 web kernel: NETDEV WATCHDOG: eth0: transmit timed out
Mar 26 16:18:09 web kernel: eth0: Transmit timed out: status 0090 0c00 at 7847937/7847997 command 000c0000.
Mar 26 16:18:09 web kernel: Command 0070 was not accepted after 20001 polls! Current status 01700000.
Mar 26 16:18:09 web kernel: eth0: wait_for_cmd_done timeout!
Mar 26 16:18:40 web last message repeated 30 times
Mar 26 16:19:13 web last message repeated 30 times

after 1 or two hours, the eth0 will drop and the server wont be accessible from the network

Thsi is RH 7.3 with the latest kernel

Any ideas ?

Pwnz3r · 03-26-2004, 05:30 PM

Have you tried not using Watchdog or is it just a logging daemon?

CRego3D · 03-26-2004, 05:54 PM

no, this is simply from /var/log/messages

weird thing is, it happens out of the blue .. the system might be up for 5 days, I look in the logs, there have been no previous eth0 errors, then all of the suddent they start, and 2 hours later after they start .. eth0 is out

Pwnz3r · 03-26-2004, 06:47 PM

Odd, but since I haven't had the same problems I suppose that I can't be of too much help.

verdeboy2k · 03-26-2004, 06:51 PM

Could the hardware be fried? (Overheat, Bad connection, loose wires, etc.)

CRego3D · 03-27-2004, 05:25 AM

very dubfull, the card is onboard (nic) .. and the system is cooled properly

verdeboy2k · 03-27-2004, 10:49 AM

hmmm, have you checked the blinking lights while all this is happening... that error message sounds like either the hardware went kaput or the device on the other end of the wire is having trouble. (I get similar errors when my cable modem connection decides to die.)