Messages in this thread | | | Date | Sun, 20 Jun 1999 05:47:57 -0400 (EDT) | From | glouis@dynamicr ... | Subject | Re: Network bug 2.2.* - reboot linux every week? |
| |
On Fri, 18 Jun 1999, Alan Cox wrote:
> tcpdump data from some clients and the server when this occurs might > make this debuggable. Otherwise it could be anything on your network, even > the switches > Our network had problems not unlike this. Eventually traced to a defective NIC on a Windows box elsewhere in the LAN. Interesting points: (1) bad packets from bad NIC never caused anything worse than slowdowns till the second 2.2 Linux box appeared on the network, after which the first 2.2 box started going crazy as described by Dietmar; (2) rebooting said first 2.2 box with 2.0.37pre10 or pre12 made the problem go away but only temporarily; within a few days the machine would crash with "too much work in interrupt" -- rebooting again with 2.2.7 or 2.2.9 would cause network flood within 5 minutes. Removing defective NIC (once we found it) eliminated the bad behaviour, but obviously there's a vulnerability in 2.2 somewhere. I have Dietmar's problem that these are production boxen and I can't experiment much (it's fiscal yearend chez nous and people are working at all hours). I've asked that the bad NIC be delivered to me so I can set up a test mini-network and see if I can reproduce the troubles. More later if I get anything interesting; just wanted to issue a heads-up that there may be something real here to worry about.
-- Greg Louis
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |