Similar problem
Instant Success but serious problems..... (Nils Hammer)
I have two systems running Michael's NAT code: one is production, the other is a backup/test system. I made the backup/test system operational first and over a 2.5 month period, I got 35 errors similar to the error Nils Hammer is reporting. My Call Trace is always the same and is different from Nils': syslog:Oct 26 01:33:25 fw2 kernel: Call Trace: [ip_build_xmit+1039/3064] [alloc_skb+90/332] [dev_queue_xmit+35/68] [arp_send+537/548] [eth_header+0/380] [arp_check_expire+279/368] [arp_check_expire+0/368] So far, this error never seems to have caused any kind of problem. When I made the production system operational, I watched for this kind of error. I've only seen one error in 1.5 months: Oct 23 12:09:40 fw1 kernel: Call Trace: [ip_rcv+947/980] [sock_wake_async+50/108] [sys_shutdown+55/96] [arp_free_entry+85/112] [kfree_skb+136/224] [raw_rcv_redo+103/124] Again, this error does not seem to have caused a problem. The backup/test system is about four years old. It's an NCR 3330 with a 33MHz 486, 16MB of memory, and a 230MB IDE drive. December, 1996, Slackware distribution. The production system is a three year old Gateway P5-90 with 40MB of memory and a 512MB IDE drive. April, 1997, Slackware distribution. Both systems are running the same version of NAT, both are running 2.0.29, but different versions of Slackware. Can anyone else shed any light on this problem? Dan |
