significant increase in ipfw pullup failed

[Randy Bush]

>> on a 1ru system with an fxp
>> yesterday, i cvsupped Apr 21 17:00 and built, first time in a few weeks
>> previously i got an ipfw pullup failed a couple of times a day
>> now i get them every few minutes
>>
>> anyone else seeing this, or should i start to send someone into
>> the colo to check cables?
> 
> What it means is that the IPFW code didn't see a long enough packet to 
> contain enough data to be logged usefully.  That could be a sign of 
> hardware problems like a flaky cable or a NIC that is going wonky: if 
> you can check "netstat -i" and look for excessive input errors or 
> collisions.
> 
> If the machine is connected to a managed switch, look for "runt 
> packets", smaller than 54 bytes (56?, whatever the minimum valid size 
> is)....

sorry, did all that before whining.  but, as i might not have,
your asking was quite appropriate.

    # netstat -m
    204 mbufs in use
    134/13120 mbuf clusters in use (current/max)
    0/4/3536 sfbufs in use (current/peak/max)
    319 KBytes allocated to network
    0 requests for sfbufs denied
    0 requests for sfbufs delayed
    0 requests for I/O initiated by sendfile

    # netstat -i
    Name    Mtu Network   Address              Ipkts Ierrs    Opkts Oerrs  Coll
    fxp0   1500 <Link#1>  00:30:48:51:c8:5e 47087068     0 46603594     0     0
    fxp0   1500 RG.NET/24 rip               47394365     - 47237789     -     -
    fxp0   1500 ssh/32    ssh                   5129     -        0     -     -
    fxp1*  1500 <Link#2>  00:30:48:51:c8:5f        0     0        0     0     0
    lo0   16384 <Link#3>                      857260     0   857260     0     0
    lo0   16384 your-net  localhost           223262     -   223262     -     -

it appears to be a plain increase in the reporting of pullup failed.

randy