freebsd stable - Dec 2006 - bge Ierr rate increase from 5.3R -> 6.1R

Hello

I recently updated two production servers from 5.3 to 6.1 via cvsup  
and buildworld. Since the upgrade I've seen an increase in the number  
of Input packet errors reported on the bge cards in on both boxes.  
One is a HP DL360g3, the other is a HP DL380g3. Both have a pair of  
2.8GHz Xeons with a SMP kernel.

So - was the old driver previously underreporting, is the new one  
over-reporting/causing the error rate or is it something else? Cables  
and cabling have not changed and the pickup in number of errors is  
quite distinct. I monitor the nightly 'periodic daily' phone homes  
closely.

Example from the DL380g3.

The 5.3->6.1 upgrade was on 8 November. 58,000 errors in 1 month  
compared to 2 errors in 1 year with 5.3.

Network interface status:
Name    Mtu Network       Address              Ipkts Ierrs    Opkts  
Oerrs  Coll
bge0   1500 <Link#1>      00:0f:20:f6:**:** 1344182650 58292  
2701993948     0     0
bge0   1500 192.168.**/25 **********        1344176611     -  
2701984851     -     -
bge1*  1500 <Link#2>      00:0f:20:f6:**:**        0     0         
0     0     0
lo0   16384 <Link#3>                           69549     0     
69549     0     0
lo0   16384 your-net      localhost            69549     -     
69549     -     -

A couple of weeks ago I turned off tx and rx check summing on this  
box as I gathered from googling it might be contributing. That had no  
effect.

Upon further investigation it appears six other boxes with bge ports  
(mostly HP DL360g4) running 6.1 started reporting errors when moved  
to 6.1. As they do only a small fraction of the traffic that the  
above box does I hadn't noticed it.

This box (a UP HP DL360g4) is on a completely different network,  
different switch, cabling etc. Again, prior to 6.1 it had never  
reported an error in 18 months of service.

Network interface status:
Name    Mtu Network       Address              Ipkts Ierrs    Opkts  
Oerrs  Coll
bge0   1500 <Link#1>      00:12:79:3b:**:** 781001814  1980  
1056534485     0     0
bge0   1500 192.168.***   192.168.***.***   783877018     -  
1061029115     -     -

I don't have a spare box with a bge interface to test 6.2 for the  
same behaviour, but would be interested if anyone had an explanation.

Best wishes.

Greg.

Greg Eden wrote:
> Hello
> 
> I recently updated two production servers from 5.3 to 6.1 via cvsup and
> buildworld. Since the upgrade I've seen an increase in the number of
> Input packet errors reported on the bge cards in on both boxes. One is a
> HP DL360g3, the other is a HP DL380g3. Both have a pair of 2.8GHz Xeons
> with a SMP kernel.
It would be quite useful at this point if you could update a box or
two to the RELENG_6_2 code base so that we can see if this problem is
solved in the latest release candidate. If it is, your problem is
solved, and if it's not, you're a lot closer to the point where we can
usefully assist you.

Doug

-- 

    This .signature sanitized for your protection

On 12/11/06, Greg Eden <greg@warprecords.com>
wrote:
> I recently updated two production servers from 5.3 to 6.1 via cvsup
> and buildworld. Since the upgrade I've seen an increase in the number
> of Input packet errors reported on the bge cards in on both boxes.
> One is a HP DL360g3, the other is a HP DL380g3. Both have a pair of
> 2.8GHz Xeons with a SMP kernel.
Just to be sure, is polling disabled?

D> > Greg Eden wrote:
D> >> Hello
D> >>
D> >> I recently updated two production servers from 5.3 to 6.1 via  
D> >> cvsup and
D> >> buildworld. Since the upgrade I've seen an increase in the
number of
D> >> Input packet errors reported on the bge cards in on both boxes.  

In 5.3-RELEASE the bge(4) driver did not read the error count from the
chip at all. So errors were not accounted.

-- 
Totus tuus, Glebius.
GLEBIUS-RIPN GLEB-RIPE