many network ierrors

[proehl]

I'm having problems with an ethernet interface and noticed many ierrors, but no oerrors or collisions.

netstat -i reports:

Name Mtu Net/Dest Address Ipkts Ierrs Opkts Oerrs Collis Queue
lo0 8232 loopback localhost 296952415 0 296952415 0 0 0

hme0 1500 bounty bounty 1197663773 600 1558594611 0 0 0
qfe0 1500 darkmoon darkmoon 220823146 26365 175770864 0 0 0

qfe1 1500 postbounty postbounty 333657959 110224 335406201 0 0 0

qfe1:1 1500 postbounty-dns postbounty-dns 0 0 0 0 0 0

qfe2 1500 zeitung zeitung 626404022 127855 733275128 0 0 0

Any ideas?
If it would be a hardware/Cable/Switch Problem I would aspect to see also oerrors?

 

[AHinMaine]

On occasion we find problems like this when we don't manually set to 100fdx. Because of this it is our standard practice to do this on all of our sun systems as well as the port config on the cisco side.

/etc/system

set hme:hme_adv_autoneg_cap=0
set hme:hme_adv_100T4_cap=0
set hme:hme_adv_100fdx_cap=1
set hme:hme_adv_100hdx_cap=0
set hme:hme_adv_10fdx_cap=0
set hme:hme_adv_10hdx_cap=0 

set qfe:qfe_adv_autoneg_cap=0
set qfe:qfe_adv_100T4_cap=0
set qfe:qfe_adv_100fdx_cap=1
set qfe:qfe_adv_100hdx_cap=0
set qfe:qfe_adv_10fdx_cap=0
set qfe:qfe_adv_10hdx_cap=0

If you don't want to reboot, you can change it on the fly with the ndd command.

ndd -set /dev/hme instance 0
ndd -set /dev/hme adv_autoneg_cap 0

...etc

 

[Annihilannic]

Which interface is giving you problems?

You are not actually having very many errors if you calculate them as a percentage of the packets that have gone through the interface. Your statistics look quite normal to me. Annihilannic.

 

[proehl]

The actual problem is a bad response time on the qfe interfaces.
When I connect a Mac-Workstation to the hme0 interfaces it will process a (serverbased) file 10 times faster than it does it on one of the qfe Interfaces.
Therefore I think the problem is very close to the hardware. If the problem would be related to Disk, Mem or CPU one should see the same timing on every interface, I think.


When you look at the netstat output you see that there are much more errors on the qfe-Interface, but only on the input side, there are no output -Errors at all!
Strange

Int Ipkts  Ierrs
hme0 1.197.663.773 600
qfe1 333.657.959 110224

 

[Annihilannic]

Have you tried naChoZ suggestion? It would make sense if your switch is set to full duplex and your qfe to half duplex for you to see more input errors than output errors.

Do the following to find out if qfe1 is at full duplex:

[tt]ndd -set /dev/qfe instance 1
ndd -get /dev/qfe link_mode[/tt]

If it's 1 you are already at full duplex. Annihilannic.

 

[mikeclark]

If you have locked the network interface to 100Mb FDX you will need to lock the switch port to 100Mb FDX as well. Sun and some switches (mostly Cisco)don't play nice with autonegotiation.

 

[proehl]

I just set the interface and the Switch to 100 MBit Full duplex. Looks OK now. Have to watch it for some time just to be shure.

Thank's for your help.