[Micronet] nic question.

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

[Micronet] nic question.

gartim
Network folk,

I have a nic that seems to work fine, but the driver does show dropped
packets, even on a 'idle' connection. ethtool shows no errors.
I do have 802.3ad turned on at the
switch, was testing bonding. Odd thing is when I run wireshark or
tcpdump the errors stop trickling in. I get about 1 every couple of
seconds. any help on this would be greatly appreciated. netpipe
performance is fine, nfs gets about 116MBs, but still worried...


# ethtool -S eth1|grep rx |grep error && ifconfig eth1
     rx_errors: 0
     rx_length_errors: 0
     rx_over_errors: 0
     rx_crc_errors: 0
     rx_frame_errors: 0
     rx_missed_errors: 0
     rx_long_length_errors: 0
     rx_short_length_errors: 0
     rx_align_errors: 0
     rx_csum_offload_errors: 0
eth1 Link encap:Ethernet  HWaddr 00:25:90:51:71:8C  
     inet addr:10.0.2.253  Bcast:10.0.2.255  Mask:255.255.255.0
     inet6 addr: fe80::225:90ff:fe51:718c/64 Scope:Link
     UP BROADCAST RUNNING MULTICAST  MTU:9000  Metric:1
     RX packets:115898464 errors:0 dropped:40139 overruns:0 frame:0
     TX packets:79694146 errors:0 dropped:0 overruns:0 carrier:0
     collisions:0 txqueuelen:1000
     RX bytes:167519905176 (156.0 GiB)  TX bytes:98232460017 (91.4GiB)
         Interrupt:16 Memory:fb900000-fb920000

device showing the error:

[root@n1 gartim]# ethtool --driver  eth1
driver: e1000e
version: 1.5.1-k
firmware-version: 2.1-2
bus-info: 0000:04:00.0
supports-statistics: yes
supports-test: yes
supports-eeprom-access: yes
supports-register-dump: yes

device on other end (shows no errors):

$ ethtool --driver eth1
driver: e1000e
version: 1.10.6-NAPI
firmware-version: 5.6-2
bus-info: 0000:01:00.0
supports-statistics: yes
supports-test: yes
supports-eeprom-access: yes
supports-register-dump: yes


 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] nic question.

Isaac Orr-2
Gary,

I'm obviously keeping folks in the network business busy with work,
since none of them seem to have responded (at least publicly) to this.

It doesn't look like you have actual errors with any frames on the link,
but that the interface/driver is choosing to drop frames rather than
pass them up the stack.  This is normally because the frame is not
considered useful.  It could be that they are 802.3ad related frames, if
you have that fully disabled on the host, but it could be other traffic too.

If your performance is fine, then it is not likely that there's a real
problem here.

You won't see any drops (which are different than errors) when you turn
on tcpdump because it sets the interface into promiscuous mode, which
means it takes and passes on every frame seen on the wire.  I believe
you can set tcpdump/wireshark to not do that.

To find the packets that are causing drops you could:

- set a span/monitor session on your switch to another host running a
sniffer, monitoring the port going to the host in question.
- run a sniffer on the host seeing drops, configured not to set the
interface into promiscuous mode.
- compare both packet traces.  What went out the interface on the switch
but did not turn up in the trace on the host is what is being dropped.

But I would try turning off 802.3ad first on the switch and see if that
changes things... (if you hadn't already).

iso


On 4/26/12 9:25 AM, Gary Artim wrote:

> Network folk,
>
> I have a nic that seems to work fine, but the driver does show dropped
> packets, even on a 'idle' connection. ethtool shows no errors.
> I do have 802.3ad turned on at the
> switch, was testing bonding. Odd thing is when I run wireshark or
> tcpdump the errors stop trickling in. I get about 1 every couple of
> seconds. any help on this would be greatly appreciated. netpipe
> performance is fine, nfs gets about 116MBs, but still worried...
>
>
> # ethtool -S eth1|grep rx |grep error&&  ifconfig eth1
>       rx_errors: 0
>       rx_length_errors: 0
>       rx_over_errors: 0
>       rx_crc_errors: 0
>       rx_frame_errors: 0
>       rx_missed_errors: 0
>       rx_long_length_errors: 0
>       rx_short_length_errors: 0
>       rx_align_errors: 0
>       rx_csum_offload_errors: 0
> eth1 Link encap:Ethernet  HWaddr 00:25:90:51:71:8C
>       inet addr:10.0.2.253  Bcast:10.0.2.255  Mask:255.255.255.0
>       inet6 addr: fe80::225:90ff:fe51:718c/64 Scope:Link
>       UP BROADCAST RUNNING MULTICAST  MTU:9000  Metric:1
>       RX packets:115898464 errors:0 dropped:40139 overruns:0 frame:0
>       TX packets:79694146 errors:0 dropped:0 overruns:0 carrier:0
>       collisions:0 txqueuelen:1000
>       RX bytes:167519905176 (156.0 GiB)  TX bytes:98232460017 (91.4GiB)
>           Interrupt:16 Memory:fb900000-fb920000
>
> device showing the error:
>
> [root@n1 gartim]# ethtool --driver  eth1
> driver: e1000e
> version: 1.5.1-k
> firmware-version: 2.1-2
> bus-info: 0000:04:00.0
> supports-statistics: yes
> supports-test: yes
> supports-eeprom-access: yes
> supports-register-dump: yes
>
> device on other end (shows no errors):
>
> $ ethtool --driver eth1
> driver: e1000e
> version: 1.10.6-NAPI
> firmware-version: 5.6-2
> bus-info: 0000:01:00.0
> supports-statistics: yes
> supports-test: yes
> supports-eeprom-access: yes
> supports-register-dump: yes
>
>
>
> -------------------------------------------------------------------------
> The following was automatically added to this message by the list server:
>
> To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:
>
> http://micronet.berkeley.edu
>
> Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.


 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
Reply | Threaded
Open this post in threaded view
|

Re: [Micronet] nic question.

gartim
Hi iso --

Thanks for the feedback...and clear explaination. I'll try these out,
haven't turned off  802.3ad  yet, and post back, just to have in
archives. Thanks,

-- gary

On Fri, Apr 27, 2012 at 09:23:08AM -0700, Isaac Orr wrote:

>Gary,
>
>I'm obviously keeping folks in the network business busy with work,
>since none of them seem to have responded (at least publicly) to this.
>
>It doesn't look like you have actual errors with any frames on the link,
>but that the interface/driver is choosing to drop frames rather than
>pass them up the stack.  This is normally because the frame is not
>considered useful.  It could be that they are 802.3ad related frames, if
>you have that fully disabled on the host, but it could be other traffic too.
>
>If your performance is fine, then it is not likely that there's a real
>problem here.
>
>You won't see any drops (which are different than errors) when you turn
>on tcpdump because it sets the interface into promiscuous mode, which
>means it takes and passes on every frame seen on the wire.  I believe
>you can set tcpdump/wireshark to not do that.
>
>To find the packets that are causing drops you could:
>
>- set a span/monitor session on your switch to another host running a
>sniffer, monitoring the port going to the host in question.
>- run a sniffer on the host seeing drops, configured not to set the
>interface into promiscuous mode.
>- compare both packet traces.  What went out the interface on the switch
>but did not turn up in the trace on the host is what is being dropped.
>
>But I would try turning off 802.3ad first on the switch and see if that
>changes things... (if you hadn't already).
>
>iso
>
>
>On 4/26/12 9:25 AM, Gary Artim wrote:
>> Network folk,
>>
>> I have a nic that seems to work fine, but the driver does show dropped
>> packets, even on a 'idle' connection. ethtool shows no errors.
>> I do have 802.3ad turned on at the
>> switch, was testing bonding. Odd thing is when I run wireshark or
>> tcpdump the errors stop trickling in. I get about 1 every couple of
>> seconds. any help on this would be greatly appreciated. netpipe
>> performance is fine, nfs gets about 116MBs, but still worried...
>>
>>
>> # ethtool -S eth1|grep rx |grep error&&  ifconfig eth1
>>       rx_errors: 0
>>       rx_length_errors: 0
>>       rx_over_errors: 0
>>       rx_crc_errors: 0
>>       rx_frame_errors: 0
>>       rx_missed_errors: 0
>>       rx_long_length_errors: 0
>>       rx_short_length_errors: 0
>>       rx_align_errors: 0
>>       rx_csum_offload_errors: 0
>> eth1 Link encap:Ethernet  HWaddr 00:25:90:51:71:8C
>>       inet addr:10.0.2.253  Bcast:10.0.2.255  Mask:255.255.255.0
>>       inet6 addr: fe80::225:90ff:fe51:718c/64 Scope:Link
>>       UP BROADCAST RUNNING MULTICAST  MTU:9000  Metric:1
>>       RX packets:115898464 errors:0 dropped:40139 overruns:0 frame:0
>>       TX packets:79694146 errors:0 dropped:0 overruns:0 carrier:0
>>       collisions:0 txqueuelen:1000
>>       RX bytes:167519905176 (156.0 GiB)  TX bytes:98232460017 (91.4GiB)
>>           Interrupt:16 Memory:fb900000-fb920000
>>
>> device showing the error:
>>
>> [root@n1 gartim]# ethtool --driver  eth1
>> driver: e1000e
>> version: 1.5.1-k
>> firmware-version: 2.1-2
>> bus-info: 0000:04:00.0
>> supports-statistics: yes
>> supports-test: yes
>> supports-eeprom-access: yes
>> supports-register-dump: yes
>>
>> device on other end (shows no errors):
>>
>> $ ethtool --driver eth1
>> driver: e1000e
>> version: 1.10.6-NAPI
>> firmware-version: 5.6-2
>> bus-info: 0000:01:00.0
>> supports-statistics: yes
>> supports-test: yes
>> supports-eeprom-access: yes
>> supports-register-dump: yes
>>
>>
>>
>> -------------------------------------------------------------------------
>> The following was automatically added to this message by the list server:
>>
>> To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:
>>
>> http://micronet.berkeley.edu
>>
>> Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.
>
>
>
>-------------------------------------------------------------------------
>The following was automatically added to this message by the list server:
>
>To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:
>
>http://micronet.berkeley.edu
>
>Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.

 
-------------------------------------------------------------------------
The following was automatically added to this message by the list server:

To learn more about Micronet, including how to subscribe to or unsubscribe from its mailing list and how to find out about upcoming meetings, please visit the Micronet Web site:

http://micronet.berkeley.edu

Messages you send to this mailing list are public and world-viewable, and the list's archives can be browsed and searched on the Internet.  This means these messages can be viewed by (among others) your bosses, prospective employers, and people who have known you in the past.