Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Strange Problems with mtu/oversized packages
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Networking & Security
View previous topic :: View next topic  
Author Message
karl001
n00b
n00b


Joined: 03 Mar 2005
Posts: 14

PostPosted: Sat Oct 27, 2007 10:07 pm    Post subject: Strange Problems with mtu/oversized packages Reply with quote

Hi!

I redid my system (new board/cpu) and switched from a 2 NIC solution to a 1 NIC solution (onboard).
To archive that I set up a vlan tagging and my 3Com Superstack II 3300 seems to work fine with it.

So i have a eth0 (r8169 module) and vlan1/vlan2 as interfaces. vlan1 is my internal network, vlan2 is my DMZ with access to my dsl modems (PPPoE, both Connections terminate on same server as ppp0/1).
The vlan1 interface is set with "REORDER_HDR" to get my dhcpd running, without no client got an ip.

Beginning with this config my network behaves very strange, e.g. I encounter errors while copying files from the server (via samba or any other protocol), the copied files are definitly corrupted, the original sources on the server are definitly ok.

I am not able to find any problem via log/dmesg/interface. The only thing I found is a increasing number of oversized packets on the 3com switch port. I think the mtu is causing this so I put all interfaces to 1472 which was the last mtu shown by a ping -f on from a windows machine.

The problem is still there, till now I found no solution to get rid of that problem.

Here the configs from the system:

For the records the kernel:

Code:
ambient ~ # uname -a
Linux ambient 2.6.22-gentoo-r8 #5 SMP Mon Oct 8 20:24:45 CEST 2007 i686 Intel(R) Core(TM)2 Duo CPU     E6750  @ 2.66GHz GenuineIntel GNU/Linux


Part of the conf.d/net file:

Code:

vlans_eth0="1 2"
config_eth0=( "null" )
mtu_eth0="1472"

vconfig_eth0=( "set_name_type VLAN_PLUS_VID_NO_PAD" )
vconfig_vlan1=( "set_flag 1" "set_egress_map 2 6" )
vconfig_vlan1=( "set_flag 1 REORDER_HDR 1" )

config_vlan1=( "10.20.1.1 netmask 255.255.255.0 brd 10.20.1.255" )
mtu_vlan1="1472"
config_vlan2=( "192.168.1.1 netmask 255.255.255.0 brd 192.168.1.255" )
mtu_vlan2="1472"


3Com Switch Port:

Code:
Port:                1             Port Speed:          100Mbps FD Auto

Received Stats                     Transmit Stats       
--------------                     --------------       
Unicast Packets:     3362106       Unicast Packets:     2241786     
Non Unicast Packets: 15779         Non Unicast Packets: 49085       
Octets:              146543689     Octets:              299843572   
Fragments:           0             Collisions:          0           

Errors
------
Undersize:           0             Oversize:            503849       
CRC Error:           0             Jabbers:             0           

Packet Size Analysis
--------------------
64 Octets:           17272         256 to 511 Octets:   46377       
65 to 127 Octets:    1534898       512 to 1023 Octets:  22692       
128 to 255 Octets:   1095096       1024 to 1518 Octets: 2448572     


Does anybody have any idea how to solve this?
Maybe mtu is not causing this but what else could cause this?
Thanks alot for your time & help!

Karl
Back to top
View user's profile Send private message
Mroofka
Guru
Guru


Joined: 25 Jan 2005
Posts: 369
Location: Poland

PostPosted: Sat Oct 27, 2007 11:57 pm    Post subject: Reply with quote

Maybe I'll not help you but I know that vlan1 is very problematic.

Code:
joshua ~ # vconfig add eth0 1
Added VLAN with VID == 1 to IF -:eth0:-
WARNING:  VLAN 1 does not work with many switches,
consider another number if you have problems.

I don't know why it is in net.example but try change it to 3.

Moreover if I were you I would try to configure it manually step by step.


what does it do? Is this part necessary?
Code:
vconfig_eth0=( "set_name_type VLAN_PLUS_VID_NO_PAD" )
vconfig_vlan1=( "set_flag 1" "set_egress_map 2 6" )
vconfig_vlan1=( "set_flag 1 REORDER_HDR 1" )


Pozdrawiam
Mroofka
_________________
"Make install not love"
registred linux User # 379143

"Ready for Anything; Prepared for everything; Surprised by Nothing !"
Back to top
View user's profile Send private message
karl001
n00b
n00b


Joined: 03 Mar 2005
Posts: 14

PostPosted: Sun Oct 28, 2007 1:46 pm    Post subject: Reply with quote

Quote:

I don't know why it is in net.example but try change it to 3.


I don't think the vlan setting is causing this. I guess the message is for switches not correctly working with vlan1 tags which
is not my problem actually.

Quote:

Moreover if I were you I would try to configure it manually step by step.
what does it do? Is this part necessary?
Code:
vconfig_eth0=( "set_name_type VLAN_PLUS_VID_NO_PAD" )
vconfig_vlan1=( "set_flag 1" "set_egress_map 2 6" )
vconfig_vlan1=( "set_flag 1 REORDER_HDR 1" )



As I wrote the reorder HDR is needed to get my dhcp server working correctly.
The egress map I got from an article but is already removed by me - no healing.

At the current point I don't think the mtu is the problem since I put down the mtu to 1472 now and no more oversized packages are visible on the switch port.
I think the module is not working correctly and this will cause the damaged traffic on the network.
The whole thing is very nicely visible while depacking a rar archive from a samba share, most times the archive seems to be corrupted, on the server itself it can be depacked locally without problems.

With the r8168 module from realtek the whole thing gets better but is still not 100% ok, the problem occurs less often but is still reproduceable.

Is there any problem of that kind know for the realtek drivers? Here the lspci output:

Code:
0000:04:00.0 Class 0200: 10ec:8168 (rev 01)
0000:04:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 01)


maybe an irq problem is causing this, but I can't see any:

Code:
ambient / # more /proc/interrupts
           CPU0       CPU1       
  0:         84          0   IO-APIC-edge      timer
  1:       4680          5   IO-APIC-edge      i8042
  6:        175          2   IO-APIC-edge      floppy
  7:          0          0   IO-APIC-edge      parport0
  8:  998718164  791149953   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 16:   72057836   65737815   IO-APIC-fasteoi   uhci_hcd:usb1, eth0
 17:  595423693  197857763   IO-APIC-fasteoi   ahci, ahci, uhci_hcd:usb5, zaphfc
 18:    2950836    4041091   IO-APIC-fasteoi   dpti0
 19: 3556296636  507462179   IO-APIC-fasteoi   uhci_hcd:usb3, uhci_hcd:usb6, ehci_hcd:usb7, zaphfc
 20:          0          0   IO-APIC-fasteoi   uhci_hcd:usb2
 21:          0          0   IO-APIC-fasteoi   uhci_hcd:usb4, ehci_hcd:usb8
NMI:          0          0
LOC:  282676161  282676160
ERR:          0
MIS:          0


Anyone other ideas?

Karl
Back to top
View user's profile Send private message
cyrillic
Watchman
Watchman


Joined: 19 Feb 2003
Posts: 7313
Location: Groton, Massachusetts USA

PostPosted: Sun Oct 28, 2007 3:59 pm    Post subject: Reply with quote

karl001 wrote:
With the r8168 module from realtek the whole thing gets better but is still not 100% ok, the problem occurs less often but is still reproduceable.

Is there any problem of that kind know for the realtek drivers?

My experience with Realtek ethernet cards (especially the gigabit ones) is that they are crap. Not only are the drivers buggy, but the hardware is too.

If you want to save yourself a lot of trouble, just replace the Realtek NIC with an Intel NIC, and be happy.
Back to top
View user's profile Send private message
karl001
n00b
n00b


Joined: 03 Mar 2005
Posts: 14

PostPosted: Sun Oct 28, 2007 4:28 pm    Post subject: Reply with quote

cyrillic wrote:
My experience with Realtek ethernet cards (especially the gigabit ones) is that they are crap. Not only are the drivers buggy, but the hardware is too.
If you want to save yourself a lot of trouble, just replace the Realtek NIC with an Intel NIC, and be happy.


Errr, I am currently trying to get a Pro 1000PT on Ebay.

But I have to say that I am very dissapointed in the linux hardware support.
I already bought a new Mobo since my one before (Nforce 680SLI chipset+additional sky2 nic's) was absolutely unstable.

forcedeth module, sky2 module all NICs were not running correctly (errors, hangups e.g.).
Additionally the sata module (AHCI or Nforce mode) was also unstable so my raid crashed all the time.

IMHO it seems that newer hardware is an neverending patch & try story under linux, not very amusing if you need the server up and running as me.

Just my 2 cents,
Karl
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Networking & Security All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum