I've been experiencing some sporadic slowness (perceived, no hard
measurement) on my local Lan subnet. When I started looking into it I
found a lot of errors on my eth1, the lan subnet in question:
# ip -s link show eth1
4: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:0c:41:e9:34:dd brd ff:ff:ff:ff:ff:ff
RX: bytes packets errors dropped overrun mcast
579054358 3829354 0 0 0 0
TX: bytes packets errors dropped carrier collsns
0 0 3643158 0 3643158 0
# ip -s -s link show eth1
4: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:0c:41:e9:34:dd brd ff:ff:ff:ff:ff:ff
RX: bytes packets errors dropped overrun mcast
579095638 3829655 0 0 0 0
RX errors: length crc frame fifo missed
0 0 0 0 0
TX: bytes packets errors dropped carrier collsns
0 0 3643405 0 3643405 0
TX errors: aborted fifo window heartbeat
0 0 0 3619749
I believe I've tracked the problem to a misconfiguration of hardware,
but would like a sanity check before I start rebuilding my network.
Here's the diagram in classic ASCII art:
+-+
| | Motorola Surfboard Cable Modem
| | (dynamic IP from Comcast)
+-+
|
| eth0:
+---------+
| | Bering uClibc/Shorewall
+---------+
eth1: | | eth2:
| |
| +-----> DMZ (Website)
|
+-------------+
| | NETGear DS108 10/100
+-------------+ 8 port HUB
Business ||| |
Computers <-----+|| |
Printer <------+| +----+
NetGear SC101 <---+ | | Linksys Etherfast 10/100
+----+ 5 port SWITCH
| |
+------+ +-------> XBox 360 (with XBox LIVE)
|
+-----+
NetGear DS104 | |
4 port HUB +-----+
| |
| |
Home Computer <-+ |
Spouse's Work Lap Top <--+
(Occasionally)
If the ASCII art gets mangled, the important detail is this:
firewall/router(eth1:)--->DS108 HUB-->EtherFast 10/100 SWITCH-->DS104 HUB
My theory is that the problem lies in having the Linksys SWITCH
between the two NetGear HUBs. I only recently (hangs head) learned
the real difference between HUBs and SWITCHes. The Linksys SWITCH
only talks full duplex upstream; My understanding of the NETGear doc
says that it only talks half duplex. So while my network is
functioning, I'm losing half the packets at the SWITCH -> HUB
connection, and that is the cause of the heartbeat errors I'm seeing.
While on the diagram it looks easy enough to reconfigure in order to
put the SWITCH directly connected to eth1: with the two NETGear HUBS
connected to the switch
firewall/router(eth1:)-->EtherFast 10/100 SWITCH-->(DS108 and DS104 in
separate SWITCH ports)
Physically, that means some movement of gear
between floors and likely purchasing another UPS (moving cable modem,
firewall and probably the SC101). Alternatively, I could replace the
DS108 HUB with an FS108 8 port SWITCH for less than a UPS would cost.
Here's the sanity check: Does the HUB/SWITCH misconfiguration theory
fit with the errors reported? Or is it really more likely a bad NIC
somewhere?
How would I track down a bad NIC?
Thanks in advance!
Ken
PS: LEAF, Bering uClibc and Shorewall are just an unbeatable
combination! Thanks to everyone who make it possible.
Details on the firewall follow:
LEAF CONFIG DETAILS:
====================
Pentium 2, 100Mhz, 168Mb RAM
Bering uClibc, v3.0.1
3 Linksys Etherfast 10/100 NICs
Fairly standard 3 card setup, Local Lan with DMZ
# uname -a
Linux 2.4.33 #1 Sun Jan 14 12:15:07 CET 2007 i686 unknown
# ip addr show
1: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
2: dummy0: <BROADCAST,NOARP> mtu 1500 qdisc noop
link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff
3: eth0: <BROADCAST,MULTICAST,NOTRAILERS,UP> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:0c:41:ec:40:78 brd ff:ff:ff:ff:ff:ff
inet 69.253.57.107/21 brd 255.255.255.255 scope global eth0
4: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:0c:41:e9:34:dd brd ff:ff:ff:ff:ff:ff
inet 192.168.225.254/24 brd 192.168.225.255 scope global eth1
5: eth2: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:0c:41:e9:36:79 brd ff:ff:ff:ff:ff:ff
inet 192.168.254.254/24 brd 192.168.254.255 scope global eth2
# ip rout show
192.168.225.0/24 dev eth1 proto kernel scope link src 192.168.225.254
192.168.254.0/24 dev eth2 proto kernel scope link src 192.168.254.254
69.253.56.0/21 dev eth0 proto kernel scope link src 69.253.57.107
default via 69.253.56.1 dev eth0
# lsmod
Module Size Used by Not tainted
softdog 1392 1
ipt_ipp2p 5908 0
ipt_state 272 31
ipt_helper 400 0 (unused)
ipt_conntrack 692 0
ipt_REDIRECT 480 0 (unused)
ipt_MASQUERADE 1024 2
ip_nat_irc 1704 0 (unused)
ip_nat_ftp 2152 0 (unused)
iptable_nat 14452 3 [ipt_REDIRECT ipt_MASQUERADE
ip_nat_irc ip_nat_ftp]
ip_conntrack_irc 2484 1
ip_conntrack_ftp 3132 1
ip_conntrack 16548 2 [ipt_state ipt_helper ipt_conntrack
ipt_REDIRECT ipt_MASQUERADE ip_nat_irc ip_nat_ftp iptable_nat
ip_conntrack_irc ip_conntrack_ftp]
tulip 36044 3
crc32 2620 0 [tulip]
vfat 8492 0 (unused)
isofs 15700 0
ide-detect 132 0 (unused)
ide-cd 26812 0
ide-disk 11436 0
ide-core 88408 0 [ide-detect ide-cd ide-disk]
cdrom 25568 0 [ide-cd]
--
J. Kenneth Gentle (Ken)
Gentle Software LLC
Phone: 484.371.8137
Mobile: 302.547.7151
Email: [EMAIL PROTECTED]
Email: [EMAIL PROTECTED]
www.gentlesoftware.com
-------------------------------------------------------------------------
SF.Net email is sponsored by:
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services
for just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
------------------------------------------------------------------------
leaf-user mailing list: leaf-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/leaf-user
Support Request -- http://leaf-project.org/
