[gentoo-user] Packet loss unknown issue

[gentoo-user] Packet loss unknown issue

[José Pedro Saraiva]

Hello,
I'm quite desperate and have tried just about anything to understand my issue.
I'll try to explain it...

I connect to the net through eth0 and have a masquerading through eth1
to my other computers, using shorewall.
My server never had any problems of this kind and all was fine until a
couple of days ago. I'm experiencing some heavy packet loss when
pinging different hosts.
I have isolated the problem (at least I think I did) and got to the
conclusion that issuing a simple 'ifconfig eth1 down' and leaving eth0
up, the packet loss vanishes.
Tests were made with all services down (shorewall, sshd, apache2,
mysqld, proftpd, sendmail, ...)
Here are the statistics:


- eth1 up -
# ping -c 200 google.com

--- google.com ping statistics ---
200 packets transmitted, 104 received, 48% packet loss, time 208227ms
rtt min/avg/max/mdev = 152.388/229.520/2016.847/279.805 ms, pipe 3

- eth1 down -
ping -c 200 google.com

--- google.com ping statistics ---
200 packets transmitted, 200 received, 0% packet loss, time 199175ms
rtt min/avg/max/mdev = 147.197/169.978/193.949/8.384 ms



Both network cards are using the same chipset (RLT8139) loaded as a
module in the kernel (2.6.11-hardened-r13).
I have tried to replace the network card (with one also using RLT8139)
and changing it from IRQ but with no effect.
Please help me, I've runned out of ideas...

Thanks in advance!
José Pedro

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Packet loss unknown issue

[Robert Svoboda]

* José Pedro Saraiva <nocive@gmail.com> [2005-07-12 23:30]:
> Hello,

Hi,

> I connect to the net through eth0 and have a masquerading
> through eth1 to my other computers, using shorewall.  

Uh, I'm not sure i understand you.

your localhost eth0 -> shorewall -> Inet
your localhost eth1 -> shorewall -> your other machines

Is above right?
What address ranges do you use?

> My server never had any problems of this kind and all was
> fine until a couple of days ago. I'm experiencing some heavy
> packet loss when pinging different hosts.  I have isolated
> the problem (at least I think I did) and got to the
> conclusion that issuing a simple 'ifconfig eth1 down' and
> leaving eth0 up, the packet loss vanishes.  Tests were made
> with all services down (shorewall, sshd, apache2, mysqld,
> proftpd, sendmail, ...)

What about routing? As root run:
$ route

or tracepath?
$ tracepath google.com

Robert
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Packet loss unknown issue

[José Pedro Saraiva]

> your localhost eth0 -> shorewall -> Inet
> your localhost eth1 -> shorewall -> your other machines
correct :)
but shorewall is not responsible for the problem, since I ran all the
tests with shorewall stopped and all other network services stopped as
well.

> What address ranges do you use?
/etc/conf.d/net
iface_eth1="192.168.0.254 netmask 255.255.0.0"
iface_eth0="dhcp"

> What about routing? As root run:
> $ route
# route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
192.168.0.0     *               255.255.0.0     U     0      0        0 eth1
loopback        *               255.0.0.0       U     0      0        0 lo
default         195-23-73-156.n 0.0.0.0         UG    0      0        0 eth0

> tracepath with eth1 up
# tracepath google.com
 1:  195-23-73-156.net.novis.pt (195.23.73.156)             0.475ms pmtu 1500
 1:  no reply
 2:  no reply
 3:  no reply
 ...

> tracepath with eth1 down
# tracepath google.com
 1:  195-23-73-156.net.novis.pt (195.23.73.156)             0.713ms pmtu 1492
 1:  router (192.168.1.254)                                 0.817ms
 2:  r6613.nr.ip.pt (195.23.0.154)                         73.028ms
 3:  195-23-31-126.nr.ip.pt (195.23.31.126)                72.102ms
 4:  195-23-31-165.nr.ip.pt (195.23.31.165)                73.625ms
 5:  194-79-92-82.nr.ip.pt (194.79.92.82)                 asymm  7  76.323ms
 6:  194-79-92-21.nr.ip.pt (194.79.92.21)                 asymm  5  77.538ms
 7:  g1-2-202.core01.lis01.atlas.cogentco.com (130.117.19.189) asymm 
6  77.528ms
 8:  p2-2.core01.mad05.atlas.cogentco.com (130.117.0.121) asymm  7  87.644ms
 9:  p5-0.core01.bcn01.atlas.cogentco.com (130.117.1.105) asymm  8  98.060ms
10:  p5-0.core01.par01.atlas.cogentco.com (130.117.1.110) asymm  8 118.240ms
11:  p5-0.core02.par02.atlas.cogentco.com (130.117.1.229) asymm 10 119.976ms
12:  p15-0.core01.fra03.atlas.cogentco.com (130.117.0.17) asymm 11 128.401ms
13:  no reply
14:  no reply
15:  no reply
16:  no reply
17:  no reply
18:  no reply
19:  no reply
20:  no reply
21:  no reply
22:  no reply
23:  no reply
24:  no reply
25:  no reply
26:  no reply
27:  no reply
28:  no reply
29:  no reply
30:  no reply
31:  no reply
     Too many hops: pmtu 1492
     Resume: pmtu 1492

192.168.1.254 is my ADSL modem that is connected to eth0.

Thanks for the reply Robert

-- 
gentoo-user@gentoo.org mailing list

RE: [gentoo-user] Packet loss unknown issue

[Dave Nebinger]

> > What about routing? As root run:
> > $ route
> # route
> Kernel IP routing table
> Destination     Gateway         Genmask         Flags Metric Ref    Use
> Iface
> 192.168.0.0     *               255.255.0.0     U     0      0        0
> eth1
> loopback        *               255.0.0.0       U     0      0        0 lo
> default         195-23-73-156.n 0.0.0.0         UG    0      0        0
> eth0

What's the route when eth1 is down?



-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Packet loss unknown issue

[Petr Kocmid]

On Wednesday 13 of July 2005 18:07, José Pedro Saraiva wrote:

> iface_eth1="192.168.0.254 netmask 255.255.0.0"

Your modem/router which is 192.168.1.254 on eth0 fits the netmask of the eth1 
network. 

-- 
Petr

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Packet loss unknown issue - SOLVED :)

[José Pedro Saraiva]

On 7/13/05, Petr Kocmid <Petr.Kocmid@project-bhairava.org> wrote:
> Your modem/router which is 192.168.1.254 on eth0 fits the netmask of the eth1
> network.
> 
Weird thing is that this configuration as always been working for me,
with no problems whatsoever... until now.

Anyways, thank you!
Problem solved!

-- 
gentoo-user@gentoo.org mailing list