MPD Issues - Errors and Mysterious Packet Loss

[Dan Hardiker]

> try from the XP Box:
>
> ping -l 1500 www.yahoo.de
>
> If you get no reply then try decreasing the ICMP packet size (-l switch),
> until you get a reply.

Pinging 192.168.1.3 (internal ip of MPD server) from 192.168.1.101 (assigned
ip of xp box) as external traffic isnt routed via vpn's gateway.

1500 - no reply x4
1400 - no reply x4
1300 - no reply x3 + 1 reply on the 3rd attempt
1200 - no reply x2 + 2 reply on the 1st / 4th attempt
1100 - no reply x3 + 1 reply on the 2nd attempt
1000 - no reply x2 + 2 reply on the 1st / 4th attempt
900  - no reply x3 + 1 reply on the 2nd attempt
800  - no reply x2 + 2 reply on the 1st / 4th attempt
700  - no reply x3 + 1 reply on the 2nd attempt
// stopped the pinging for about 5 mins here //
600  - no reply x1 + 3 reply on the 1st / 2nd / 3rd attempt
500  - no reply x1 + 3 reply on the 1st / 2nd / 3rd attempt
400  - no reply x2 + 2 reply on the 1st / 2nd attempt
300  - no reply x2 + 2 reply on the 1st / 4th attempt
200  - no reply x2 + 2 reply on the 1st / 4th attempt
100  - no reply x2 + 2 reply on the 1st / 2nd attempt
50   - no reply x1 + 3 reply on the 1st / 2nd / 3rd attempt

Heres a bit of info for you - while I doing some of the duplicate tests
(pinging 1000 more than just 1 set of 4 for example) I though I would try
pinging 192.168.1.101 from the mpd server to see if I could get through.
Heres what I found:

[10:29:32][root at armadillo]:~$ ping 192.168.1.101
PING 192.168.1.101 (192.168.1.101): 56 data bytes
ping: sendto: No buffer space available
ping: sendto: No buffer space available
ping: sendto: No buffer space available
64 bytes from 192.168.1.101: icmp_seq=3 ttl=128 time=0.922 ms
64 bytes from 192.168.1.101: icmp_seq=4 ttl=128 time=0.889 ms
64 bytes from 192.168.1.101: icmp_seq=5 ttl=128 time=0.882 ms
64 bytes from 192.168.1.101: icmp_seq=6 ttl=128 time=0.877 ms
64 bytes from 192.168.1.101: icmp_seq=7 ttl=128 time=0.871 ms

NOTE: the TTL on the pings from the XP box are notabley 64 where as the mpd
server has them set to 128

Buffer space seems to be running out somewhere?

PS: I applied the TCP MSS patch, and even with TCP web traffic, Im still
clocking up receiving errors by the dozen.

I hope that helps

--
Dan Hardiker [dhardiker at eorigen.com]
ADAM Systems Engineer
eOrigen Ltd