[Xen-users] network hangs / packet losses - can't find a solution

[Tomasz Chmielewski <mangoo@xxxxxxxx>]

I saw some other users on the list also had "network hangs" / packet 
losses, but I didn't find a solution.
What happens?

There are several symptoms:

1. When logged into a Guest from some other host in LAN via SSH, every 
few minutes (sometimes seconds), you can see that the Guest doesn't 
respond (i.e., you type, and it doesn't appear on the screen); after a 
few seconds it appears
2. Can's SSH from domain0 to guest domains:

# ssh -v 192.168.111.171
OpenSSH_4.2p1, OpenSSL 0.9.7g 11 Apr 2005
debug1: Reading configuration data /etc/ssh/ssh_config
debug1: Applying options for *
debug1: Connecting to 192.168.111.171 [192.168.111.171] port 22.
debug1: Connection established.
debug1: permanently_set_uid: 0/0
debug1: identity file /root/.ssh/identity type -1
debug1: identity file /root/.ssh/id_rsa type -1
debug1: identity file /root/.ssh/id_dsa type -1
debug1: Remote protocol version 2.0, remote software version OpenSSH_4.2
debug1: match: OpenSSH_4.2 pat OpenSSH*
debug1: Enabling compatibility mode for protocol 2.0
debug1: Local version string SSH-2.0-OpenSSH_4.2
debug1: SSH2_MSG_KEXINIT sent

(Here it waits forever)

I can log in from one guest domain to the other without problems.

3. Packet losses (with big packet size):

a) from domain0 to other hosts in LAN, and the way around - no losses:

ping -f -c 1000 -s 20000 192.168.111.133

b) from/to Guest domains - about 7% losses:

# ping 192.168.111.171 -f -s 20000 -c 1000
PING 192.168.111.171 (192.168.111.171) 20000(20028) bytes of data.
........................................................................
--- 192.168.111.171 ping statistics ---
1000 packets transmitted, 928 received, 7% packet loss, time 10817ms
rtt min/avg/max/mdev = 6.155/10.317/695.355/38.685 ms, pipe 3, ipg/ewma 
10.828/7.298 ms
No losses with smaller packet sizes though.


The issue 1) is particularly annoying ("hangs" when SSHing), but the 
issue 2) is also interesting (no ability to SSH from domain0).
It all reminds me some mysterious MTU issues, but I tried setting 
(guessing?) different values on both domain0 and guestes, but it didn't 
help.
I use Xen 3.0 + bridging.

--
Tomasz Chmielewski
http://wpkg.org

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users