[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-users] arbitrary network unreachable problems


  • To: xen-users@xxxxxxxxxxxxxxxxxxx
  • From: samuel <samu60@xxxxxxxxx>
  • Date: Tue, 29 Sep 2009 13:41:53 +0200
  • Delivery-date: Tue, 29 Sep 2009 04:43:27 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:date:message-id:subject:from:to:content-type; b=dNLGbqhzt1hlggQ8UJEW/KM/Fn98Jx/FAe1z6i8vDH2crdlrqHwscUi7HMtwShtW3P 4jYC7585noDil1DjzD0xmeZVYoibtpn6NZ/XeDYETocxbfaFYHI/BjKfUTGPF7GyDOLU DNpaU6wPlqRwAG+OF+OWbclIlDYAWztmSj+nQ=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>

Hi all,

Recently I've faced a trouble about some networking problem in a Xen server.

dom0
debian lenny kernel 2.6.26-1-xen-686 #1 SMP
xen_caps               : xen-3.0-x86_32p

4 domU all the same
debian lenny kernel  2.6.26-1-xen-686 #1 SMP
with 256M memory

dom0 has access to 4 CPUs:
xm vcpu-list
Name                                ID  VCPU   CPU State   Time(s) CPU Affinity
Domain-0                             0     0     1   r--   47960.4 any cpu
Domain-0                             0     1     0   -b-    1531.2 any cpu
Domain-0                             0     2     1   -b-    1335.4 any cpu
Domain-0                             0     3     2   -b-    2513.8 any cpu
firstdomU                            6     0     3   -b-   39686.3 any cpu
seconddomU                          8     0     3   -b-   20195.7 any cpu
thriddomU                        7     0     3   -b-   19494.5 any cpu
fourthdomU                         9     0     3   -b-   20209.4 any cpu

The issue is that 3 out of the 4 domU, randomly appear Network is unreachable problems as if there were no routes to the destinations although there is. The dom0 is acting as a router for all the domains with next option:
(network-script network-multiple-bridge)

dom0 has 2 interfaces, one internal and an external

pings to the destination from domUs works after the Network is unreachable message appear.

ip_conntrack table is not full at dom0:
# cat /proc/net/ip_conntrack | wc -l
1744
# cat /proc/sys/net/ipv4/netfilter/ip_conntrack_max
65536

netperf checks looked ok, showing around 2G throughpout between domUs and between dom0

routes are ok in all the 4 nodes, pointing at dom0 "internal" address.


The most annoying thing is that one domU does not show the problem and the other 3 do....anyone can point any method to check where the problem may reside? I can provide more info because I've tried to make the e-mail not very long and I think I didn't succeed :(

Any hint will be higly appreciated,
Samuel.

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

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.