Re: [Xen-users] Vif dropping packets

[Peter Booth <peter_booth@xxxxxx>]

Basically i wanted to know if any resource consyraints exist.

Do packet drops occur all the time?

1. What about with a steady state load?
2. What about no load?
3. What about a heavy load?




There's an IBM study that shows xen systems with more than 12vms AND  
CPU overcommitnent can have unusual delays responding to requests.
I've seen similar behavior but only  on a busy host with busy vms.

I think that you need more data.



On Jun 26, 2009, at 6:26 PM, Mike Lovell <mike@xxxxxxxxxxxx> wrote:

> I'm not entirely sure what you are asking with that second question.  
> sar is installed in the guests. I did `sar -u1 1000` to watch the  
> status for a while. Here is an abbreviated output with the averages  
> and the line I saw with the highest %steal
> 04:13:12 PM       CPU     %user     %nice   %system   %iowait     
> %steal     %idle
> 04:13:16 PM       all      0.00      0.00      0.00      0.00       
> 0.99     99.01
> Average:          all      0.01      0.00      0.02      0.55       
> 0.05     99.37
> Overall, these vms aren't doing a lot. Mainly sitting idle waiting  
> for a QA engineer to test code functionality. I should probably say  
> a little more about the config. The host boxes are dual quad-core  
> Opteron 2346HE systems with 24GB of memory and a ton of disks.  
> Currently there are 28 virtual machines running on the one I have  
> been checking but most of them idle.
> Hope that helps or is what you were looking for.
>
> mike
>
> Peter Booth wrote:
> > Mike,
> >
> > You don't say anything about the workload on your system or the  
> > resource consumption that it causes.
> > Do you have sar installed on the domUs? Does a one second vmstat  
> > show a %st that's above 1%?
> > Sent from my iPhone
> >
> > On Jun 26, 2009, at 4:22 PM, Mike Lovell <mike@xxxxxxxxxxxx> wrote:
> >
> > > I have a problem with packets being dropped on some vif  
> > > interfaces. I currently have a box running Debian Lenny with Xen  
> > > 3.2.1. The host is running the Xen-ified kernel from the Debian  
> > > repos. 2.6.26-1-xen-amd64. The host has a single bridge, 'vmnet',  
> > > that connects to one physical interface and all of the guest's  
> > > interfaces connect to. If I do a 'ifconfig vifX.0', I see dropped  
> > > TX packets. Looking at the guest does not show any dropped packets  
> > > on its interface.
> > > From the host
> > > # cat /etc/network/interfaces
> > > auto vmnet
> > > iface vmnet inet static
> > >      address 10.135.2.71
> > >      netmask 255.255.255.224
> > >      bridge_ports eth1
> > >      bridge_stp off
> > >      bridge_fd 9
> > >      bridge_hello 2
> > >      bridge_maxage 12
> > >
> > >
> > > # brctl show vmnet
> > > bridge name    bridge id        STP enabled    interfaces
> > > vmnet               8000.003048c8166d    no        eth1
> > >     
> > >     
> > >                                                                   
> > > vif22.0
> > >     
> > >     
> > >     
> > >                                                                ...  
> > > <lots of interfaces>
> > > # ifconfig vif22.0
> > > vif22.0   Link encap:Ethernet  HWaddr fe:ff:ff:ff:ff:ff           
> > > inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
> > >        UP BROADCAST RUNNING PROMISC MULTICAST  MTU:1500  Metric:1
> > >        RX packets:7946252 errors:0 dropped:0 overruns:0 frame:0
> > >        TX packets:8282858 errors:0 dropped:160 overruns:0 carrier:0
> > >        collisions:0 txqueuelen:32
> > >        RX bytes:5758663699 (5.3 GiB)  TX bytes:5887860418 (5.4 GiB)
> > >
> > >
> > > From the guest
> > > # ifconfig eth0
> > > eth0      Link encap:Ethernet  HWaddr 00:16:3E:02:00:91           
> > > inet addr:10.135.2.91  Bcast:10.135.2.127  Mask:255.255.255.192
> > >        inet6 addr: fe80::216:3eff:fe02:91/64 Scope:Link
> > >        UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
> > >        RX packets:8283072 errors:0 dropped:0 overruns:0 frame:0
> > >        TX packets:7946364 errors:0 dropped:0 overruns:0 carrier:0
> > >        collisions:0 txqueuelen:1000
> > >        RX bytes:5887884877 (5.4 GiB)  TX bytes:5869927068 (5.4 GiB)
> > >
> > > # cat /etc/network/interfaces
> > >
> > > auto eth0
> > > iface eth0 inet dhcp
> > > post-up ethtool -K eth0 tx off
> > >
> > >
> > >
> > >
> > > I am a stumped on this and it is causing some problems for some of  
> > > the applications that are being tested across the virtual  
> > > machines. For example, one app on one vm needs to connect to a  
> > > database, on another vm, but it can't and complains. Do any of you  
> > > know what might be causing this or any ways to fix this? Thanks in  
> > > advance.
> > > mike
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-users
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users