Operating System - HP-UX
1748022 Members
5116 Online
108757 Solutions
New Discussion

HPivm guest intermitent packet loss

 
likid0
Honored Contributor

HPivm guest intermitent packet loss

Hi,

 

We ha a HPIVM 4.2.5 host in a rx7640:

 

> model

ia64 hp server rx7640

 

> swlist -l bundle | grep VM

B.11.31.1211.01 HPVM Host AVIO Storage Software HostAvioLan B.11.31.1211.01 HPVM Host AVIO LAN Software

T2767CC B.04.20.05 Integrity VM

T8718AC B.04.20.05 Integrity VM Online Migration Software VMGuestLib B.04.20.05 Integrity VM Guest Support Libraries VMGuestSW B.04.20.05 Integrity VM Guest Support Software VMKernelSW B.04.20 Integrity VM Kernel Software

 

> swlist -l bundle | grep OE

HPUX11i-DC-OE B.11.31.1009 HP-UX Data Center Operating Environment

 

> swlist -l bundle | grep Patches

FEATURE11i B.11.31.1009.353a Feature Enablement Patches for HP-UX 11i v3, September 2010

HWEnable11i B.11.31.1009.353a Hardware Enablement Patches for HP-UX 11i v3, September 2010

QPKAPPS B.11.31.1009.353a Applications Patches for HP-UX 11i v3, September 2010

 

CM>sysrev

Core IO MP-0 : B.004.002.003

ED-0 : 002.011.000

MP-1 : B.004.002.003

ED-1 : 002.011.000

Cell 0 PDHC : B.023.003.040 - Active

PDHC : B.023.003.030

IPF_FW : 005.024.000

IPF_FW : 009.066.000 - Active

 

> swlist -l bundle | grep GigEthe

GigEther-00 B.11.31.0903 PCI GigEther;Supptd HW=A4926A/A4929A/A6096A;SW=J1642AA

GigEther-01 B.11.31.1403 PCI GigEther;Supptd HW=A6825A/A6794A/A6847A/A8685A/A9782A/A9784A/A7109A/AB465A

 

The host has 4 VMs, 3 VMs have no network problems and work fine, a fourth VM has almost constant bad ping times over 800ms on the local(and also the default gw) network, and some times it looses all network conectivity during short spells. This happens in all the NICs it has attached that are part of Vswitches in different Lans.

 

Cheking the VMs resources, cpu/mem/disk, they are almos IDLE all the time on the guest VM, so it doesn't seem it's running out of resources

 

The network stats for the VM guest on the vswiches from the host, show no problems.

 

 

Info from the guest:

 

# swlist -l bundle | grep OE

HPUX11i-HA-OE B.11.31.1009 HP-UX High Availability Operating Environment

# swlist -l bundle | grep -i patch

BUNDLE B.2012.01.31 Patch Bundle

FEATURE11i B.11.31.1009.353a Feature Enablement Patches for HP-UX 11i v3, September 2010

HWEnable11i B.11.31.1009.353a Hardware Enablement Patches for HP-UX 11i v3, September 2010

QPKAPPS B.11.31.1009.353a Applications Patches for HP-UX 11i v3, September 2010

 

# swlist -l bundle | grep AVI

GuestAVIOStor B.11.31.1211.01 HPVM Guest AVIO Storage Software GuestAvioLan B.11.31.1211.01 HPVM Guest AVIO LAN Software

 

it's giving on the guest:

Dead gateway detection can't ping the last remaining default gateway at 0xa714001 .See ndd -h ip_ire_gw_probe for more info Dead gateway detection can't ping the last remaining default gateway at 0xa714001 .See ndd -h ip_ire_gw_probe for more info

 

I have disabled the gw probing, but problems still continue, i have a ping running and you can check the high rtt times, on all nics on the guest:

03/08/14 21:11 ping  14688 ms  local ping-gw.txt

03/08/14 21:11 ping 13678 ms  ping-local.txt

03/08/14 21:11 ping  12658 ms ping-ext.txt

 

This doesn't happen in the other VMs, the other machines have the same OE,AVIO and patches,etc.

 

Any idea where I can keep looking?.

 

Thanks

 

Windows?, no thanks
1 REPLY 1
likid0
Honored Contributor

Re: HPivm guest intermitent packet loss

After watching packet loss in other interfaces of the virtual machine, I took a longer look at the VM OS resources, and this time I spotted very short burst of cpu and disk activiy, that leave the VM during and interval of one or 2 seconds unresponsive.

 

Here is the sar -d output with service times over the roof:

 

10:35:53 disk6 100.00 3.00 1 2 5794.79 22189.95

10:35:55 disk6 100.00 1.50 1 2 11593.92 2645.25

10:35:56 disk5 100.00 7.40 1 12 3624.67 12190.01

                   disk6 91.16 0.40 2 98 23967.29 4553.04

 

 And the cpu is all in sys:

 

10:35:53 0 100 0 0

10:35:55 0 100 0 0

 

This coincides in time with the high pings:

 

06/08/14 10:35 ping  18090 ms  ping-local.txt

06/08/14 10:35 ping 14281 ms ping-gw.txt

 

If I do a read/write test to the disks using dd, there is no problem, the service times are good, so there has to be a short lived process or somthing that is hogging the system, but because it only lasts seconds I can't get hold of it.

 

Any ideas?.

 

thnx

Windows?, no thanks