OpenVZ Forum


Home » General » Support » *SOLVED* Slow Network
*SOLVED* Slow Network [message #14899] Thu, 12 July 2007 17:54 Go to next message
moorejon is currently offline  moorejon
Messages: 4
Registered: July 2007
Junior Member
I am attempting to setup a HA cluster using OpenVZ and the guide posted in the wiki. During the process we noticed extremely slow network connections even from the host OS. iperf to another server on the same 100 mbps LAN reports

[ ~]# iperf -c ##############
------------------------------------------------------------
Client connecting to ##############, TCP port 5001
TCP window size: 16.0 KByte (default)
------------------------------------------------------------
[ 3] local ########### port 39098 connected with ####### port 5001
[ 3] 0.0-10.2 sec 3.72 MBytes 3.05 Mbits/sec


Servers with the identical hardware on the same subnet running CentOS 4.x 64 bit report around 96 Mbps

Inside a VE we get the same or lower iperf results, so I am assuming this is an issue with the host OS rather than VE configuration options.

I have two test setups that provide the same results. One is CentOS 4.3 64 bit with kernel:

cat /proc/version
Linux version 2.6.18-8.1.4.el5.028stab035.1 (root@rhel5-64-build) (gcc version 4.1.1 20070105 (Red Hat 4.1.1-52)) #1 SMP Fri Jun 8 21:14:18 MSD 2007

The other is CentOS 4.5 32bit (couldn't get 4.3 32 bit to work with our kickstart network installer) kernel:

Linux version 2.6.18-8.1.4.el5.028stab035.1PAE (root@rhel5-32-build) (gcc version 4.1.1 20061011 (Red Hat 4.1.1-30)) #1 SMP Sat Jun 9 02:27:12 MSD 2007


Server hardware is Supermicro 1U, 2 x dual core Woodcrest 5110, 2 GB RAM, 2 x SATA 250 GB SATA drives, 2 x Intel(R) PRO/1000 Network

Any ideas what would cause this issue?

[Updated on: Fri, 13 July 2007 06:32] by Moderator

Report message to a moderator

Re: Slow Network [message #14900 is a reply to message #14899] Thu, 12 July 2007 17:57 Go to previous messageGo to next message
rickb is currently offline  rickb
Messages: 368
Registered: October 2006
Senior Member
Hi, if you reboot the node into the standard Centos kernel and run the exact same test, are the results vastly different?

There have been many reports similar to yours (some reported much higher speeds!) but whenever I asked them to perform this test, the problem was shown not to be the kernel, as the speed was approximately the same using the stock (non vz) kernel.

Rick


-------------
Common Terms I post with: http://wiki.openvz.org/Category:Definitions

UBC. Learn it, love it, live it: http://wiki.openvz.org/Proc/user_beancounters
Re: Slow Network [message #14902 is a reply to message #14900] Thu, 12 July 2007 19:09 Go to previous message
moorejon is currently offline  moorejon
Messages: 4
Registered: July 2007
Junior Member
I bow to your wisdom. Rebooted into stock smp kernel and get the same results. That will teach me to check the obvious first. I'll try a reload of the base OS no changes just to confirm these results....

Definitely something with the hardware. Appears we have about 4-5 servers slipped into the rack with this issue that only showed after trying to do something really network intensive like drbd.

Previous Topic: urgent *all VEs not accessible
Next Topic: Time issues, logs in the future
Goto Forum:
  


Current Time: Mon Aug 26 05:12:58 GMT 2024

Total time taken to generate the page: 0.03452 seconds