topic Re: Network performance in Operating System - HP-UX

Network performance

Shivkumar — Sun, 11 Sep 2005 00:27:38 GMT

What are the commands to find out the network performance and how to determine that there exist bottleneck due to network ?

Thanks,
Shiv

Re: Network performance

Thayanidhi — Sun, 11 Sep 2005 01:17:21 GMT

Hi,
lanadmin gives statistics of the interface.
glance could be better option if you want analyse the performance.
There are also contributed tcpdump and/or similar utilities if you want capture the packets for analysis.

Regds
TT

Re: Network performance

saju_2 — Sun, 11 Sep 2005 01:31:00 GMT

Hi shiv

I believe with netstat command can help u in pointing out the exect problem in network.
There are lot of switches for netstat

netstat -v
netstat -r --routing table
netstat -i --interface status

more info at man netstat

Also lanadmin is a good administration tool for network monitoing and trouble shoot

It is a menu based tool which has variety of options to look into

I am pasting a section from "HP-UX performance cook-book"

Here are some things to try if you suspect a network bottleneck: Run the command
netstat â s twice, spaced 30 seconds apart. Look at the change in tcp sent data packet
retransmissions, and any udp socket overflows or ip fragments dropped after timeout
(reassembly timeouts). You can grab a copy of the utility beforeafter from the HP
Networking contrib archive (see References section) under the tools directory: it will
help parse the output.
For lanadmin, you can watch for inbound and outbound discard and error counts and
excessive collisions. Be careful, because on fast networks like gigabit ethernet, the 32-bit
counters shown by these tools roll over frequently.
If you use NFS a lot, see the configuration tips above in the System Setup section. The
nfsstat command and Glance's NFS Reports can be helpful in monitoring traffic,
especially on the server. If the NFS By System report on the server shows one client
causing lots of activity, run Glance on that client and see which processes may be causing
it. We've seen users on clients doing repeated (and unnecessary) find commands across
NFS mounts, which can drag a server down. On the client side, use nfsstat to watch for
retransmits and timeouts that can indicate a network or server problem. See our NFS
performance reference (References section below).

Full book available at h21007.www2.hp.com/dspp/files/unprotected/devresource/Docs/TechPapers/UXPerfCookBook.pdf

Regards
CS

Re: Network performance

Devesh Pant_1 — Sun, 11 Sep 2005 16:39:18 GMT

Shiv,

lanadmin is a good real time monitoring tool as is glanceplus.

To have historical data to analyse, you might want to consider measureware which of course comes with a price tag attached to it.

thanks
DP

Re: Network performance

morganelan — Sun, 11 Sep 2005 22:37:05 GMT

Important early step in troubleshooting any Network performance issues:
#How many network hops (i.e. bridges, hubs, routers, switches, etc.) do network packets traverse between the client and the server systems?
#What is the speed of each link separating these systems?
#Does your network equipment use auto-negotiation to set speed and duplex settings?
#Are your network interfaces configured for half-duplex or full-duplex mode?
#Do your switch port settings match the speed and duplex settings of your host interfaces?
#What is the maximum transmission unit (MTU) size of the links between these systems?
#If the links are using different MTU sizes, how are the packets being translated? For example, if the NFS client resides in an FDDI ring and uses an MTU size of 4352 and the NFS server uses a 100BT interface with an MTU size of 1500, how are the 4352 byte packets from the client being fragmented into 1500 byte packets for the server?
#Do packets sent from the client to the server take the same route through the network as the packets sent from the server to the client?

You can man netstat to find what you want to know about your network statistics.The traceroute(1M) tool provides a simple means of determining the path through the network taken by packets sent from one system to another.Another tool shipping with HP-UX that can simplify the process of collecting network topology information is ping(1M).The lanadmin command allows a system administrator to display many useful statistics kept by the LAN driver subsystem, regardless of the interface type.

You can try using tools such as:
ttcp:http://ftp.arl.mil/ftp/pub/ttcp
netperf:http://netperf.org

Re: Network performance

Steven E. Protter — Sun, 11 Sep 2005 23:39:19 GMT

This script set, provided elsewhere I believe provides plenty of data on network performance.

Attached.

SEP

Re: Network performance

Yogeeraj_1 — Mon, 12 Sep 2005 02:43:07 GMT

hi shiv,

On you hp-ux server,
if you run glance and go to "Network by Interface" by typing "l" you can get much more details.

for example, if you type "S" and select your main interface and get information such as:
Collision Rate, Error Rate, Packets In, Packets Out, Packet Rate In and Packet rate Out.

Then, you further query at your switch level. Depending on the switch manageability, you can quesry the input and output rates...

Further down, you can go into you router and verify the inbound and out bound traffics.

We are using a great software from Solarwinds.

hope this helps too!

kind regards
yogeeraj

Re: Network performance

Rajesh SB — Mon, 12 Sep 2005 02:51:42 GMT

Hi,

When you are interested watch about network performance for bottleneck use

# glance -l -f
-l option captures the utilization of Network By Interface.
-f option writes the output to specified file.

Later you can analyze output for instances and can be traced easily.

Regards,
Rajesh