mounted device appears slow and also scp

[MohitAnchlia]

We have Sun OS that has a mounted file system from HP box. All other file system are pretty fast. Except when I try to compile the code on the mounted FS from HP box it is so slow that simple program compilation takes 20 mts. How can I track where the problem is.

Also when I try to scp the file to this HP box in LAN it's speed is in 200 kbps which used to be in mbps.

Only recent change was that IP address of this machine got changed.

[A. Clay Stephenson]

Almost certainly this is a speed-duplex mismatch between the NIC and the corresponding ethernet switch port. If, for example, one end of the connection is set to auto-negotiate, the other end MUST also be set to auto-negotiate. Similarly, if obe end is hard-set to 100FD; the other end MUST be hard-set to 100FD. A speed/duplex mismatch will almost work well and would not be apparent in a telnet session, for example, but an NFS connection, an FTP session, or an scp session would crawl.

[MohitAnchlia]

But how do I trace it. Is there any way I could tell ? I am not sys admin and I am not sure if my sysadmin will understand.

[rariasn]

Hi,

/usr/sbin/lanscan

/usr/sbin/ifconfig interface_name

/usr/sbin/lanadmin -s -x PPA

sample:

/usr/sbin/lanadmin -s -x 0


$/usr/sbin/lanadmin -s -x 0
Speed = 100000000
Current Config = 100 Full-Duplex MANUAL

rgs

[MohitAnchlia]

This is what I get. It looks like lan1 is down but how do I know that mounted file system on other box, which is Sun, is going over or trying to access this lan1 ? Below is the output:

:root#lanscan
Hardware Station Crd Hdw Net-Interface NM MAC HP-DLPI DLPI
Path Address In# State NamePPA ID Type Support Mjr#
0/0/0/0 0x00504F1AB70F 0 UP lan0 snap0 1 ETHER Yes 119
0/5/0/0 0x002245BHG49F 1 UP lan1 snap1 2 ETHER Yes 119

:root#ifconfig lan0
lan0: flags=843

:root#ifconfig lan1
ifconfig: no such interface

:root#lanadmin -s -x 0
Speed = 100000000
Current Config = 100 Full-Duplex MANUAL

:root#lanadmin -s -x 1
Speed = 0
The link is down. The speed and other link state information
are undefined. The configuration specified for this card is:
Speed = 1000 or 100 or 10 Mbps.
Autonegotiation = On.

[MohitAnchlia]

In the above output that I just pasted I forgot to mention that when I issue mount it's using the ip address of lan0 so I guess it's not even using lan1. Now how do I track it further because output of lan0 looks pretty good.

[MohitAnchlia]

In the above output, in above post that I just pasted I forgot to mention that when I issue mount it's using the ip address of lan0 so I guess it's not even using lan1. Now how do I track it further because output of lan0 looks pretty good.

[Steven E. Protter]

Shalom,

There could be a problem with one of the disks on the HP box.

Take a minute and run cstm,mstm or xstm on each of the disks and look for problems.

dmesg

Take a look for lbolt messages or timeouts in /var/adm/syslog/syslog.log

I favor a network problem for this, but its either the network, a switch problem or disk problem. At least that is the first three places I'd look.

SEP

[MohitAnchlia]

I don't see any error messages. Would you be able to tell me how to determine if it's the switch that's acting weird.

[VK2COT · Accepted solution]

Hello,

a) What is the result of command:

netstat -i

on HP-UX and SunOS servers?

Note, I said SunOS because that is
the correct operating system name.
Solaris is the marketing name that
includes SunOS and other things.

You should not see any or very small
number of network errors. Otherwise,
you have a problem (pyiscal cable, faulty
ports, speed and duplex mismatch, and
so on).

b) Why do not you use wonderful tools
like nttcp or ttcp (they exist for
both operating systems) and check raw
speed between Ethernet cards on two servers.

It is an excellent tool.

c) Check your routing table. Are you
using proper interfaces between the
two servers.

d) NFS statistics:

nfsstat -s (on the server side, HP-UX)
nfsstat -c (on the client side, SunOS)

Cheers,

VK2COT

[MohitAnchlia]

netstat -i on HP box shows high number in Oerrs column. What's Oerrs ? How can I see why these errors are occuring.

[D Block 2]

and your Solaris box ?

http://www.princeton.edu/~unix/Solaris/troubleshoot/netstat.html

[MohitAnchlia]

Solaris box looks good. Only few around 800 ierrs as compared to 20K oerrs on HP box

[D Block 2]

you might want to tweat or play with the numer of io readers for NFS on the hp-ux side. study the nfs stats, the cache of directories/files. if you have a window where you can increase the number of these, this might help. but I don't know.

also, study the NFS version 3, I assume.

I would seek help with PATCHes. For there might be some concerns on patching that might really help. Please investigate this after you have studied the statistics on nfstat.

[MohitAnchlia]

I am little confused about netstat -i output. I get different results from different login. From root I see different output, but when I login why my user id I see different output.