low response after replacing IP and netmask!

[leereg_6]

Hi, Team,
I changed the IP and netmask and gateway of an host to reconnect it to a small network within my LAN. Like change it from 135.252.x.x to 172.12.x.x but after that, it was too slowly to telnet to another 172 host, and it took me more than 10 minutes to FTP a 10k file.
What's wrong with it? How to resolve?

Thanks a lot!

[Karthik S S · Accepted solution]

Hi,

Check hosts entry in the nsswitch.conf file. If the earlier configuration was using DNS change it to files and then try. Did you update your /etc/hosts file with the new IP?

Karthik

[T G Manikandan]

What about the dns?
Did you change the /etc/resolv.conf file
after the change in ip.
Is it pointing to the same dns after the ip change.

Probably you can do one thing
1.rename your /etc/resolv.conf file to resolv.conf.old
2.Modify your /etc/nsswitch.conf file like

hosts : files

just to look into files.

Now try ftpying as before and check the speed.

Also I would go about checking your network collisions using lanadmin

#lanadmin
lan
dis

check for the errors and collisions.

[RAC_1]

check network spped settings on both the ends -switch and nic

[harry d brown jr]

And if dns isn't the issue look at your network speeds. Are you running at full duplex on a half duplex router or the other way around? How about your default gateway? Is it right?

Can you post the output of lanadmin - select lan - select display ??


live free or die
harry

[Adam J Markiewicz]

Mayby I"m wrong, but I think that DNS is used only once per session.
Another thing is that changing only IP should not have influence on router configuration if there was no physical reconnection (or my knowledge is again not deep enough).
So I would bet routing problem, so the first thing to check are gateways in routing tables.

Good luck

Adam.

[harry d brown jr]

Adam,

You are right about DNS - it's used to resolve name to IP then IP is used from there.

He changed the IP and the NETWORK the server USED TO BE ON, therefore the router/switch IS an issue.

Plus the NUMBER of HOPS to the main gateway!

leereg, Try a traceroute to a server and see if the hops are an issue.


live free or die
harry

[Adam J Markiewicz]

Yes, Harry, it is written...

I must had let my brains somewhere, while i read this... :)

But if another 172. host were not 172.12. default gateway could be used. And most probabilly the default gateway would be 135.252.x.x, so every datagram could fly into the Universe and then return back.

Or not. :)

traceroute would be the judge.


Good luck

Adam

die as a free man. ;D

[John Bolene]

Gotta be DNS on the telnet, probably reverse lookup.

FTP, sounds like router or network collisions.

[rick jones]

When you changed the IP, netmask and default route did you also move the cable(s) from one switch port to another? If so, then there may indeed be a duplex mismatch - lanadmin stats will show the duplex setting of the NIC, and you should look for FCS errors or Late Collisions. (The other types of collisions are normal and expected for a half-duplex link).

Since you changed subnets, you are now likely reaching the 172 hosts through your router - perhaps it is not able to keep-up with the traffic. I would check the netstat -p tcp statistics on the machine sending the data and look for retransmissions. If there are retransmissions and the lanadmin statistics are clean, then you need to check the stats on the router and the receiver.

You should take a snapshot of netstat -p tcp before the transfer and one after, and run them through "beforeafter" from ftp://ftp.cup.hp.com/dist/networking/tools/

and it will subtract the one from the other for you.

I concurr with the other assesments that DNS problems would only manifest themselves as slow session (telnet or FTP) establishment, they would not be causitive of a 10 minute transfer of a 10K file.