LAG ISSUE

[winterstide]

OK this problem happens on both mine and my friends Redhat Machines... from Redhat 6 to 7.1 and on different hardware configurations.

This problem happens with all types of services, from telnet to sendmail, and it comes as ittermittent...

The problems is when trying to use a service such as telnet or connectiong to a pop3 server on the machine, sometimes it is instant ( like telnet will connect staright away) but all too often you will have a lag delay of up to 3 mins...

As I said before, this happens all lots of version installs and hardware, so it must be some network configuration problem? We thought it was a routing problem, but that all seems ok...

Any Ideas???

Cheers

Luke

[element-x]

There was a similar post in the Web Serving/Security forum, the suggestions included were

ping the machine(s) which are slow, and see if there is significant packet loss. (or any at all, there shouldn't be)
[response timed out]

The second was that the cabling used wasn't the crossover needed (this isn't needed if using a hub) or that the cabling was bad.

and lastly, if it's neither of those, that it was "b0rk3d" hardware.

I hope this gives you a better idea. It would be a better idea to do a quick search for your problem on the message boards before posting. (it would save a lot of time)

[winterstide]

Hmmm but i dont think its that.... thats why i cant search for it... you see it cant be hardware cos it happens on a wide range of new and old hardware, including stuff that has been proven to work. The thing is it only happens SOMETIMES... sometimes it will be sweet, and work instantly... then others it will just sit there??? i thought it could be some type of power saving feature in linux? but i dont know...

[element-x]

have you taken into account the other suggestions made?

bad crossover cabling?

I can't see linux have some sort of 'power saving' on ethernet, it wouldn't make sense to me.

[winterstide]

like i said, its not hardware cos its been tested on about 10 different setups of different hardware which has been proven to work...

im talking from a crossover cable to 300 computer switched network...

[jeo]

Hi

I've got the same problem on a IBM PIII (standard HW). It's running RH7.2 and I'm experiancing the the problem when using FTP or Telnet. I'm sure its not the cabling since a connection (when obtained) runs with out problems.
It seems like when applying rules for login the problem appears. BUT it will not disapear when resetting hosts.allow, .deny etc.


Hope someone has a solution!!


/JeO

[chrism01]

could be DNS? if the address is not already cached, could take a while to find if DNs prob. Once cached, should be instant (ish).
Try turning everything off, then bring up and test initial vs subsequent connection attempts timings..

[bandwidth_pig]

Just curious, but what does the output of ifconfig show? Do you see any errors? If not, perhaps due to the large size of the network you are on (300 machines you were saying) and the fact that the problem is not regular but intermitant, could it be that you are simply out of bandwidth? What is the utilization rate on the Ethernet Switch that you are attached to? Could be perhaps that when you experience the problem just happens to be at network peak times and since the bandwidth is not available, the packets are simply getting dropped. If you ping out during one of these episodes, do you get any packet loss?

[jeo]

Hi

U are right Chris it is a DNS problem. Acctually I found another thread that suggested DNS. I deleted DNS on the server and the problem disapeared - both for telnet and Ftp.


Thanx
JeO