Hello,

the problem is as follows:

I have two boxes within local network -
'A' - CentOS
'B' - Debian

they are part of the infrastructure, and all the network settings are made for them to see each other (or so I think).

The problem is:

STEP 1 : I go to 'A' and 'ping B' - no result, all packets are lost.

STEP 2 : I go to 'B' and do 'ping A' - and everything fine, the connection works.

STEP 3 : I go back to 'A' and do again 'ping B' - and this time - everything is ok and the connection works from this direction also.


so, connection 'A->B' works only after successful connection 'B->A', and after some time of inactivity is lost.

May be anyone will be able to explain this... as I've ran out of ideas..

thanks.

Very strange. If you have the time, I would get a third computer to test. Put CentOS on it and see if the same thing happens. If it does, put Debian on it and try it again. If you still have the problem, try it with Debian to Debian, and CentOS to CentOS to try and narrow down the operating system with the problem.

I actually have many boxes on the same network - all CentOS, and no such problems. But it was not I, who configured them, so, I thought may be I missed some configuration... though I searched everywhere already ... - this is networking - not may places to search .. but may be I managed to overlook something...

So pinging from other CentOS computers to the Debian computer works? Is it just the one CentOS computer with the problem?

well - it is not that simple .. yes, from other CentOS boxes - the ping works... but - I have several IP ranges used, each of them has its own set of network/gateway/broadcast IP, the CentOS box in question ('A') is in separate rage than Debian and other CentOS boxes, but it has no problems connecting to the CentOs boxes which are in the same range as Debian - it has routes set to be able to see the target range... but for other CentOS it works - for Debian- as I've said above. And, everything is ok with connections CentOS<->Debian within the same IP range.

pgrd, what are the IP addresses and netmasks of the two computers? And what are their routing tables set to?

I'm guessing that B's IP address may be on A's local network range, but not the other way around.

Alternatively, it could be a routing problem with IP update information not being refreshed somehow in the Ethernet layer? Are you able to try switching out hubs or switches at all?

everything is fine with ip configuration and routes, as I'be already mentioned - there are other CentOS boxed in the same range with Debian, and there are no problems connecting to them. I've checked routes and IPs many times already...

About refreshing IPs ... may be ... I'm not able to switch hubs or change any hardware ... but I will look into this .. only - not sure how I can affect this ..