Messages in this thread | | | Date | Fri, 15 Oct 1999 13:46:04 -0700 | From | Benjamin Suto <> | Subject | TCP/IP Packet Loss (2.2.13pre17) |
| |
We've been having some problems with some machines with a Samba/Email server. It's a Dell Poweredge 4300, with 2 processors, with 1GB RAM, MegaRaid drivers, (We tried it with an etherexpress pro 100, and are currently using a tulip-based card, with the same problem).. This machine has at least 200-300 concurrent connections at any one time, and it usually works fine.
However, with some workstations connecting to the server, we've had problems pinging/accessing the server. The first few packets get dropped, then packets coming after that go through without a hitch. If we keep pinging it afterwards, all the subsequent pings go through flawlessly.
Sometimes, when a pop3 mail check goes through, it will connect, do nothing for 6-7 seconds, then finally finish. Afterwards, if we do it from the same machine, it happens instantaneously.
If we wait 10 minutes or more on the same machine, the cycle is repeated again. It's become quite a problem.
We've run tcpdump on that network segment and noticed a few things. Most of the time, the connections are initiated and the mail checks go through instantaneously. Sometimes, however it seems like the connection is initiated, there is a delay of 6-7 seconds, then the linux server makes an ARP request for the machine, the machine responds, and things progress. We do not seem to have a network problem.
Does anyone have any suggestions, or any other troubleshooting we can do to make this problem easier to solve?
Ben
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |