I have a DNS server with a lot of traffic. After just a couple of minutes of resetting the netstat -s counters, I get the following output:

        130670 datagrams received
        0 with incomplete header
        0 with bad data length field
        1 with bad checksum
        30 with no checksum
        0 dropped due to no socket
        12 broadcast/multicast datagrams dropped due to no socket
        4226 dropped due to full socket buffers
        0 not for hashed pcb
        126431 delivered
        136753 datagrams output

As time goes by, the "dropped due to full socket buffers" number goes up.

I have increased the sysctl variable net.inet.udp.recvspace to 100000, up from about 40000. This did not make any difference, the number goes up with time.

The server is running FreeBSD 4.6 RELEASE with Bind 9.2.2.

I went through the newsgroups and did not find any solutions to this problem.

Any help will be greatly appreciated.

Arie Kachler

[EMAIL PROTECTED] mailing list
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Reply via email to