I was downloading a 13 MB file from an FTP site in Linux Netscape today
and
suddenly I couldn't do anything else on the Internet. I pinged various
hosts, and finally realized that although I could still make connections,
the packet turnaround time was more than 13 seconds. So to contact a
host, I would have to spend 13 seconds contacting my DNS, then get the
hostname, then spend another 13 seconds contacting the host. I thought to
redial, but as soon as I cancelled the download, everything went back to
normal. I thought the problem might be Netscape's fault, so I attempted
the same download in Lynx, and had the same result. The same thing occurs
when downloading from other hosts. This leaves me with 2 questions:
1. Why does pppd (I'm using 2.3.5) allow one socket to hog the entire ppp
route? I would rather download at half the speed and be able to make
basic telnet connections than to download at this speed and be unable to
do anything else.
2. I was only downloading at 2 kb/s on a 56k modem. 2kb/s shouldn't be
enough to kill every other socket, especially since I've done things like
this before, without any problem.
Is this a previously seen bug? Thanks very much. :)
mark
-
To unsubscribe from this list: send the line "unsubscribe linux-ppp" in
the body of a message to [EMAIL PROTECTED]
