Date: Fri, 15 Mar 1996 22:13:06 -0800 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: current@freebsd.org Cc: peter@freebsd.org Subject: Very strange packet-forwarding behavior! Message-ID: <1093.826956786@time.cdrom.com>
next in thread | raw e-mail | index | archive | help
I have an ISDN line hanging off of whisker.cdrom.com, my little 486 DX2 based router. It runs an older (pre-instabilities) version of -current and has been happy in its role as a router for my internal net for some time. I routinely ftp huge files across it, run sups, you name it. All of this came to a screeching halt this evening when I started using ssh with `remote cvs' to check out a tree straight from freefall's ncvs tree onto my development box (time.cdrom.com). The ISDN TA's receive light is fairly well pegged solid, which has always generally been the case while receiving large files and such - I routinely get 10.5K/sec from ftp, which is pretty much the theoretical max for an async 115.2K ISDN link. With ssh eating the line, however, everything else grinds to a screeching halt. I can't even *ping* any other hosts through whisker, nor do DNS queries or any other forms of traffic get through. It's like ssh is not only eating all the bandwidth, it's eating it at such a "high priority" that absolutely nothing else can get through, not even pings. If I ^Z the cvs op that's using ssh, everything suddently starts working again. Once I foreground it, it's wedge city again. I dunno, I've never had to suspend a process just to get name lookups to work before! :-) Any ideas? Jordan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1093.826956786>