From owner-freebsd-questions Sat Aug 23 15:32:36 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id PAA18185 for questions-outgoing; Sat, 23 Aug 1997 15:32:36 -0700 (PDT) Received: from srv.net (snake.srv.net [199.104.81.3]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id PAA18178 for ; Sat, 23 Aug 1997 15:32:32 -0700 (PDT) Received: from darkstar.home (dialin3.anlw.anl.gov [141.221.254.103]) by srv.net (8.8.5/8.8.5) with SMTP id QAA10087 for ; Sat, 23 Aug 1997 16:22:12 -0600 (MDT) Date: Sat, 23 Aug 1997 15:21:39 -0700 (MST) From: Charles Mott X-Sender: cmott@darkstar.home To: freebsd-questions@freebsd.org Subject: 2.2.2 lockups Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@freebsd.org X-Loop: FreeBSD.org Precedence: bulk I recently upgraded from 2.1.0 to 2.2.2. It was a clean re-install (I reformatted the disk). It is on an Intel 386 (25 or 33 MHz, I can't remember) with 8mb of RAM and a 500mb IDE hard drive. I have been using 2.1.0-R with no problems for over a year. Before I go on to the problems, I will say a few good things about 2.2.2. Using an antiquated 386, efficiency improvements in the operating system tend to be a little more apparent than for a late model pentium. The first thing I noticed is that FTP transfers of ~300 kilobyte files were almost maxing out the ethernet bandwidth at 900 to 1060 kbytes/sec, whereas I was used to seeing 180 to 300 kbytes/sec with 2.1.0-R. Wow! Samba performance also seems disctinctly more crisp and lower latency, although I can't quantify it. The bad news is that I have had several lockups or crashes in a variety of circumstances. There is no common thread really except to say that the machine is doing something when it crashes rather being in an idle state. I will give two examples. (1) A "host -l" to certain domains over pppd seemed to very reliably crash the system. This problem does not happen with user ppp, although the "host -l" process itself might die. (2) Creating a tar archive directly a floppy disk, "tar -cvf /dev/fd0 ..." also seemed to crash the system. I don't like crashing the system, so I have learned to avoid activities which do so, and now things seem to be working well. I don't expect any help with specific problems I mentioned, but what I am wondering is if the main FreeBSD systems programmers are all doing their work on reasonably up-to-date and fast machines, is it possible that they are missing some timing problems that might be picked up on older, slower computers? Charles Mott