Date: Tue, 28 Mar 1995 12:35:21 -0700 (MST) From: Brad Midgley <brad@pht.com> To: freebsd-hackers@FreeBSD.org Subject: 950322 SNAP and flakey networking Message-ID: <Pine.LNX.3.91.950328120237.10871A-100000@exodus.pht.com>
next in thread | raw e-mail | index | archive | help
950322 seems much better than plain 2.0 as far as kernel panics go. (no fatal errors yet.) However, networking works, then doesn't, then works. The symptoms are the same as you would expect if someone occasionally walked over to the machine and unplugged it from the (coax) network, waited a while, and plugged it back in: nothing can get in or out of the machine through the net for a while, netstat shows outgoing stuff queueing up, then the net operates again. (In case you're wondering, I have checked, and there's nobody back there unplugging it :) I have a monitoring script which beeps about every three seconds if it can't ping the machine next to it. It just beeped about 15 times and then stopped. It does this about every ten minutes. I have the snap system running on a Pentium-60. It has an SMC net card, PCI Buslogic SCSI, oak "otivga" pci video, 48M ram. It's primarily being used for NFS service, (also configured as a client), wu-ftp, and cern_httpd. I reconfigured the GENERIC kernel with userlimit 64 and with scsi devices sd1-sd6. It actually has 5 drives on it: 4 seagates and a quantum. Other items of interest/annoyance: "mb_map: full" messages about once an hour, "in_rtqtimo: adjusted rtq_reallyold to 2400/1600/1066" messages, syslogd dies when started in /etc/rc but can be started up manually later. about 3/4 of the times I try netstat, it dies before finishing the "Active Internet Connections" section and prints ??? Any pointers which could make let me run the system more smoothly would really be appreciated of course. brad@pht.com
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.3.91.950328120237.10871A-100000>