Date: Tue, 6 Apr 1999 15:59:49 -0700 From: Rahul Dhesi <dhesi@rahul.net> To: freebsd-stable@freebsd.org Subject: 3.1-STABLE + specific hardware crashes, 2.2.8 ok on same hardware Message-ID: <199904062259.AA13422@waltz.rahul.net> References: <3709e0fd.7404424@news.esat.net> <370A246B.AF62056E@dyndns.org> <370a2a35.26152099@news.esat.net> <7edbf0$nnn$1@artemis.backbone.ou.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
I am encountering an odd situation. I am doing experiments with two machines and two OSs. Configuration I: ASUS P2B-S m/b, Pentium II 400 MHz, 256M ECC memory Configuration Ia: Using onboard SCSI. Configuration Ib: Onboard SCSI disabled, Symbios-based CSC wide SCSi card added. Configuration II: FIC PA-2013 m/b, AMD K6-2 380 MHz, 128M ECC memory, Symbios-based CSC wide SCSi card. Both have Netgear FA310TX (Tulip) network cards. The CSC SCSI card shows up at boot time as 'ncr 53c875 fast20 wide scsi'. The operating systems are: - FreeBSD 3.1-STABLE updated via cvsup on April 3rd. (Called 3.1 below.) I used the exact same OS disk in all experiments, i.e., moved the disk from machine to machine. - FreeBSD 2.2.8-19990120-SNAP. (Called 2.2.8 below.) The 'intensive processing' below involves doing the following concurrently: - using rsync to copy fileystem A to filesystem B, repeated in a loop - rm -rf on filesystem B, repeated in a loop Thus we have lots of file copies and directory creates going on, concurrently with rm -rf. Here is what I find. 1. Configuration Ia + 3.1: Repeatedly crashes every 1-2 hours under intensive processing. Crash dump backtrace is completely different each time. 2. Configuration Ib + 3.1: Crashes the same as above. 3. Configuration II + 3.1: No crashes under intensive processing. 4. Configuration Ib + 2.2.8: No crashes under intensive processing. Let's explore two theories. Theory A: Bad hardware in Configurations Ia and Ib. This is indicated by the fact that the crash dump is different each time. However, see item 4 above -- the identical hardware is rock solid with 2.2.8. Theory B: A bug in 3.1. However, the identical OS disk with 3.1 on it is rock solid with Configuration II, but shows repeated crashes with Configurations Ia and Ib. So, if there is an OS problem, it's dependent on the motherboard and CPU -- which seems a bit unusual to me. So, I am puzzled as to why 3.1 crashes with the PII motherboard but not with the K6, while 2.2.8 is solid with both motherboards. What sort of hardware problem could it be that would be triggered by 3.1? I didn't save any actual crash dumps, but I can probably regenerate them quite easily. I did save the backtraces, and they are all at this URL: http://www.rahul.net/dhesi/freebsd/crash3/ I would welcome any suggestions as to what other tests to do. Regrettably, I do not have any other machine available at this time which is identical to Configuration Ia. -- Rahul Dhesi <dhesi@spams.r.us.com> To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199904062259.AA13422>