From owner-freebsd-bugs Tue Jul 17 16: 0:10 2001 Delivered-To: freebsd-bugs@hub.freebsd.org Received: from freefall.freebsd.org (freefall.freebsd.org [216.136.204.21]) by hub.freebsd.org (Postfix) with ESMTP id 8B16937B401 for ; Tue, 17 Jul 2001 16:00:07 -0700 (PDT) (envelope-from gnats@FreeBSD.org) Received: (from gnats@localhost) by freefall.freebsd.org (8.11.4/8.11.4) id f6HN07O04851; Tue, 17 Jul 2001 16:00:07 -0700 (PDT) (envelope-from gnats) Date: Tue, 17 Jul 2001 16:00:07 -0700 (PDT) Message-Id: <200107172300.f6HN07O04851@freefall.freebsd.org> To: freebsd-bugs@FreeBSD.org Cc: From: Bill Moran Subject: Re: i386/29045: Heavy disk usage causes panic in ffs_blkfree Reply-To: Bill Moran Sender: owner-freebsd-bugs@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org The following reply was made to PR i386/29045; it has been noted by GNATS. From: Bill Moran To: Ian Dowse Cc: freebsd-gnats-submit@FreeBSD.org Subject: Re: i386/29045: Heavy disk usage causes panic in ffs_blkfree Date: Tue, 17 Jul 2001 18:51:41 -0400 Ian Dowse wrote: > Could you try printing > out some more information including the contents of the free block > bitmap from frame 2 on the stack - i.e. something like Errr ... I tried, but frame 2 considers those symbols undefined. Did I misunderstand? > A good test for hardware/driver faults is to take some large > directory tree that does not change, especially one with lots of > huge files, and run > > find /whatever -type f -print0 |xargs -0 md5 > /tmp/md5.1 > find /whatever -type f -print0 |xargs -0 md5 > /tmp/md5.2 > find /whatever -type f -print0 |xargs -0 md5 > /tmp/md5.3 > > etc while the system is under heavy load. Then diff the /tmp/md5.X > files to see if anything has changed. You should try this with > trees on different disks in case there is a driver/disk dependent > corruption problem. Also try leaving quite a long gap between > running the finds; data could be getting corrupted as it sits in > the buffer cache. Here's what I did: Started a "make buildworld" and then ran the md5 routines you displayed above on the /data partition. The machine only has 1 physical disk drive. The /data partition had 15000 files on it comprising 9.3G at this point. systat showed disk usage ~99% during this. I ran two tests during the "make buildworld" (one right after the other) I ran a diff on the two resultant files and Lo and Behold! there are a slew of differences in the hashes. Now, there's no doubt in my mind that this is A Bad Thing (tm) but the big question is, does it indicate filesystem problems or ata problems? The mobo is an Asus A7 with 2 ata66 controllers and 2 ata100 controllers. The drive is an IBM 76G and is currently connected to the ata100 controller. Bill To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message