Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Sat, 16 Dec 2000 11:40:07 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Sat, 16 Dec 2000 11:39:58 -0500 Received: from netsrvr.ami.com.au ([203.55.31.38]:13106 "EHLO netsrvr.ami.com.au") by vger.kernel.org with ESMTP id ; Sat, 16 Dec 2000 11:39:42 -0500 Message-Id: <200012160320.eBG3JQZ16646@emu.os2.ami.com.au> X-Mailer: exmh version 2.1.1 10/15/1999 To: Linux kernel Subject: System lockup in 2.4.0-test12; loop dev & ext2 suspect Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Sat, 16 Dec 2000 11:19:37 +0800 From: John Summerfield Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I'm setting up a filesystem for user-mode linux, but uml is not directly involved. I have a script (enhanced version of some supplied with uml) which is best run as root. It creats a 100 Mbyte file (dd from /dev/zero) and makes a filesystem (mke2fs). It then mounts the filesystem, and proceeds to populate it using rpm to pull packages from my CD (scsi cd writer, mounted via autofs); it puts a few essential ones first (filesystem, bash, setup and a couple of others) and then a great heap at once. The system has locked up three times at this point; I was wondering about the SCSI for a while. However, the fourth time it locked was during an "rpm --rebuild" A common feature of each time is that the filesystem on the loop device was getting a hammering. It's characteristic of this lockup that I can ping the box, it does I/O etc. I even found that it logged some information connected with my efforts with system-request. I could see nothing relevant to the problem. I tried to ssh into the box from across the LAN; the effort was logged as successful. What I cannot do is run programs. Although the log said ssh was successful, I couldn't see that from my terminal. I mostly can't run programs - I found out that the script was stuck in "rpm --rebuilddb" using the ps command; maybe it was cached? In the STAT column, it showd "D" and nothing I could do with the KILL command would get id of it. I've had XFree lock solid (I use GNOME); the last time I created the problem from a VC. Well, nearly solid - I was able to get out of it with ctrl-alt-F1 so I could do things from a VC. I have on each occasion been able to reboot using the sysrq feature. Software versions: [summer@possum summer]$ rpm -q mount e2fsprogs losetup mount-2.10f-1 e2fsprogs-1.19-0 losetup-2.9u-4 [summer@possum summer]$ I'm running Red Hat Linux 6.2. -- Cheers John Summerfield http://www2.ami.com.au/ for OS/2 & linux information. Configuration, networking, combined IBM ftpsites index. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/