Date: Thu, 15 Feb 2001 16:49:51 -0800 From: Mark <mark@etinc.com> To: Mike Smith <msmith@FreeBSD.ORG> Cc: freebsd-questions@FreeBSD.ORG, dennis@etinc.com Subject: Re: FreeBSD cant boot from ZIP - Is it true? Message-ID: <3.0.32.20010215164950.0080b360@129.45.17.190>
next in thread | raw e-mail | index | archive | help
>> >> Whats the "trick"? > >There shouldn't be any trick; as long as you have the 'atapifd' driver in >your kernel, and the correct entries in /etc/fstab on the disk, you >should be fine. > >Where does the confusion arise? When you're trying to find the kernel, >or when you're mounting root? Are you installing to the disk, or >building a root filesystem manually? > >If you're having trouble mounting root, check what the kernel says it's >trying to mount and if it's not /dev/afd0-something then you should >suspect your /etc/fstab's / entry. > The confusion arises between the boot loader and the kernel. The ATAPI Zip drives emulate an IDE hard drive, so for example one installed as the primary IDE master would show up as "ad0c" to the loader. If you then select 0:ad(0,c) kernel It'll boot fine until it tries to remount / as r-w. The devices have been probed and now / is being recognized as afd0c, not ad0c. I've tried both entries in /etc/fstab, neither work. This wouldn't seem to be a fatal problem, since FreeBSD will prompt you for a manual location of /. But overriding the fstab doesn't work - here's an old message containing a capture of this behavior, using a Zip configured as /dev/afd0a. -------------------- <<CAPTURE> <fontfamily><param>Times New Roman</param>afd0: 239MB <<IOMEGA ZIP 250 ATAPI> [239/64/32] at ata0-master using PIO3 Mounting root from ufs:ad0s1a Root mount failed: 6 Mounting root from ufs:ad0sa Root mount failed: 6 Manual root filesystem specification: <<fstype>:<<device> Mount <<device> using filesystem <<fstype> eg. ufs:/dev/da0s1a ? List valid disk boot devices <<empty line> Abort manual input mountroot> /dev/afd0a Mounting root from /dev/afd0a Root mount failed: 22 mountroot> ufs:/dev/afd0a Mounting root from ufs:/dev/afd0a spec_getpages:(#afd/0) IO read failure: (error=0) bp 0xc1c78438 vp 0xc5ae1d40 size: 53248, resid: 32768, a_count: 53248, valid: 0x0 nread: 20480, reqpage: 7, pindex: 51, pcount: 13 vm_fault: pager read error, pid 1 (init) Nov 28 10:31:15 init: setlogin() failed: Bad address spec_getpages:(#afd/0) IO read failure: (error=0) bp 0xc1c78438 vp 0xc5ae1d40 size: 57344, resid: 32768, a_count: 57344, valid: 0x0 nread: 24576, reqpage: 7, pindex: 73, pcount: 14 vm_fault: pager read error, pid 6 (sh) pid 6 (sh), uid 0: exited on signal 11 Nov 28 10:31:16 init: /bin/sh on /etc/rc terminated abnormally, going to single user mode Enter full pathname of shell or RETURN for /sbin/sh: </fontfamily> <</CAPTURE> As you can see, manually selecting "/dev/afd0a" seems to be a valid option. However, once / is mounted, I get a series of read errors. There's more, though. If I hit "enter" a few times, it manages to execute /bin/sh and give me a prompt. At this point, most any command I type that isn't a shell builtin results in a few inital read failures followed by any number of successful operations. For example, if I tried to 'newfs' a partition, the first two attempts fail, and then I can 'newfs' as many partitions as I need. Seems like a caching issue, as if the latency of the drive exceeds the driver's expectation. This should not be an issue if it's using the "afd" driver. I'm wondering if this kind of operation is even supported? There don't seem to be any problems with the drive or the media, I can read and write files w/o errors with the drive mounted on a running system. Thanks, Mark Staudinger mark@etinc.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3.0.32.20010215164950.0080b360>