Date: Mon, 18 Aug 1997 10:46:23 -0400 From: Mark Mayo <mark@quickweb.com> To: Simon Bennet <support@natsoft.com.au> Cc: stable@FreeBSD.ORG Subject: Re: FreeBSD Partition Over 2048 Mega Bytes Message-ID: <19970818104623.39790@vinyl.quickweb.com> In-Reply-To: <33F86CE6.6334@natsoft.com.au>; from Simon Bennet on Mon, Aug 18, 1997 at 04:40:22PM %2B0100 References: <33F86CE6.6334@natsoft.com.au>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Aug 18, 1997 at 04:40:22PM +0100, Simon Bennet wrote: > [SNIP] > I am attempting to install FreeBSD 2.2.2-RELEASE on a Pentium 150MHz > with a 3Gig IDE Hard Drive. > > When you run the bad block scan (bad144), at install time, it fails at > the 2048 MByte point with every block past there being reported as > an error. This happens whether the drive is in LBA or NORMAL mode. > > Please note that MSDOS cannot have a partition greater than 2048MByte, > this is only possible with the lasest version of windows. > > Will FreeBSD support a single partition over 2048MBytes in length, > or is this just a problem with the bad block scanning program? Hmmm. Well, my SCSI disks are definately bigger than 2048MB: Filesystem 1K-blocks Used Avail Capacity Mounted on /dev/sd0a 31775 13558 15675 46% / /dev/sd0s1f 1913091 1078904 681140 61% /usr /dev/sd0s1e 29727 12884 14465 47% /var procfs 4 4 0 100% /proc /dev/sd1c 4108717 687694 3092326 18% /d2 So FreeBSD can definately support >2048MB partitions! I suspect your problem comes from the fact that it's an IDE drive - although I'm pretty sure the IDE code does allow for gigantic drives, I think I remember Mike Smith making some changes to support even the 8GB IDE monsters from Maxtor... It's probably just a block scanning problem. Have you tried newfs'ing it and mounting it up? As for the stuff below, I have no idea... If I saw a drive doing this, I would probably assume a hardware problem, but I'd probably be wrong :-) Good Luck! -Mark > > Also if you scan an IDE hard drive over 512MByte without LBA mode > (not all computers have LBA mode), then the bad block table gets written > above the 512MByte boundry (1024cylinder) making the drive unbootable. > Is there a way to get the bad track table to reside below cylinder 1024, > or is it safe not to use bad blocking on large IDE drives? > > I have been running a test on a 2.5GB IDE Drive in a Intel 486/66 with > FreeBSD 2.2.2-RELEASE to see if a partition over 2048MBytes works ok. > The kernel used was as loaded onto the system from the CD-ROM. > I was creating a file large enough to fill the 2.5GB partition, using > the c write() function wtiting 1024 bytes of data per write call, when > the IDE drive started to spin down and then up again, the drive > continued to spin down and up as though it was being turned off and > back on again. There were error messages displayed on the console, but > unfortunately I did not write them down. I reset the computer system > and then the hard drive failed to boot with DRIVE NOT READY, > all partition information had been lost and I had to reinstall FreeBSD, > therefore I could not consult the log file for error messages. > Is there a way FreeBSD can cause a hard drive to spin down and backup > again continually or have I encountered a hardware problem? I did > rebuild a kernel with FAILSAFE set and ran the above program again, > and everything worked ok this time, indicating that FreeBSD can handle > the large partition. > > Thanking you in advance. > > > Regards Craig. > National Software Pty Ltd > ------------------------- -- ---------------------------------------------------------------------------- Mark Mayo mark@quickweb.com RingZero Comp. http://vinyl.quickweb.com/mark finger mark@quickweb.com for my PGP key and GCS code ---------------------------------------------------------------------------- University degrees are a bit like adultery: you may not want to get involved with that sort of thing, but you don't want to be thought incapable. -Sir Peter Imbert
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19970818104623.39790>