FIXED - Re: OS hallucinates hard drive geometry
On Wednesday 19 June 2002 03:25, Michel Loos wrote: Did you format (mkfs) those 2 partitions before trying to mount them? Seems the kernel sees your old formatation on hda2 and no formatation on hda3 D'oh, THAT'S what I was missing :) mkfs'ing the partitions solved the problem. Thanks to everyone who replied. Aaron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OS hallucinates hard drive geometry
Aaron Maxwell wrote: shiznit:~# df -h FilesystemSize Used Avail Use% Mounted on /dev/hda1 11G 4.5G 6.0G 43% / /dev/hda2 926M 65M 814M 8% /mnt/hda2 Note that: 1) /dev/hda2 is smaller than it should be. 2) /dev/hda3 could not be mounted at all. I tried the partitioning with fdisk, cfdisk, and parted (I decided not to try sfdisk yet). Same results, except parted produced this warning: Information: The operating system thinks the geometry on /dev/hda is 2495/255/63. Therefore, cylinder 1024 ends at 8032.499M. I'm not clear on what to try next. Anyone? Thanks in advance. What does your BIOS think the size and geometry of your drive are? I'm thinking the problem is that Linux is getting confused by the BIOS. This happened to me when I tried to install a 60G drive into my 3 year old PC, until I flashed the BIOS to the latest version. I think there's also a way to pass the real disk geometry to the Linux kernel using a LILO (or GRUB, etc) parameter, ie. to override the BIOS values. You'd have to investigate this solution yourself though, as I've never tried it. Matthew -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OS hallucinates hard drive geometry
On Tuesday 18 June 2002 21:03, Matthew Dalton wrote: What does your BIOS think the size and geometry of your drive are? Bios believes CHS is 2495/255/63, and size is about 2e10 bytes (I didn't write down the exact number). I'm thinking the problem is that Linux is getting confused by the BIOS. I don't think so, though I did before. First, this problem didn't exist a few months ago, and I've had the same bios. Second, 2495/255/63 should be correct: 2495*255*63*(512 bytes/sector) * (1GB/2^30 bytes) = 19.1GB (the actual capacity of my HD). (it IS 512 bytes per sector, right?) Thanks anyway though - I didn't think to check this out until you suggested it, and that's valuable. Any other ideas? P.S.: I think there's also a way to pass the real disk geometry to the Linux kernel using a LILO (or GRUB, etc) parameter, ie. to override the BIOS values. You'd have to investigate this solution yourself though, as I've never tried it. You are correct on this, incidentally. The correct param is hd=C/H/S (from [kernel-source]/Documentation/kernel-parameters.txt) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OS hallucinates hard drive geometry
I just realized something interesting. Before I tried repartitioning as described in my original message (below), I had hda2 as a smaller partition containing the HURD. When I mount the new /dev/hda2, its size is 926 MB, and it contains all the old hurd files. I did not write it down and can't remember, but I think this size - just under a gig - is what I made the HURD partition long ago. So the partition table is correct; however, when the OS actually is up and running, the kernel somehow is confused, and acts as if it still has the OLD partition table. (It's not the geometry at all like I originally thought.) What could cause this behavior? Thanks, Aaron On Tuesday 18 June 2002 20:37, Aaron Maxwell wrote: Hi, I'm running woody with kernel 2.4.18. (This is a custom kernel - I downloaded the source, and built it with make-kpkg.) My hard drive is 20 GB with three primary partions: Device BootStart EndBlocks Id System /dev/hda1 * 1 1459 11719386 83 Linux /dev/hda2 1460 1945 3903795 83 Linux /dev/hda3 1946 2495 4417875 83 Linux This means that hda1 is 11.7GB, hda2 is 3.9GB, and hda3 is 4.4 GB. (hda1 has existed for a while: hda2 and hda3 are new partitions.) However, when I mount them, problems ensue: (hda1 is already mounted on /) shiznit:~# mount /dev/hda2 /mnt/hda2 shiznit:~# mount /dev/hda3 /mnt/hda3 mount: you must specify the filesystem type shiznit:~# mount -t ext2 /dev/hda3 /mnt/hda3 mount: wrong fs type, bad option, bad superblock on /dev/hda3, or too many mounted file systems shiznit:~# df -h FilesystemSize Used Avail Use% Mounted on /dev/hda1 11G 4.5G 6.0G 43% / /dev/hda2 926M 65M 814M 8% /mnt/hda2 Note that: 1) /dev/hda2 is smaller than it should be. 2) /dev/hda3 could not be mounted at all. I tried the partitioning with fdisk, cfdisk, and parted (I decided not to try sfdisk yet). Same results, except parted produced this warning: Information: The operating system thinks the geometry on /dev/hda is 2495/255/63. Therefore, cylinder 1024 ends at 8032.499M. I'm not clear on what to try next. Anyone? Thanks in advance. Aaron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OS hallucinates hard drive geometry
I just realized something interesting. Before I tried repartitioning as described in my original message (below), I had hda2 as a smaller partition containing the HURD. When I mount the new /dev/hda2, its size is 926 MB, and it contains all the old hurd files. I did not write it down and can't remember, but I think this size - just under a gig - is what I made the HURD partition long ago. So the partition table is correct; however, when the OS actually is up and running, the kernel somehow is confused, and acts as if it still has the OLD partition table. (It's not the geometry at all like I originally thought.) What could cause this behavior? Thanks, Aaron On Tuesday 18 June 2002 20:37, Aaron Maxwell wrote: Hi, I'm running woody with kernel 2.4.18. (This is a custom kernel - I downloaded the source, and built it with make-kpkg.) My hard drive is 20 GB with three primary partions: Device BootStart EndBlocks Id System /dev/hda1 * 1 1459 11719386 83 Linux /dev/hda2 1460 1945 3903795 83 Linux /dev/hda3 1946 2495 4417875 83 Linux This means that hda1 is 11.7GB, hda2 is 3.9GB, and hda3 is 4.4 GB. (hda1 has existed for a while: hda2 and hda3 are new partitions.) However, when I mount them, problems ensue: (hda1 is already mounted on /) shiznit:~# mount /dev/hda2 /mnt/hda2 shiznit:~# mount /dev/hda3 /mnt/hda3 mount: you must specify the filesystem type shiznit:~# mount -t ext2 /dev/hda3 /mnt/hda3 mount: wrong fs type, bad option, bad superblock on /dev/hda3, or too many mounted file systems shiznit:~# df -h FilesystemSize Used Avail Use% Mounted on /dev/hda1 11G 4.5G 6.0G 43% / /dev/hda2 926M 65M 814M 8% /mnt/hda2 Note that: 1) /dev/hda2 is smaller than it should be. 2) /dev/hda3 could not be mounted at all. I tried the partitioning with fdisk, cfdisk, and parted (I decided not to try sfdisk yet). Same results, except parted produced this warning: Information: The operating system thinks the geometry on /dev/hda is 2495/255/63. Therefore, cylinder 1024 ends at 8032.499M. I'm not clear on what to try next. Anyone? Thanks in advance. Aaron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OS hallucinates hard drive geometry
On Wednesday 19 June 2002 07:18, Aaron Maxwell wrote: I just realized something interesting. Before I tried repartitioning as described in my original message (below), I had hda2 as a smaller partition containing the HURD. When I mount the new /dev/hda2, its size is 926 MB, and it contains all the old hurd files. I did not write it down and can't remember, but I think this size - just under a gig - is what I made the HURD partition long ago. So the partition table is correct; however, when the OS actually is up and running, the kernel somehow is confused, and acts as if it still has the OLD partition table. (It's not the geometry at all like I originally thought.) What could cause this behavior? I had a similar problem once (some time ago, don't know the system specs anymore). After some fdisk orgies, linux seemed to lose partition numbers. So I ended up with an extended partition that containded something like partitions 5,8,11 or something. The /dev/sdXY's were numered correctly, though. The only way I was able to fix this was to start from scratch. -- Embedded Linux -- True multitasking! TWO TOASTS AT THE SAME TIME! -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OS hallucinates hard drive geometry
In article [EMAIL PROTECTED], Aaron Maxwell [EMAIL PROTECTED] wrote: I tried the partitioning with fdisk, cfdisk, and parted (I decided not to try sfdisk yet). Same results, except parted produced this warning: Information: The operating system thinks the geometry on /dev/hda is 2495/255/63. Therefore, cylinder 1024 ends at 8032.499M. I'm not clear on what to try next. Anyone? Thanks in advance. You do know that when you repartition a drive of which at least one partition is in use you must reboot ? Otherwise the kernel will not pick up the new partition table. Mike. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OS hallucinates hard drive geometry
Em Qua, 2002-06-19 às 02:18, Aaron Maxwell escreveu: I just realized something interesting. Before I tried repartitioning as described in my original message (below), I had hda2 as a smaller partition containing the HURD. When I mount the new /dev/hda2, its size is 926 MB, and it contains all the old hurd files. I did not write it down and can't remember, but I think this size - just under a gig - is what I made the HURD partition long ago. So the partition table is correct; however, when the OS actually is up and running, the kernel somehow is confused, and acts as if it still has the OLD partition table. (It's not the geometry at all like I originally thought.) What could cause this behavior? Did you format (mkfs) those 2 partitions before trying to mount them? Seems the kernel sees your old formatation on hda2 and no formatation on hda3 Michel. Thanks, Aaron On Tuesday 18 June 2002 20:37, Aaron Maxwell wrote: Hi, I'm running woody with kernel 2.4.18. (This is a custom kernel - I downloaded the source, and built it with make-kpkg.) My hard drive is 20 GB with three primary partions: Device BootStart EndBlocks Id System /dev/hda1 * 1 1459 11719386 83 Linux /dev/hda2 1460 1945 3903795 83 Linux /dev/hda3 1946 2495 4417875 83 Linux This means that hda1 is 11.7GB, hda2 is 3.9GB, and hda3 is 4.4 GB. (hda1 has existed for a while: hda2 and hda3 are new partitions.) However, when I mount them, problems ensue: (hda1 is already mounted on /) shiznit:~# mount /dev/hda2 /mnt/hda2 shiznit:~# mount /dev/hda3 /mnt/hda3 mount: you must specify the filesystem type shiznit:~# mount -t ext2 /dev/hda3 /mnt/hda3 mount: wrong fs type, bad option, bad superblock on /dev/hda3, or too many mounted file systems shiznit:~# df -h FilesystemSize Used Avail Use% Mounted on /dev/hda1 11G 4.5G 6.0G 43% / /dev/hda2 926M 65M 814M 8% /mnt/hda2 Note that: 1) /dev/hda2 is smaller than it should be. 2) /dev/hda3 could not be mounted at all. I tried the partitioning with fdisk, cfdisk, and parted (I decided not to try sfdisk yet). Same results, except parted produced this warning: Information: The operating system thinks the geometry on /dev/hda is 2495/255/63. Therefore, cylinder 1024 ends at 8032.499M. I'm not clear on what to try next. Anyone? Thanks in advance. Aaron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]