Stephan Shuichi Haupt <stephan@bios.t.u-tokyo.ac.jp>
Hi
I have noticed that there is not much information about
magneto-optical disks in the howto, which may be due to the fact that
these are not very popular in general. In Japan, MO drives are very
common, especially the 3.5' variety with media in 128MB (maybe not
available anymore), 230MB, and recently 640MB sizes. I suppose there
is plenty of info on usage of these drives with Linux in Japanese -
but that does not help most people for some reason ;-) MODs can be
used very much like any removable media and are handy for smaller
backups as the media are relatively inexpensive (about 10US$ / 640MB
as of 10-98). I can only comment on the usage of 230MB drives with
SCSI interface.
Drives used: several, no problems encountered (Olympus, Epson, currently
Mitsubishi MK230LK3). Drives may have strange jumper setting like "Mac
Mode" or such - naturally, disable.
If you decide to get a drive, pay attention to the
cache size - It can speed things up enormously, still speed will be
soso compared to hard disks, of course.
SCSI controllers: NCR53C810-based (Asus PCI-200), Adaptec APA-1460A,
Adaptec AHA2940.
Just install the drive as you would do with an additional SCSI hard
disk. It will show up as such. You don't need a disk in the drive when
booting.
There are two ways to format the disks:
a) A bit like a floppy. Just run mkfs on the raw device i.e. something like
sdb or sdc. I don't recommend this in general (see below).
b) Like a hard disk. Do fdisk on the raw device and then mkfs on the
partition as you would for a hard disk (like sdc0, I have never made
multiple partitions on a MOD).
What I have not tried is to boot from MOD, yet I cannot see why it
should not work. I would only recommend it for emergency system
recovery, however, due to MO drive performance.
Note: Purchased disks for Doze or Windog may be formatted "like
floppies" and cannot be used with either O(gre)S right away while MODs
formatted under linux as hard disks (partition FAT16 / type 6 and
mkdosfs) will work fine (only tested with NT 3.5/4.0). Fdisk will
issue a warning upon exit that concerned FAT16 partitions and you do
better to take it seriously (look at the fdisk man-page). The sector
size will not be automatically set properly for mkdosfs. Use "mkdosfs
-s 8". That came from some Japanese Web site in mid 1995 (Thanks to Ken
Kawabata for finding and deciphering it). Using the vfat file-system
with the disks works fine. I have only used FAT/DOSfs or Linux/ext2
formatted disks so far.
Additional Note: The media are probably a bit sensitive. Of course to
magnetic fields, but also to mechanical stress, some formats seem
to be more fragile than others (Mac format seemingly worst, data loss has
occurred when dropping disks during sneaker net traffic).
Though this does not steer anyone through particularly dense
jungle, it may be nice for completeness.
Steve
--
***********************cut*here*or*do*not********************************
S. Shuichi Haupt
email stephan@bios.t.u-tokyo.ac.jp
http://www.bios.t.u-tokyo.ac.jp/~stephan/
---------------- December 11 1998 update from Steve -------------------
OK, some problems will arise with MO disks occasionally. the safest
way to avoid them is not to use the disks "off the shelf". trying to
mount disks can even result in kernel panics. i accidentally tried to
mount a 640MB disk (format windows95 it said, so maybe FAT32) as -t
vfat, this is not a thing to try.
also, 2.0.x kernels don't support 2048b block size (also 640MB disks).
a patch for 2.0.3x kernels seems to float around somewhere in Japan,
but i have not yet gotten hold of it. here a link that certainly has
an English description:
http://elektra.e-technik.uni-ulm.de/~mbuck/linux/patches.html
or search the u-tokyo.ac.jp domain. the page of the developers is
hidden somewhere.
the best way to use these 640MB disks is therefore to do fdisk and
mkfs first. i have only done this with mke2fs on type 83 partitions:
mke2fs -b 2048 /dev/sdxy
i will check it out for FAT16 partitions and mkdosfs when i have some
spare time and disks.
my kernel version used is 2.1.124 (for all of the above).
Steve
--
***********************cut*here*or*do*not********************************
Stephan Shuichi
office: Dept. for Mechano-Informatics, Yoshizawa Lab.
Faculty for Engineering, University of Tokyo
Tel 03-3812-2111 ext 6390, FAX 03-5802-2957
email stephan@bios.t.u-tokyo.ac.jp
http://www.bios.t.u-tokyo.ac.jp/~stephan/
private: --
You've probably already received a number of messages regarding the Fujitsu DynaMO 640 - I have the 640SZI, which is the internal version; the model number given in a SCSI probe is M2513-MCC3064SS. I recently installed this drive practically without a hitch. I say practically because the sector size of the 640 MB disks is 2048 bytes, which is not supported in the Linux 2.0.x kernel but is supported in the development kernels. A patch for 2.0.x is available at http://wwwcip.informatik.uni-erlangen.de/~orschaer/mo/ -- also at this site is a patched fdisk to use in conjunction with it. Otherwise, installing the drive was no different from installing a SCSI hard drive. It runs well, and I'm very happy with it. Phil Garcia
philip_kerr_at_wmc__brsf2@wmcmail.wmc.ac.uk
Dear Skip
In your Optical HOWTO, you asked for anyone else's experiences of
installing optical drives under Linux.
Please find below details of how I managed to get a Panasonic LF-7010
(SCSI) working on my Sparc Classic.
I'm using Redhat, 4.2 and 5.1
Regards
Philip Kerr
philip.kerr@wmc.ac.uk
ps I'm now trying to get the drive to work under Solaris 2.6... it's
not an easy a job as it was under Linux!!
------------------------
plugged the drive in (on id5)...
powered up the Sparc...
the following came up....
scsi0 : Sparc ESP100A-FAST
scsi : 1 host.
Vendor: SAMSUNG Model: WN32162U Rev: 0100
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sda at scsi0, channel 0, id 3, lun 0
Vendor: MATSHITA Model: LF-7010 (00:06) Rev: 1.42
Type: Optical Device ANSI SCSI revision: 02
Detected scsi removable disk sdb at scsi0, channel 0, id 5, lun 0 scsi
: detected 2 SCSI disks total.
esp0: target 3 [period 100ns offset 15 10.00MHz FAST SCSI-II]
SCSI device sda: hdwr sector= 512 bytes. Sectors= 4236661 [2068 MB]
[2.1 GB]
esp0: target 5 [period 248ns offset 4 4.03MHz synchronous SCSI] sdb :
READ CAPACITY failed.
sdb : status = 0, message = 00, host = 0, driver = 28 sdb : extended
sense code = 2
sdb : block size assumed to be 512 bytes, disk size 1GB.
sunlance.c:v1.9 21/Aug/96 Miguel de Icaza (miguel@nuclecu.unam.mx)
eth0: LANCE 08:00:20:04:3d:cf
eth0: using auto-carrier-detection.
Partition check:
sda: sda1 sda2 sda3 sda4 sda5 sda6 sda7 sda8
sdb:scsidisk I/O error: dev 08:10, sector 0, absolute sector 0 unable
to read partition table
I edited my fstab, adding the entry for the drive (on sdb)
==========
/etc/fstab
==========
/dev/sda1 / ext2 defaults 1 1
/dev/sda2 swap swap defaults 0 0
/dev/fd0 /mnt/floppy msdos noauto,user 0 0
/dev/sr0 /mnt/cdrom iso9660 noauto,ro,user 0 0
/dev/sdb /mnt/optical ext2 noauto,rw,user 0 0
none /proc proc defaults 0 0
Then mkfs'ed a blank disc as follows...
[root@localhost me]# /sbin/mkfs -t ext2 /dev/sdb
mke2fs 1.10, 24-Apr-97 for EXT2 FS 0.5b, 95/08/09 /dev/sdb is entire
device, not just one partition! Proceed anyway? (y,n) y
Linux ext2 filesystem format
Filesystem label=
118320 inodes, 472448 blocks
23622 blocks (5.00%) reserved for the super user First data block=1
Block size=1024 (log=0)
Fragment size=1024 (log=0)
58 block groups
8192 blocks per group, 8192 fragments per group 2040 inodes per group
Superblock backups stored on blocks:
8193, 16385, 24577, 32769, 40961, 49153, 57345, 65537, 73729, 81921,
90113, 98305, 106497, 114689, 122881, 131073, 139265,
147457,
155649, 163841, 172033, 180225, 188417, 196609, 204801,
212993, 221185,
229377, 237569, 245761, 253953, 262145, 270337, 278529,
286721, 294913,
303105, 311297, 319489, 327681, 335873, 344065, 352257,
360449, 368641,
376833, 385025, 393217, 401409, 409601, 417793, 425985,
434177, 442369,
450561, 458753, 466945
Writing inode tables: done
Writing superblocks and filesystem accounting information: done
rebooted...
mounted the drive...
I've since then edited the fstab, adding the following mount-point...
/dev/sdb /mnt/dostical msdos noauto,rw,user 0 0
I can now mount ext2 or dos formatted optical carts by mounting either
optical or dostical.
Dear Skip, hoping that this is interesting for other Linux-Users, I want to tell You about my experiences with optical disks under Linux: I use an internal 640 MB MO-Drive with IDE-Interface from Fujitsu with Linux-Kernel 2.2.x and in the meantime it works fine. At Germany this drive is sold as DYNAMO 640 AI but according to it's firmware it is a MCC 3064 AP. Booting kernel 2.2.x the drive is detected like "hdx: FUJITSU MCC3064AP, ATAPI OPTICAL drive". No driver is loaded, as there still is no ATAPI- driver for optical disks. Older kernels (2.0.x) do not detect the drive correctly and surely need some patches. To use the drive I need kernel support for SCSI-emulation. So I compile this (ide-scsi.o) as a module together with the SCSI-disk-support (sd_mod.o). Making a "modprobe ide-scsi", the drive shows up in /proc/scsi/scsi. If it isn't done by kerneld I have to make "modprobe sd_mod" to be ready to mount the preformatted MO-disk. If I want to use the disks with Dos/Windows, I use the Dos/Windows-tools for formatting. I tried mkdosfs under Linux too, but then most files on the disk seemed to be corrupted for Dos. They were still o.k. for Linux and could be restored without problems. With the Dos-tools I prefer the Superfloppy-Format as this can be used with most operating-systems and it is slightly faster in comparison to a partitioned disk. This disks can be mounted like other Windows-Disks (e.g. "mount -t vfat /dev/sdx /mountpoint"). Disks for Linux should be in ext2-Format. The 640 MB disks are hardsectored with 2048 bytes/sector (smaller media aren't.) . This is no problem for kernel 2.2.x, but fdisk and mke2fs do not agree in how to manage this geometry. So I don't use fdisk anyway and format the disks with "mke2fs -b 2048 /dev/sdx". I have to tell mke2fs about the 2048 bytes/sector with the "-b"-option, otherwise the format will fail. Mke2fs than asks to really do his job, as it has do format the whole disk not a single partion and I answer with "y". Now the disk can be mounted with "mount -t ext2 /dev/sdx /mountpoint", which gives a warning in /var/log/messages about a nonexisting partion-table. This is o.k. as fdisk wasn't used and I can now use the disk. The MO-Disks are slow, but the most reliable media available. Smaller disks (230 MB) are hardsectored for 512 bytes/sector and can be partioned with fdisk. before formatting. This should be true for the 512 MB disks, but I didn't test it. Best regards and thanks for Your support for Linux, Guido Brunner
Dear Skip, I recently aquired a LF-7010 for a project. My experience getting it up for Linux under ext2 mirrors what you already have. The msdos and vfat file systems also worked. The project I was working on was based on a SunOS/Solaris formatted MO disk. While it *should* have worked under the ufs file system, using Redhat 5.2 and the stock 3.0.36 kernel it didn't. I got, installed, debugged and compiled the u2fs into the kernel and it DID mount the SunOS/Solaris MO disk. Please write if you need/want additional details. -Donald Donald Kerns <dkerns@cruzio.com>
Hi Skip, following your request of info from everyone else who is playing with optical disks under Linux, here I am. I'm an happy user of a PIONEER DE-C7001, mounted in my linux box controlled by a PCI SYMBIOS (NCR) 53c815 SCSI controller and running slackware 4. After having set the jumpers in the correct way ( all the technical docs are still available in the Pioneer web pages) and configured the kernel for SCSI support ( generic SCSI support and support for SCSI disks), it works. Here you have the boot details: sym53c8xx: at PCI bus 0, device 8, function 0 sym53c8xx: not initializing, device not supported ncr53c8xx: at PCI bus 0, device 8, function 0 ncr53c8xx: 53c815 detected ncr53c815-0: rev=0x04, base=0xe8000000, io_port=0xe000, irq=11 ncr53c815-0: ID 7, Fast-10, Parity Checking ncr53c815-0: restart (scsi reset). scsi0 : ncr53c8xx - version 3.2a-2 scsi1 : SCSI host adapter emulation for IDE ATAPI devices scsi : 2 hosts. Vendor: PIONEER Model: DE-C7001 Rev: 0500 Type: Direct-Access ANSI SCSI revision: 01 CCS Detected scsi removable disk sda at scsi0, channel 0, id 1, lun 0 Vendor: ARCHIVE Model: VIPER 150 21247 Rev: -005 Type: Sequential-Access ANSI SCSI revision: 01 Detected scsi tape st0 at scsi0, channel 0, id 2, lun 0 Vendor: PHILIPS Model: CDD3600 CD-R/RW Rev: 2.00 Type: CD-ROM ANSI SCSI revision: 02 Detected scsi CD-ROM sr0 at scsi0, channel 0, id 6, lun 0 scsi : detected 1 SCSI tape 1 SCSI cdrom 1 SCSI disk total. ncr53c815-0-<6,*>: FAST-10 SCSI 10.0 MB/s (100 ns, offset 8) sr0: scsi3-mmc drive: 2x/6x writer cd/rw xa/form2 cdda tray Uniform CDROM driver Revision: 2.55 ncr53c815-0-<1,*>: FAST-5 SCSI 5.0 MB/s (200 ns, offset 8) sda : READ CAPACITY failed. sda : status = 1, message = 00, host = 0, driver = 28 sda : extended sense code = 2 sda : block size assumed to be 512 bytes, disk size 1GB. PPP: version 2.3.7 (demand dialling) TCP compression code copyright 1989 Regents of the University of California PPP line discipline registered. 3c59x.c:v0.99H 11/17/98 Donald Becker http://cesdis.gsfc.nasa.gov/linux/drivers/vortex.html Partition check: sda:scsidisk I/O error: dev 08:00, sector 0 unable to read partition table hda: hda1 hda2 hda3 Once the drive has been recognized by the system, I inserted a new disk and I created a partition on it with: # fdisk /dev/sdanew file system on it : After this step I created a new file system : # mkfs -t ext2 /dev/sda1 Linux ext2 filesystem format Filesystem label= 79872 inodes, 318448 blocks 15922 blocks (5.00%) reserved for the super user First data block=1 Block size=1024 (log=0) Fragment size=1024 (log=0) 39 block groups 8192 blocks per group, 8192 fragments per group 2048 inodes per group Superblock backups stored on blocks: 8193, 16385, 24577, 32769, 40961, 49153, 57345, 65537, 73729, 81921, 90113, 98305, 106497, 114689, 122881, 131073, 139265, 147457, 155649, 163841, 172033, 180225, 188417, 196609, 204801, 212993, 221185, 229377, 237569, 245761, 253953, 262145, 270337, 278529, 286721, 294913, 303105, 311297 Writing inode tables: 0/39 ..... 38/39 done Writing superblocks and filesystem accounting information: done and then the disk is usable and accessible as a normal SCSI disk. I don't have modified the /etc/fstab and /etc/mtab files, because I prefer to mount the disk manually when needed. Now I need your help. In my office, I have 3 LMS ( Philips ) LF4500 rapid changer that are completely not used. The LF4500 holds 5 12" WORM optical disk. double sided 5,6 GB capacity. They were originally used with a SPARC 1+ controller and a software written for solaris 1.x. The controllers and the original sw are completely gone ( nobody knows where they are !). I will try to connect one of thisLF4500 to my linux box. I will let you know the results but I think I'll be able to see only the drive but not the disk exchange device. Could please help in find out more info/SW in order to fully drive this rapid changer ? :-) I work for Philips Medical Systems but I'm not able to find more info on this formerly Philips product. Regards Paolo Droghetti <paolo.droghetti@philips.com>
Hi Skip, I've used your 'Linux-Optical Disk HOWTO' to setup our magneto-optical drive. You mentioned somewhere in the HOWTO that you'd like to receive additional informations, and since I've used a drive which was not included, I'd like to tell you about it. Hope it can help someone! Used hardware: INTEL Pentium 90 SCSI-Controller ADAPTEC 2940 MO-Drive Fujitsu MCD3130SS (1.3 GB Capacity) Software: S.u.S.E.-LINUX 6.1, Kernel-Version 2.2.5 There is no "native" driver for the 2940AU, so I used the "aic7xxx" which I load as a module during bootup (I didn't want to compile a new kernel, because I need many other features, and expect of the MO-Drive, everything worked fine before. So, why "change a running system"?!) I can mount the MO-Disk, no matter what filesystem is used, entirely. In addition to that, I set up autofs to ease my work: in /etc/auto.misc, I added the line: ==================/SNIP/=============== /misc/mo-disk -fs auto /dev/sda1 ==================/SNAP/============== With that, I even don't need to mount the drive, I can access it whenever I want, no matter what filesystem is used (tested with MSDOS and ext2-fs) Finally, I used SAMBA to export the drive to our WIN95-Clients with the following inserted in /etc/smb.conf: =======================/SNIP/====================== [mo-disk] path = /misc/mo-disk public = yes writeable = yes ;write-only can of course still be controlled by flipping the ;write-protect-switch at the MO-Disk! readable = yes browseable = yes =======================/SNAP/======================= (for further details abt. SAMBA, refer to the excellent HOWTO) Now, every WIN-Client can use the MO-Drive as if it was a local hdd, with one (minor) caveat: When you map the exported SAMBA-Drive to a drive letter in WIN Explorer, it's impossible to umount it under LINUX! Everytime you try, you get a "device busy"... So, unfortunately I can't map the drive during startup in WIN95, but I think with some hacking in the SAMBA-Code this problem could be solved... I don't have the time at the moment, but perhaps somewhat later I will try to "dig into the code" to do the hack. Of course you can include my e-amil-address in the HOWTO, but please use my private one: dh9dat@cityweb.de instead of ds@leiterplattentechnik.de! with regards, Harald Husemann LINUX - the operating system for people whose IQ is greater than 98... Harald Husemann <ds@www.leiterplattentechnik.de>
Hi, Just stumbled across your page on Tucows (dated Dec '98 so I hope this still reaches you). You asked if anyone had any experience with optical storage etc. under Linux, which I have, so here it is! I worked for Ericsson (UK) Ltd. and some of their telephone switches use optical media for system backups. I have used these optical drives on i386 Linux boxes for some years now, with no problems whatsoever. The units in question are the Ricoh RO-5031E (scsi) and it's bigger brother, which unfortunately I cannot remember that name of (also Ricoh + scsi). The RO-5031E is a full-height, 5.25in magneto-optical drive that uses 650Mb disk cartridges (325Mb per side), such as Sony's EDM-650B. The other drive has similar spec but can use both 1.3Gb and 650Mb disks. Ricoh's website may have more on these drives, but they're quite long in the tooth now and may not feature anymore. Usage was very simple - The drives were treated almost as scsi fixed disks. Pop a new disk in, use fdisk to create your filesystem (I've tried both ext2 & msdos) then format with mkfs. That's it! The one weird thing I did find was that a RedHat 6.x system (2.2.x kernel) would not read a filesystem that had been created on an old Slackware (2.0.x kernel) system, and vice versa. Other than that, 100 million re-writes... thankyou very much!!! All the best // Jem P.S. Please feel free to include my email if I've been of any help. jem <jem@monty.ericsson.se>
I have used a Maxoptix T6-5200 with re-writable MO media without any problems. Donovan Allen admin@robot-factory.net
Maxoptix TMT3-1300 Magneto Optical drive. Accepts 1Gb and 1.3Gb magneto-optical read/write cartridges, these are double sided, so half the capacity is on each side, and the cartridge needs to be ejected to access the opposing side.
When configuring a Maxoptix drive for Linux, it should be configured such that the Removable Media Report Disable switch is OFF (the dip switch bank S2, switch 3 is OFF, i.e. in down position).
When configuring a Maxoptix drive for Linux, it should NOT be configured such that Removable Media Report Disable is switched ON (the dip switch bank S2, switch 3 is ON, i.e. in up position).
Setting this switch ON will set the RMB (removable media bit) to 0. This would indicate to Linux that the media is NOT removable, but Linux of course still allows one to eject it using the command eject /dev/sda or by pressing the invitingly large button on the drive itself. This could have the consequence of accidentally corrupting any of the good data stored on cartridges inserted subsequently, since Linux has still cached the directory information of the previous disk.
When the RMB is 0, even after the cartridge is ejected, it is possible to perform the mount command, and have it 'succeed', since Linux has still got the directory structure of the previous disk cached in its buffers, which have not been flushed. One way to force Linux to flush its buffers in this situation, is to do the following sequence: eject /dev/sda (do NOT insert a new cartridge before performing the next two steps)
mount /dev/sda5 /mnt/max umount /mnt/max Note that this will show the following on the console (Ctrl+Alt+F10): Dec 14 19:32:14 kernel: SCSI disk error : host 1 channel 0 id 6 lun 0 return code = 28000000 Dec 14 19:32:14 kernel: [valid=0] Info fld=0x0, Current sd08:05: sense key Not Ready Dec 14 19:32:14 kernel: Additional sense indicates Medium not present Dec 14 19:32:14 kernel: scsidisk I/O error: dev 08:05, sector 2 Dec 14 19:32:14 kernel: SCSI disk error : host 1 channel 0 id 6 lun 0 return code = 28000000 Dec 14 19:32:14 kernel: [valid=0] Info fld=0x0, Current sd08:05: sense key Not Ready Dec 14 19:32:14 kernel: Additional sense indicates Medium not present Dec 14 19:32:14 kernel: scsidisk I/O error: dev 08:05, sector 2 Now we can insert the new disk, and the new disk's directory structure will appear correctly as expected. This method is very error-prone, and not recommended as it is very easy to forget to perform the 'dummy' mount before inserting a disk, with the consequence of wiping out good data on the new disk. The reason this method works is that Linux normally only calls invalidate_buffers() (see the file sd.c in the routine check_media_change()) if the RMB is set to 1, and the above sequence forces Linux to call invalidate_buffers() once it notices that it can't mount the filesystem. Performing the 'dummy' mount/unmount after ejecting forces Linux to call invalidate_buffers(). For Linux, to avoid using the above workaround, we should always have the Removable Media Report Disable switch OFF (dip switch bank S2, switch 3, OFF, i.e. in down position). When the Removable Media Report Disable switch is correctly set to OFF, then attempting to mount the drive when the cartridge is not present will show the following on the console (Ctrl+Alt+F10): Dec 14 21:35:16 kernel: sda : READ CAPACITY failed. Dec 14 21:35:16 kernel: sda : status = 0, message = 00, host = 0, driver = 28 Dec 14 21:35:16 kernel: sda : extended sense code = 2 Dec 14 21:35:16 kernel: sda : block size assumed to be 512 bytes, disk size 1GB. Dec 14 21:35:16 kernel: sda:scsidisk I/O error: dev 08:00, sector 0 Dec 14 21:35:16 kernel: unable to read partition table Dec 14 21:35:16 kernel: sda : READ CAPACITY failed. Dec 14 21:35:16 kernel: sda : status = 0, message = 00, host = 0, driver = 28 Dec 14 21:35:16 kernel: sda : extended sense code = 2 Dec 14 21:35:16 kernel: sda : block size assumed to be 512 bytes, disk size 1GB. Dec 14 21:35:16 kernel: sda:scsidisk I/O error: dev 08:00, sector 0 Dec 14 21:35:16 kernel: unable to read partition table -------------------------------------------------------------------------------- When using the drive by means of an Adaptec PCMCIA card in slot 0, if disk is in drive, when the command: # cardctl insert 0 is performed, we see the following message on the console (Ctrl+Alt+F10): Dec 14 19:05:31 kernel: aha152x: processing commandline: ok Dec 14 19:05:31 kernel: aha152x: BIOS test: passed, detected 1 controller(s) Dec 14 19:05:31 kernel: aha152x0: vital data: PORTBASE=0x140, IRQ=3, SCSI ID=7, reconnect=enabled, parity=enabled, synchronous=disabled, delay=100, extended translation=disabled Dec 14 19:05:31 kernel: aha152x: trying software interrupt, ok. Dec 14 19:05:31 kernel: scsi1 : Adaptec 152x SCSI driver; $Revision$ Dec 14 19:05:31 kernel: scsi : 2 hosts. Dec 14 19:05:32 kernel: Vendor: Maxoptix Model: T3-1304 Rev: 1.1c Dec 14 19:05:32 kernel: Type: Direct-Access ANSI SCSI revision: 02 Dec 14 19:05:32 kernel: Detected scsi disk sda at scsi1, channel 0, id 6, lun 0 Dec 14 19:05:32 kernel: SCSI device sda: hdwr sector= 512 bytes. Sectors= 904995 [441 MB] [0.4 GB] Dec 14 19:05:33 kernel: sda: sda1 < sda5 > If no disk is inserted when booting Linux, get the following message on the console: Dec 14 18:55:23 kernel: Detected scsi disk sda at scsi1, channel 0, id 6, lun 0 Dec 14 18:55:24 kernel: sda: Spinning up disk....<7>ROM image dump: Dec 14 18:57:02 kernel: ...................................................................................................not responding... Dec 14 18:57:02 kernel: sda : READ CAPACITY failed. Dec 14 18:57:02 kernel: sda : status = 0, message = 00, host = 0, driver = 28 Dec 14 18:57:02 kernel: sda : extended sense code = 2 Dec 14 18:57:02 kernel: sda : block size assumed to be 512 bytes, disk size 1GB. Dec 14 18:57:02 kernel: sda:SCSI disk error : host 1 channel 0 id 6 lun 0 return code = 28000000 Dec 14 18:57:02 kernel: [valid=0] Info fld=0x0, Current sd08:00: sense key Not Ready Dec 14 18:57:02 kernel: Additional sense indicates Medium not present Dec 14 18:57:02 kernel: scsidisk I/O error: dev 08:00, sector 0 Dec 14 18:57:02 kernel: unable to read partition table Dec 14 18:57:03 kernel: scsi2 : Adaptec AHA274x/284x/294x (EISA/VLB/PCI-Fast SCSI) 5.1.25/3.2.4 Dec 14 18:57:03 kernel: <Adaptec PCMCIA SCSI controller> Dec 14 18:57:03 kernel: scsi : 3 hosts. Dec 14 18:57:08 kernel: scsi : 2 hosts. -------------------------------------------------------------------------------- When using the disks which come pre-formatted with 1024 bytes per sector, it's important to use the -b 1024 flag with fdisk, otherwise the partitioning isn't correctly written by the hardware, and mke2fs hangs, so that the system cannot be shutdown cleanly. When making the filesystem using mke2fs, don't use the defaults, since depending on the media size, the block size which is chosen by mke2fs might only be 1024, which is way too small. We want to always use 4096 as the block size, otherwise writing to the disk becomes very slow indeed. So for a disk with 1024-bytes per sector, the sequence of commands would be: # fdisk -b 1024 /dev/sda # mke2fs -b 4096 -m 0 /dev/sda5 # e2fsck -f -B 4096 /dev/sda5 -b 98304 (-b 98304 uses an alternate superblock) # mke2fs -b 4096 -m 0 /dev/sda5 mke2fs 1.18, 11-Nov-1999 for EXT2 FS 0.5b, 95/08/09 Filesystem label= OS type: Linux Block size=4096 (log=2) Fragment size=4096 (log=2) 79680 inodes, 159216 blocks 0 blocks (0.00%) reserved for the super user First data block=0 5 block groups 32768 blocks per group, 32768 fragments per group 15936 inodes per group Superblock backups stored on blocks: 32768, 98304 Writing inode tables: done Writing superblocks and filesystem accounting information: done # time (cp ~guest/kernel/linux-2.2.15.tar.gz .;sync;sync;sync) real 0m28.411s user 0m0.010s sys 0m2.590s # time (cp ~guest/kernel/*.gz .;sync;sync;sync) real 3m54.046s user 0m0.060s sys 0m2.910s ----------------------------------------------------------------------- For 512-byte sectors, the output of mke2fs is as follows for a 1G disk: # mke2fs -m 0 /dev/sda5 mke2fs 1.18, 11-Nov-1999 for EXT2 FS 0.5b, 95/08/09 Filesystem label= OS type: Linux Block size=1024 (log=0) Fragment size=1024 (log=0) 112896 inodes, 451552 blocks 0 blocks (0.00%) reserved for the super user First data block=1 56 block groups 8192 blocks per group, 8192 fragments per group 2016 inodes per group Superblock backups stored on blocks: 8193, 24577, 40961, 57345, 73729, 204801, 221185, 401409 Writing inode tables: done Writing superblocks and filesystem accounting information: done # time (cp ~guest/kernel/*.gz .;sync;sync;sync) real 3m54.046s user 0m0.060s sys 0m2.910s -------------------------------------------------------------------------------- Note the difference in storage of a disk formatted for 1024 bytes per sectos, and 512 bytes per sector, for the same 1.3GB capacity claimed on the disk label. It is about 9.5% more for the 1024 format. 1.3 GB 512 bytes/sector: df -k . Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda5 572436 740 571696 0% /mnt/max 1.3 GB 1024 bytes/sector, using 4096 block size in mke2fs: df -k . Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda5 626840 112208 514632 18% /mnt/max -------------------------------------------------------------------------------- This is the output of fdisk partitioning created by Win98 on the 1.3GB 1024 bytes/sector disk: Disk /dev/sda: 64 heads, 32 sectors, 311 cylinders Units = cylinders of 2048 * 1024 bytes Device Boot Start End Blocks Id System /dev/sda1 ? 937477 1203315 544437093 20 Unknown Partition 1 has different physical/logical beginnings (non-Linux?): phys=(356, 97, 46) logical=(937476, 3, 15) Partition 1 has different physical/logical endings: phys=(357, 116, 40) logical=(1203314, 30, 19) Partition 1 does not end on cylinder boundary: phys=(357, 116, 40) should be (357, 63, 32) /dev/sda2 ? 649505 912677 538976288 6b Unknown Partition 2 has different physical/logical beginnings (non-Linux?): phys=(288, 110, 57) logical=(649504, 0, 11) Partition 2 has different physical/logical endings: phys=(269, 101, 57) logical=(912676, 1, 10) Partition 2 does not end on cylinder boundary: phys=(269, 101, 57) should be (269, 63, 32) /dev/sda3 ? 263179 945973 1398362912 53 OnTrack DM6 Aux3 Partition 3 has different physical/logical beginnings (non-Linux?): phys=(345, 32, 19) logical=(263178, 26, 16) Partition 3 has different physical/logical endings: phys=(324, 77, 19) logical=(945972, 51, 15) Partition 3 does not end on cylinder boundary: phys=(324, 77, 19) should be (324, 63, 32) /dev/sda4 * 680971 680981 21337 49 Unknown Partition 4 has different physical/logical beginnings (non-Linux?): phys=(87, 1, 0) logical=(680970, 34, 16) Partition 4 has different physical/logical endings: phys=(335, 78, 2) logical=(680980, 61, 8) Partition 4 does not end on cylinder boundary: phys=(335, 78, 2) should be (335, 63, 32) Command (m for help): q Similarly, this is the output of fdisk partitioning created by Win98 on the 1GB 512 bytes/sector disk: # fdisk -b 512 /dev/sda Command (m for help): p Disk /dev/sda: 64 heads, 32 sectors, 441 cylinders Units = cylinders of 2048 * 512 bytes Device Boot Start End Blocks Id System /dev/sda1 ? 937477 1203315 272218546+ 20 Unknown Partition 1 has different physical/logical beginnings (non-Linux?): phys=(356, 97, 46) logical=(937476, 3, 15) Partition 1 has different physical/logical endings: phys=(357, 116, 40) logical=(1203314, 30, 19) Partition 1 does not end on cylinder boundary: phys=(357, 116, 40) should be (357, 63, 32) /dev/sda2 ? 649505 912677 269488144 6b Unknown Partition 2 has different physical/logical beginnings (non-Linux?): phys=(288, 110, 57) logical=(649504, 0, 11) Partition 2 has different physical/logical endings: phys=(269, 101, 57) logical=(912676, 1, 10) Partition 2 does not end on cylinder boundary: phys=(269, 101, 57) should be (269, 63, 32) /dev/sda3 ? 263179 945973 699181456 53 OnTrack DM6 Aux3 Partition 3 has different physical/logical beginnings (non-Linux?): phys=(345, 32, 19) logical=(263178, 26, 16) Partition 3 has different physical/logical endings: phys=(324, 77, 19) logical=(945972, 51, 15) Partition 3 does not end on cylinder boundary: phys=(324, 77, 19) should be (324, 63, 32) /dev/sda4 * 680971 680981 10668+ 49 Unknown Partition 4 has different physical/logical beginnings (non-Linux?): phys=(87, 1, 0) logical=(680970, 34, 16) Partition 4 has different physical/logical endings: phys=(335, 78, 2) logical=(680980, 61, 8) Partition 4 does not end on cylinder boundary: phys=(335, 78, 2) should be (335, 63, 32) Question: Which partition is the data in, and how do we mount it? None of the start/end values make any sense to Linux. -------------------------------------------------------------------------------- Comparison of media sizes and speeds: ===================================== Using the following files: # ls -l ~guest/kernel/*gz -rw-r----- 1 guest users 16371764 May 24 2000 /home/guest/kernel/linux-2.2.15.tar.gz -rw-rw-rw- 1 guest users 17106471 Jun 15 2000 /home/guest/kernel/linux-2.2.16.tar.gz -rw-r--r-- 1 guest users 20881782 May 25 2000 /home/guest/kernel/linux-2.3.99-pre9.tar.gz -rw-rw-rw- 1 guest users 21085275 Jun 16 2000 /home/guest/kernel/linux-2.4.0-test1.tar.gz -rw-r--r-- 1 guest users 22888582 Oct 3 08:36 /home/guest/kernel/linux-2.4.0-test9.tar.gz The tests consisted of simply copying the above files to the drive. Sometimes the media is new, sometimes previously written to. It seems that new media is slightly faster to write to. %-------------------------------------------------------------------------------% On media identified as follows: P/N 2015382-0010 Max-GL (Optical Glass) Jukebox Certified Rewritable 1.3GB 512 # fdisk /dev/sda Command (m for help): p Disk /dev/sda: 64 heads, 32 sectors, 568 cylinders Units = cylinders of 2048 * 512 bytes Device Boot Start End Blocks Id System /dev/sda1 1 568 581616 5 Extended /dev/sda5 1 568 581600 83 Linux # mke2fs -m 0 -b 4096 /dev/sda5 mke2fs 1.18, 11-Nov-1999 for EXT2 FS 0.5b, 95/08/09 Filesystem label= OS type: Linux Block size=4096 (log=2) Fragment size=4096 (log=2) 72800 inodes, 145400 blocks 0 blocks (0.00%) reserved for the super user First data block=0 5 block groups 32768 blocks per group, 32768 fragments per group 14560 inodes per group Superblock backups stored on blocks: 32768, 98304 Writing inode tables: done Writing superblocks and filesystem accounting information: done Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda5 572436 20 572416 0% /mnt/max # time (cp ~guest/kernel/linux-2.2.15.tar.gz .;sync;sync;sync) real 0m29.099s user 0m0.010s sys 0m1.430s # time (cp ~guest/kernel/*.gz .;sync;sync;sync) real 4m18.446s user 0m0.010s sys 0m3.880s %-------------------------------------------------------------------------------% On media identified as follows: P/N 1015386RW Max-GL (Optical Glass) Jukebox Certified Rewritable 1GB 512 fdisk shows: Disk /dev/sda: 64 heads, 32 sectors, 441 cylinders Units = cylinders of 2048 * 512 bytes Device Boot Start End Blocks Id System /dev/sda1 1 441 451568 5 Extended /dev/sda5 1 441 451552 83 Linux # mke2fs -m 0 -b 4096 /dev/sda5 mke2fs 1.18, 11-Nov-1999 for EXT2 FS 0.5b, 95/08/09 Filesystem label= OS type: Linux Block size=4096 (log=2) Fragment size=4096 (log=2) 112896 inodes, 112888 blocks 0 blocks (0.00%) reserved for the super user First data block=0 4 block groups 32768 blocks per group, 32768 fragments per group 28224 inodes per group Superblock backups stored on blocks: 32768, 98304 Writing inode tables: done Writing superblocks and filesystem accounting information: done Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda5 437384 20 437364 0% /mnt/max # time (cp ~guest/kernel/linux-2.2.15.tar.gz .;sync;sync;sync) real 0m34.220s user 0m0.100s sys 0m5.610s # time (cp ~guest/kernel/*.gz .;sync;sync;sync) real 4m8.846s user 0m0.280s sys 0m28.500s %-------------------------------------------------------------------------------% On media identified as follows: (obtained from eBay:Borismcbin) MaxEP Rewritable 1GB 512 (Tahiti P/N 1015387-0040) Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda5 437384 20 437364 0% /mnt/max # time (cp ~guest/kernel/linux-2.2.15.tar.gz .;sync;sync;sync) real 0m30.321s user 0m0.010s sys 0m1.340s # time (cp ~guest/kernel/*.gz .;sync;sync;sync) real 3m32.851s user 0m0.010s sys 0m4.340s %-------------------------------------------------------------------------------% On media identified as follows: (obtained from eBay:Surpuluseller) MaxEP Rewritable 1.2GB 512 (P/N: PN2015383RW) "Maxoptix PN2015383RW FORMATTED ERASABLE OPTICAL CARTRIDGE 1.2 GIGABYTE 512 BYTES/SECTOR" # fdisk /dev/sda Command (m for help): p Disk /dev/sda: 64 heads, 32 sectors, 568 cylinders Units = cylinders of 2048 * 512 bytes Device Boot Start End Blocks Id System /dev/sda1 1 568 581616 5 Extended /dev/sda5 1 568 581600 83 Linux # mke2fs -m 0 -b 4096 /dev/sda5 mke2fs 1.18, 11-Nov-1999 for EXT2 FS 0.5b, 95/08/09 Filesystem label= OS type: Linux Block size=4096 (log=2) Fragment size=4096 (log=2) 72800 inodes, 145400 blocks 0 blocks (0.00%) reserved for the super user First data block=0 5 block groups 32768 blocks per group, 32768 fragments per group 14560 inodes per group Superblock backups stored on blocks: 32768, 98304 Writing inode tables: done Writing superblocks and filesystem accounting information: done Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda5 572436 20 572416 0% /mnt/max # time (cp ~guest/kernel/linux-2.2.15.tar.gz .;sync;sync;sync) real 0m28.979s user 0m0.000s sys 0m2.600s # time (cp ~guest/kernel/*.gz .;sync;sync;sync) real 4m0.486s user 0m0.000s sys 0m1.400s %-------------------------------------------------------------------------------% On media identified as follows: HEWLLET PACKARD REWRITABLE OPTICAL DISK (Type R/W - CC Format) 1.3 Gbytes 1024 Byte/Sector Reorder No: 92280T Made in Japan. # fdisk -b 1024 /dev/sda Command (m for help): p Disk /dev/sda: 64 heads, 32 sectors, 311 cylinders Units = cylinders of 2048 * 1024 bytes Device Boot Start End Blocks Id System /dev/sda1 1 311 636896 5 Extended /dev/sda5 1 311 636864 83 Linux Command (m for help): q Filesystem 1k-blocks Used Available Use% Mounted on /dev/sda5 626840 20 626820 0% /mnt/max # mke2fs -m 0 -b 4096 /dev/sda5 mke2fs 1.18, 11-Nov-1999 for EXT2 FS 0.5b, 95/08/09 Filesystem label= OS type: Linux Block size=4096 (log=2) Fragment size=4096 (log=2) 79680 inodes, 159216 blocks 0 blocks (0.00%) reserved for the super user First data block=0 5 block groups 32768 blocks per group, 32768 fragments per group 15936 inodes per group Superblock backups stored on blocks: 32768, 98304 Writing inode tables: done Writing superblocks and filesystem accounting information: done # time (cp ~guest/kernel/linux-2.2.15.tar.gz .;sync;sync;sync) real 0m28.411s user 0m0.020s sys 0m1.570s # time (cp ~guest/kernel/*.gz .;sync;sync;sync) real 4m19.854s user 0m0.010s sys 0m2.350s %-------------------------------------------------------------------------------% Peter Knaggs <Peter.Knaggs@oracle.com>
Alexander supplied me with some very informative points. He has ask me to pass that information on to you. Without fully understanding all areas that he knows, I am attempting to pass it on to you as best I can;
There are two kind of 3.5" Magneto Optical drives : DynamMO and GigaMO
DynaMO : 128Mb, 230Mb, 540Mb and 640Mb (this drive has 2048Kb/sector)
GigaMO : 1.3Gb, 2.3Gb (2048Kb/sector)
(Sony, Fujitsu) has a models with different interfaces : SCSI, IDE/ATAPI, USB, FireWire, PCMCIA. I think, we should expand your HOWTO to this models.
For example, to work with IDE/ATAPI interface you should install "ide-scsi" kernel module :
# modprobe ide-scsi # dmesg .... sscsi1 : SCSI host adapter emulation for IDE ATAPI devices Vendor: IomintSA Model: MCE3130AP-MO1300 Rev: 0011 Type: Optical Device ANSI SCSI revision: 02 Attached scsi removable disk sdb at scsi1, channel 0, id 1, lun 0 SCSI device sdb: 605846 2048-byte hdwr sectors (1241 MB) sdb: Write Protect is off sdb:
You could use "mtool" package to access a FAT/VFAT on MO. Mtools page to FAT and VFAT File systems
First of all, you should define a new drive M: in /etc/mtools.conf ===== drive m: file="/dev/sdb" ===== # minfo m: # mformat m: # mdir m: You should use -I key to format MO media with mkfs.vfat : # mkfs.vfat -I -v /dev/sdb Unfortunately, Linux kernel uses bogus MO drive geometry even for SCSI drives. It makes sense only for BPB FAT filesystem (ext2fs does not depend on drive geometry). To compatibility with Windows system you should use real geometry (1 Head) while formatting. (See cartridge specifications for real geometry.) For example for 230Mb DynaMO you should define in /etc/mtools.conf ===== drive m: file="/dev/sdb" cylinders=17853 heads=1 sectors=25 mformat_only ===== # mformat m: ... ## minfo m: Alexander Voropay <alec@vmb-service.ru>