Linux's persistent naming scheme for disks

At OLS this year, I mentioned about the persistent device naming scheme that Kay demoed. This is now implemented in three distros (OpenSUSE, Gentoo, and Debian) for people to play around with (maybe four, haven't looked at Red Hat rawhide in a while.) So, if you want to see how this is solved for Linux, try one of those distros out and look at /dev/disk:

$ tree /dev/disk/
/dev/disk/
|-- by-id
|   |-- ata-TOSHIBA_MK4018GAP_23388192T -> ../../hda
|   |-- ata-TOSHIBA_MK4018GAP_23388192T-part1 -> ../../hda1
|   |-- ata-TOSHIBA_MK4018GAP_23388192T-part2 -> ../../hda2
|   |-- ata-TOSHIBA_MK4018GAP_23388192T-part3 -> ../../hda3
|   |-- ata-TOSHIBA_MK4018GAP_23388192T-part4 -> ../../hda4
|   |-- usb-Generic_USB_Flash_Drive_200506062389 -> ../../sdd
|   |-- usb-Generic_USB_Flash_Drive_200506062389-part1 -> ../../sdd1
|   |-- usb-Lexar_Media_Inc._CF_20030128160726900 -> ../../sdb
|   |-- usb-Lexar_Media_Inc._CF_20030128160726900-part1 -> ../../sdb1
|   |-- usb-Lexar_Media_Inc._SD.MS_20030128160726900 -> ../../sdc
|   `-- usb-Lexar_Media_Inc._SM.xD_20030128160726900 -> ../../sda
|-- by-label
|   |-- 1YET57WW -> ../../sdd1
|   |-- EOS_DIGITAL -> ../../sdb1
|   `-- home -> ../../hda4
|-- by-path
|   |-- pci-0000:00:0f.0-ide-0:0 -> ../../hda
|   |-- pci-0000:00:0f.0-ide-0:0-part1 -> ../../hda1
|   |-- pci-0000:00:0f.0-ide-0:0-part2 -> ../../hda2
|   |-- pci-0000:00:0f.0-ide-0:0-part3 -> ../../hda3
|   |-- pci-0000:00:0f.0-ide-0:0-part4 -> ../../hda4
|   |-- usb-0x05dc-0xb18d:0:0:0 -> ../../sda
|   |-- usb-0x05dc-0xb18d:0:0:1 -> ../../sdb
|   |-- usb-0x05dc-0xb18d:0:0:1-part1 -> ../../sdb1
|   |-- usb-0x05dc-0xb18d:0:0:2 -> ../../sdc
|   |-- usb-0x1043-0x8006:0:0:0 -> ../../sdd
|   `-- usb-0x1043-0x8006:0:0:0-part1 -> ../../sdd1
`-- by-uuid
    |-- 1553-17E8 -> ../../sdd1
    |-- 599b4559-2591-4649-a0e2-2547ef07cf33 -> ../../hda2
    |-- 7e80a6b0-918f-407b-a9d5-70858980a4ce -> ../../hda1
    `-- d6d10936-82a6-4884-8317-784c90a63830 -> ../../hda4

This machine has:

  • a single ide hard drive with 4 partitions, one of them labeled home.
  • a USB storage reader (3 slots, 6 or so storage types), with a Compact Flash card in it.
  • a USB flash stick with the label 1YET57WW on its first partition (why, I have no idea, it was a give-away at OSCON this year...)

So, another huge bureaucratic mess gives no real help and the developers go off and solve the problem on their own. When will they ever learn...

posted Thu, 18 Aug 2005 in [/diary]

oscon talk

My OSCON 2005 talk was reported on. Pretty good summary of it, and even the OS News thread about it was pretty sane. Not sure what this all means...

posted Thu, 18 Aug 2005 in [/diary]

todo lists

Going to take the advice here and try to put everything into one big file to see if that helps out any. Right now I use my email client to handle my todo list (different folders mean different things left to attend to.) Not the best scheme, but it's worked out so far alright.

posted Thu, 18 Aug 2005 in [/diary]

meta-bugs

Wee, this is a great idea. I've now created a kernel meta bug for USB and used it to clean out a whole lot of crufty old reports. Hopefully it will make it easier to track stuff that other developers are assigned easier (once the bug gets assigned away from me, I just forgot about it, which isn't always the best thing over time.)

posted Thu, 18 Aug 2005 in [/diary]

mimed patches

Yeah Dave, I know what you mean about getting patches all crusty with mime. You should see the ones that are always base-64 encoded. Anyway, James passed on a little perl script that he wrote to handle this when you absolutly-can't-get-the-sender-to-fix-their-email-client. Have fun with it.

posted Thu, 18 Aug 2005 in [/diary]


   



My Linux Stuff


RSS