Uploaded image for project: 'Lustre'
  1. Lustre
  2. LU-10599

Possible for ID_FS_TYPE to be lustre / ldiskfs?

Details

    • Improvement
    • Resolution: Unresolved
    • Minor
    • None
    • None
    • None
    • 9223372036854775807

    Description

      consider a command like:

      mkfs.ext4 /dev/disk/by-id/ata-VBOX_HARDDISK_OST0PORT100000000000-part1

      This yields a UDEV event with:

      ID_FS_TYPE=ext4
      

       
      Creating a ldiskfs also sets

      ID_FS_TYPE=ext4
      

      Possible to set that to something other than ext4, I.E. lustre or ldiskfs?

      Attachments

        Issue Links

          Activity

            [LU-10599] Possible for ID_FS_TYPE to be lustre / ldiskfs?

            I agree with James in that the format of ldiskfs is as close as possible to ext4, so the ID_FS_TYPE cannot be used to distinguish them. Some options include using the ID_FS_LABEL to check for $fsname[:-]TGTxxxx, where TGT is MDT or OST or MGS/MGT.

            adilger Andreas Dilger added a comment - I agree with James in that the format of ldiskfs is as close as possible to ext4, so the ID_FS_TYPE cannot be used to distinguish them. Some options include using the ID_FS_LABEL to check for $fsname [:-] TGTxxxx , where TGT is MDT or OST or MGS/MGT .

            one day ext4 will be ldiskfs so we need to think of another way.

            simmonsja James A Simmons added a comment - one day ext4 will be ldiskfs so we need to think of another way.
            utopiabound Nathaniel Clark made changes -
            Link New: This issue is related to LU-12530 [ LU-12530 ]
            ofaaland Olaf Faaland added a comment -

            It would be useful for the purposes of LU-12530 if there was a property of an ldiskfs-formatted device that definitely distinguished it from an ext-formatted device.

            ofaaland Olaf Faaland added a comment - It would be useful for the purposes of LU-12530 if there was a property of an ldiskfs-formatted device that definitely distinguished it from an ext-formatted device.
            simmonsja James A Simmons made changes -
            Link New: This issue is related to LU-10756 [ LU-10756 ]
            joe.grund Joe Grund made changes -
            Description Original: consider a command like:

            {{mkfs.ext4 /dev/disk/by-id/ata-VBOX_HARDDISK_OST0PORT100000000000-part1}}

            This yields a UDEV event with:

            {code}
            ID_FS_TYPE=ext4
            {code}
             
            Creating a ldiskfs FS also sets

            {code}
            ID_FS_TYPE=ext4
            {code}

            Possible to set that to something other than {{ext4}}, I.E. {{lustre}} or {{ldiskfs}}?
            New: consider a command like:

            {{mkfs.ext4 /dev/disk/by-id/ata-VBOX_HARDDISK_OST0PORT100000000000-part1}}

            This yields a UDEV event with:

            {code}
            ID_FS_TYPE=ext4
            {code}
             
            Creating a ldiskfs also sets

            {code}
            ID_FS_TYPE=ext4
            {code}

            Possible to set that to something other than {{ext4}}, I.E. {{lustre}} or {{ldiskfs}}?
            joe.grund Joe Grund made changes -
            Description Original: consider a command like:

            {{mkfs.ext4 /dev/disk/by-id/ata-VBOX_HARDDISK_OST0PORT100000000000-part1}}

            This yields a UDEV event like:

            {code}
            UDEV [1191.465945] change /devices/pci0000:00/0000:00:0d.0/ata4/host3/target3:0:0/3:0:0:0/block/sdb/sdb1 (block)
            ACTION=change
            DEVLINKS=/dev/disk/by-id/ata-VBOX_HARDDISK_OST0PORT100000000000-part1 /dev/disk/by-path/pci-0000:00:0d.0-ata-2.0-part1 /dev/disk/by-uuid/a0a50f45-ed91-47c9-b0f9-87c1bc7fb03e
            DEVNAME=/dev/sdb1
            DEVPATH=/devices/pci0000:00/0000:00:0d.0/ata4/host3/target3:0:0/3:0:0:0/block/sdb/sdb1
            DEVTYPE=partition
            ID_ATA=1
            ID_ATA_FEATURE_SET_PM=1
            ID_ATA_FEATURE_SET_PM_ENABLED=1
            ID_ATA_SATA=1
            ID_ATA_SATA_SIGNAL_RATE_GEN2=1
            ID_ATA_WRITE_CACHE=1
            ID_ATA_WRITE_CACHE_ENABLED=1
            ID_BUS=ata
            ID_FS_TYPE=ext4
            ID_FS_USAGE=filesystem
            ID_FS_UUID=a0a50f45-ed91-47c9-b0f9-87c1bc7fb03e
            ID_FS_UUID_ENC=a0a50f45-ed91-47c9-b0f9-87c1bc7fb03e
            ID_FS_VERSION=1.0
            ID_MODEL=VBOX_HARDDISK
            ID_MODEL_ENC=VBOX\x20HARDDISK\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20\x20
            ID_PART_ENTRY_DISK=8:16
            ID_PART_ENTRY_NUMBER=1
            ID_PART_ENTRY_OFFSET=2048
            ID_PART_ENTRY_SCHEME=dos
            ID_PART_ENTRY_SIZE=10483712
            ID_PART_ENTRY_TYPE=0x83
            ID_PART_TABLE_TYPE=dos
            ID_PATH=pci-0000:00:0d.0-ata-2.0
            ID_PATH_TAG=pci-0000_00_0d_0-ata-2_0
            ID_REVISION=1.0
            ID_SERIAL=VBOX_HARDDISK_OST0PORT100000000000
            ID_SERIAL_SHORT=OST0PORT100000000000
            ID_TYPE=disk
            MAJOR=8
            MINOR=17
            SEQNUM=1816
            SUBSYSTEM=block
            TAGS=:systemd:
            USEC_INITIALIZED=151903
            {code}
             
            It would be useful if {{mkfs.lustre}} would also emit a UDEV event when creating a filesystem.
            New: consider a command like:

            {{mkfs.ext4 /dev/disk/by-id/ata-VBOX_HARDDISK_OST0PORT100000000000-part1}}

            This yields a UDEV event with:

            {code}
            ID_FS_TYPE=ext4
            {code}
             
            Creating a ldiskfs FS also sets

            {code}
            ID_FS_TYPE=ext4
            {code}

            Possible to set that to something other than {{ext4}}, I.E. {{lustre}} or {{ldiskfs}}?
            joe.grund Joe Grund made changes -
            Summary Original: Add ability for Lustre to update ID_FS UDEV properties. New: Possible for ID_FS_TYPE to be lustre / ldiskfs?
            pjones Peter Jones made changes -
            Labels Original: LTS
            joe.grund Joe Grund added a comment -

            Sorry, let me update ticket to reflect my intent better

            joe.grund Joe Grund added a comment - Sorry, let me update ticket to reflect my intent better

            People

              utopiabound Nathaniel Clark
              joe.grund Joe Grund
              Votes:
              0 Vote for this issue
              Watchers:
              7 Start watching this issue

              Dates

                Created:
                Updated: