Failed install changed /dev names

System dual boots Windows 10 and Linux. Linux had been openSUSE for quite a while, then Mint, now replaced Mint with openSUSE Leap 15. The first try at Leap installation ended with an error. I rebooted and started installation again. When it showed the partitions, the two hard drives were named sde and sdf. Always before, including the first Leap 15 installation try, the drives were sda and sdb. I aborted installation, and rebooted to a live gparted session. It also showed the two hard drives as sde and sdf. I restarted installation and got it done correctly.

System seems to behave correctly, booting Win10 or Leap. Windows has all partitions with their correct drive letters. Leap has sde and sdf.

What became of sda and sdb? Is there any problem continuing with sde and sdf?

Info on drives is:

Drives:    HDD Total Size: 1000.2GB (18.8% used)
           ID-1: /dev/sde model: ST3750640NS size: 750.2GB
           ID-2: /dev/sdf model: ST3250620AS size: 250.1GB
Partition: ID-1: / size: 30G used: 7.2G (26%) fs: ext4 dev: /dev/sde5
           ID-2: /home size: 9.8G used: 133M (2%) fs: ext4 dev: /dev/sde6
           ID-3: swap-1 size: 17.18GB used: 0.00GB (0%) fs: swap dev: /dev/sde8

Show us the complete output of

sudo fdisk -l
sudo blkid
howard@HP-oS15KDE:~> sudo fdisk -l
[sudo] password for root: 
Disk /dev/sde: 698.7 GiB, 750156374016 bytes, 1465149168 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x9bae337b

Device     Boot      Start        End    Sectors  Size Id Type
/dev/sde1  *          2048  208783413  208781366 99.6G  7 HPFS/NTFS/exFAT
/dev/sde2        208785408  209715199     929792  454M 27 Hidden NTFS WinRE
/dev/sde3        209717248 1258293247 1048576000  500G  7 HPFS/NTFS/exFAT
/dev/sde4       1258293248 1465147391  206854144 98.7G  5 Extended
/dev/sde5       1258295296 1321209855   62914560   30G 83 Linux
/dev/sde6       1321211904 1342183423   20971520   10G 83 Linux
/dev/sde7       1342185472 1431586815   89401344 42.6G 83 Linux
/dev/sde8       1431588864 1465147391   33558528   16G 82 Linux swap / Solaris


Disk /dev/sdf: 232.9 GiB, 250059350016 bytes, 488397168 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x0008460b

Device     Boot     Start       End   Sectors   Size Id Type
/dev/sdf1            2048 294750207 294748160 140.6G  7 HPFS/NTFS/exFAT
/dev/sdf2       294750208 488396799 193646592  92.3G 83 Linux

howard@HP-oS15KDE:~> 


howard@HP-oS15KDE:~> sudo blkid
/dev/sde1: LABEL="Win10_64C" UUID="E682268582265A77" TYPE="ntfs" PARTUUID="9bae337b-01"
/dev/sde2: UUID="0A520E02520DF2ED" TYPE="ntfs" PARTUUID="9bae337b-02"
/dev/sde3: LABEL="Storage" UUID="681A714B1A711772" TYPE="ntfs" PARTUUID="9bae337b-03"
/dev/sde5: LABEL="Root" UUID="8831e0c7-c380-4647-9904-d72d0fd522f3" TYPE="ext4" PARTUUID="9bae337b-05"
/dev/sde6: LABEL="Home" UUID="98c6cb68-09c6-474f-9f64-d86bb58cd604" TYPE="ext4" PARTUUID="9bae337b-06"                                                                             
/dev/sde7: LABEL="Data" UUID="1880576f-aa73-4593-b637-7a9e7c2a1f4f" TYPE="ext4" PARTUUID="9bae337b-07"                                                                             
/dev/sde8: LABEL="Swap" UUID="ac587534-5ccd-4c8d-8cf3-38c01bc94e53" TYPE="swap" PARTUUID="9bae337b-08"                                                                             
/dev/sdf1: LABEL="Big_Storage" UUID="D8460C24460C05CA" TYPE="ntfs" PARTUUID="0008460b-01"                                                                                          
/dev/sdf2: LABEL="Backups" UUID="13e91daf-1628-4b76-a3a0-9e538336d80b" TYPE="ext4" PARTUUID="0008460b-02"                                                                          
howard@HP-oS15KDE:~>    

Show dmesg output immediately after boot (upload to http://susepaste.org).

See SUSE Paste
Regards,

sda through sdd is multifunctinal card reader.

Yes, here’s the evidence…




  1.     4.162602] scsi 8:0:0:0: Direct-Access     Generic  USB SD Reader    1.00 PQ: 0 ANSI: 0

  1.     4.162870] sd 8:0:0:0: Attached scsi generic sg0 type 0

  1.     4.216467] scsi 8:0:0:1: Direct-Access     Generic  USB CF Reader    1.01 PQ: 0 ANSI: 0

  1.     4.216736] sd 8:0:0:1: Attached scsi generic sg1 type 0

  1.     4.280465] scsi 8:0:0:2: Direct-Access     Generic  USB SM Reader    1.02 PQ: 0 ANSI: 0

  1.     4.280577] sd 8:0:0:2: Attached scsi generic sg2 type 0

  1.     4.354464] scsi 8:0:0:3: Direct-Access     Generic  USB MS Reader    1.03 PQ: 0 ANSI: 0

  1.     4.354722] sd 8:0:0:3: Attached scsi generic sg3 type 0

  1.     4.722464] sd 8:0:0:0: [sda] Attached SCSI removable disk

  1.     4.762463] sd 8:0:0:1: [sdb] Attached SCSI removable disk

  1.     4.792456] sd 8:0:0:2: [sdc] Attached SCSI removable disk

  1.     4.812587] sd 8:0:0:3: [sdd] Attached SCSI removable disk




So it’s OK to leave the system with device names as they are? It seems to be functioning fine, just unusual to have the hard drives as sde and sdf.

Is it even possible to change the device names?

Thanks,

This happens because USB storage module(s) load before HDD storage module(s), and enumeration finds the card reader before the HDDs.

If you really want sda & sdb back on HDDs, check if a BIOS upgrade is available. You may be able to get them back simply by reconfiguring BIOS USB settings, including getting USB out of or to the bottom of boot priorty. Otherwise, reconfigure dracut to exclude USB from the initrds. Back up the prior initrds before generating new ones. If no USB in initrds works, restoring the backed up initrds and moving the USB card reader cable to different motherboard connectors might also have the desired effect. Or, you might start with moving the card reader cable.

I’ll leave the naming as is for now. I’ve done numerous fresh installations of both openSUSE and Mint on this hardware, always keeping the same partitioning. This is the first time the hard drives are not sda and sdb. Both openSUSE and Windows perform fine, so it seems there is not a problem that needs fixing.
Regards,

The reason that openSUSE uses mount by UUID instead of direct by device file, is to avoid problems like your situation where the the detection sequence changes. And as you report your openSUSE is still OK, that proves that this approach works.

Yes, it should be fine.

At one time, I had a laptop where the hard drive would be “/dev/sdb” if there was a USB plugged in.

The naming appears to depend on weirdness of hardware and firmware.

My current desktop is fine with device names. However, if I use encryption with “/boot” as part of the encrypted root file system, then grub gives a message about an I/O error on disk 1 (I actually have 3 hard drives). With some experimenting, I was able to work out that grub does not give that error if there is a CD or DVD in DVD reader. So apparently, grub sees the DVD as disk 1. So I switched the SATA connectors, so that it wouldn’t be disk 1. But that did not work. Apparently the firmware always wants to make the DVD to be disk 1, no matter where it is plugged in.

no, it has nothing to do with BIOS.

Otherwise, reconfigure dracut to exclude USB from the initrds.
yes, that will probably “fix” it, except the whole is solution in search of problem. Device names are not guaranteed to be persistent across reboot by definition; nothing should rely on them.

It shouldn’t, but I’ve seen BIOS updates change device enumeration timings and thus order, hence the suggestion.

yes, that will probably “fix” it, except the whole is solution in search of problem. Device names are not guaranteed to be persistent across reboot by definition; nothing should rely on them.

That’s a good enough paradigm for automated operations, but human brains work differently than computers. My brain knows sda1 is the first partition on the first HD. It doesn’t know and will never remember a randomly generated 36 character UUID string, much less type it in a reasonable length of time.

Computers can count however they want, but humans learn in the first grade if not before: alphabetical order, 1 = first, 0 means none, e is fifth, and existence of a fifth depends on a first, second, third and fourth. Thus it logically follows that if any sdX devices exist, the first is a, second b, third c, etc. If any sdX devices exist, then sda must exist for sde to exist. That software can violate this order as OP sees is mentally discombobulative, and justification for a workaround to avoid being exposed to it.

Today, my old brain is dealing with hard drives as sde and sdf with only a twinge of discombobulation (e.g., why did it do it this time and never before?). I’ll let it ride.
Cheers,

Indeed. And, to extropolate so other Users understand, this is the reason why mounting with UUID (or, if done carefully, by Label) is the only sure way of avoiding a problem caused by device names moving around.

Might have something to do with the mushrooming kernel roughly doubling in size since systemd usurped sysvinit. That’s a lot of changes.

30123973 Apr 13  2011 ./112/kernel-desktop-2.6.31.14-0.8.1.x86_64.rpm
33902071 Apr 14  2011 ./113/kernel-desktop-2.6.34.8-0.2.1.x86_64.rpm
29111834 Feb 18  2014 ./114/kernel-desktop-3.0.101-79.1.x86_64.rpm
38123156 Nov  5  2011 ./121/kernel-desktop-3.1.0-1.2.1.x86_64.rpm
40292045 Aug  4  2012 ./122/kernel-desktop-3.4.6-2.10.1.x86_64.rpm
41542619 Dec 19  2014 ./123/kernel-desktop-3.7.10-1.45.1.x86_64.rpm
42837080 Feb  8  2016 ./131/kernel-desktop-3.11.10-34.2.x86_64.rpm
48450181 Dec  8  2016 ./132/kernel-desktop-3.16.7-53.1.x86_64.rpm
50954586 May  8  2017 ./421/kernel-default-4.1.39-56.1.x86_64.rpm
55184414 Jan  5  2018 ./422/kernel-default-4.4.104-18.44.1.x86_64.rpm
55047820 Jul 28 04:02 ./423/kernel-default-4.4.140-62.2.x86_64.rpm
53047716 Jul 28 03:55 ./150/kernel-default-4.12.14-lp150.12.7.1.x86_64.rpm
58886019 May  3  2017 ./Factory/kernel-default-4.10.13-1.2.x86_64.rpm
58571199 Aug 13  2017 ./Factory/kernel-default-4.11.8-2.4.x86_64.rpm
64982767 Aug 30  2017 ./Factory/kernel-default-4.12.9-1.2.x86_64.rpm
61065066 Nov 12  2017 ./Factory/kernel-default-4.13.12-1.1.x86_64.rpm
62547184 Jan 31 21:11 ./Factory/kernel-default-4.14.15-2.3.x86_64.rpm
64949060 Apr  4 22:24 ./Factory/kernel-default-4.15.13-2.4.x86_64.rpm
66152028 Jun 13 21:10 ./Factory/kernel-default-4.16.12-3.5.x86_64.rpm
66683292 Jul 26 06:27 ./Factory/kernel-default-4.17.9-1.5.x86_64.rpm