0

My SSD RAID setup just doesn't "feel" right for the following reasons:

  1. sdb and sdc only show at 238G. These are 256G drives
  2. There are two /boot's per RAID drive. One is /boot/EFI. When I boot my device I'm presented with 6-7 options for booting. Should be just 2-3.
  3. My boot partition is full and i'm getting errors.

Computer Specs:

  • ASUS G750JZ
  • 2 256GB SSD's * Named "RAIDER"
  • 1 1TB HDD
  • No dual boot, all Ubuntu

Results of lsblk: lsblk

Anyone want to weigh in on issues they see?

3
  • 2
    The 256/238 factor is most likely just GiB versus GB Sep 30, 2016 at 23:29
  • Some SSDs reserve space for management. Sep 30, 2016 at 23:59
  • @WinEunuuchs2Unix Your comment is misleading, all HDDs and SSDs have an amount of reserved blocks for bad sectors. If he bought a drive and didn't configure anything special lsblk should give him the the size in GiB: 256×10^9÷1024^3=238,418579102.
    – LiveWireBT
    Oct 1, 2016 at 9:10

1 Answer 1

1
  1. sdb and sdc only show at 238G. These are 256G drives

256×10^9÷1024^3=238,418579102

Why are hard drives never as large as advertised?

lsblk's representation is correct. Posting text output of commands like lsblk as text or code is much more appreciated by the way.

  1. There are two /boot's per RAID drive. One is /boot/EFI. When I boot my device I'm presented with 6-7 options for booting. Should be just 2-3.

I'm not sure about your setup, a dm software raid should represent both disks as one to processes like update-grub, but it reads like another grub.cfg is also picked up, a screenshot or photo of the grub menu might help clarifying things. You could add GRUB_DISABLE_OS_PROBER=true to /etc/default/grub to not look for other operating systems bootloaders and run update-grub to see if that helps.

  1. My boot partition is full and i'm getting errors.

Which errors exactly? No it's not always obvious and there may be the cause of you other problems buried deep down there.


Your setup looks complex (dm-raid, efi and boot partitions, LVM on LUKS with nested cryptswap), are you sure do need all of this and that you want to main all of the possible breakage (distribution upgrades)? My guess is you checked all of the boxes in the installer after setting up dm-raid before, not knowing the implications and also possibly setting up encryption twice (LUKS + eCryptfs).

Instead of dm-raid you could use LVM or btrfs' RAID functionality, though I have been unable to get a working system using the desktop install media for the past few hours testing LVM and btrfs individually several times.

There is one advice I can give: You won't need more than 100MB¹ for the UEFI partition. You could create a new FAT16 partition mark it as boot and esp in gparted and copy the data from the backup of the old partition onto the new one. This will gain you 412 MB more space for the boot partition. Since 488 MB seems to be the default for all LVM setups on Ubuntu you should check why the partition got filled up, usually old kernels are removed automatically on recent releases and you can also remove them with apt autoremove.


  1. Fedora's LVM setup configures 200MB for EFI and 500MB for /boot.
2
  • thanks for the knowledge! 1. I should have remembered that from school. Lame! 2. Here is a screenshot of my ROM and BIOS screens 3. I can't remember how I setup the SDD's. I know I encrypted them, just not sure of the LUKS piece. ONE. QUICK. NOTE. > apt autoremove fixed my /boot partition issue. Removed all the old kernels, now no more error.
    – SeaDude
    Oct 7, 2016 at 5:25
  • Sorry for the late reply. Thanks for the screenshot, I didn't knew that Grub generates boot entries for the .efi binaries setting GRUB_DISABLE_OS_PROBER=true and running update-grub could probably remove then but there is not much to worry about these entries.
    – LiveWireBT
    Oct 9, 2016 at 21:02

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .