Skip to content

Housecleaning…

These are all the defective harddisks in the household – I finally got to remove them, grab the files I still needed to have (in one case) and replace them.

Todays catch: all defective HDDs in the household

  • 750 GB HDD from a Buffalo USB drive, used as a time machine in the living room Mac mini
    • Date Code 07231 corresponds to 2006-12-06.
  • 1000 GB HDD in a WD Elements drive, used as a local time machine in Sammy’s computer
    • 18-Jul-2008
  • 1000 GB HDD in a WD Elements drive, used as a local time machine in my Computer.
    • I managed to get the accumulated license code archive from this one. Due to an accident, additional replica in a centralised form didn’t exist (I still have the mail archive, but there is a lot of other stuff in it)
    • 10-Sep-2007
  • 1500 GB HDD from the file server. Removed that one due to smart warnings, and a 3000 GB sister drive, placed 2x 3000 GB WD Red. Used in a BTRFS compound as a backup for the other backups.
    •  22-Apr-2010
    • Drive has been opened to show the little one the insides.

Adding backblaze external backup now, too.

The Fileserver/Data Grave now looks like this:

[root@server:~] # lsblk --nodeps --list --output NAME,VENDOR,MODEL,SIZE
NAME VENDOR   MODEL             SIZE
sda  ATA      SAMSUNG HD154UI   1.4T ?
sdb  ATA      WDC WD30EFRX-68E  2.7T btrfs  /backup
sdc  ATA      SAMSUNG HD154UI   1.4T ?
sdd  ATA      WDC WD30EFRX-68E  2.7T btrfs  /backup
sde  ATA      TOSHIBA DT01ABA3  2.7T R5/xfs /export
sdf  ATA      TOSHIBA MD04ACA4  3.7T R5/xfs /export
sdg  ATA      WDC WD30EFRX-68A  2.7T R5/xfs /export
sdh  ATA      WDC WD30EZRX-00D  2.7T R5/xfs /export
sdi  ATA      M4-CT512M4SSD2    477G SSD boot/root

The sdi is the system SSD, it just carries a fileserver Linux installation.

The sde-sdh are a RAID5 with LVM2, and a lot of ‘/export/$PURPOSE’ filesystems on them. Many of them are remote time machines for all the mobile computers in the household (all have local time machine + remote time machine).

Due to a purchase error, sdf is one TB too large, the overshoot is being used as a scratch partition outside of the MD compound.

Previously, sda, sdb, sdc and sdd have been a large 4x 1.5TB BTRFS. This is now the two new WD RED in a 2x 3.0TB BTRFS. The two 1.5TB Samsung are free and unused, I need to see what I will be doing with them.

Published inSmall computers

7 Comments

  1. Kann das Backblaze auch rsync?

    • kris kris

      I am on Linux. The backup programs used by Backblaze under their personal backup program work on Windows and Mac. Backblaze also offers B2 cloud storage, and that is supported by rclone (rsync for the cloud) and duplicity.

  2. Mike

    >The sdi is the system SSD, it just carries a fileserver Linux installation.

    Was für eine Linux-Distribution verwendest Du dort?

    • kris kris

      The SSD is under LVM2 control and contains a Suse Linux (old) and an Ubuntu (new).

  3. Disk platters make for excellent flat thin mirrors that do not immediately break when they fall on the ground.

    Ideal for some do-it-your-self stuff in the household to view hard to reach places (like installing a dishwasher this weekend).

    • kris kris

      I have been using BTRFS now as an archive disk for many years, and it worked flawlessly. That is, with the first 1.5T disks coming in, I have been making a 4-device btrfs (no mirror or other redundancy) and rsync’ing the production xfs to it once a day. That worked and never had problems, until the one 1.5T physically failed (that would make it seven years of flawless operation)

      This usage pattern does not exercise any of the patterns that are known to be problematic in BTRFS.

Leave a Reply

Your email address will not be published. Required fields are marked *