Linux and the hard drive, LVM, SSD, USB, capacity, speed and lifespan

Mass External Storage

For mass external storage I use Seagate Expansion USB3 drives, I have 2 5TB and 4 8TB drives.  I thought they were the same drive except for capacity but I'm here waiting for 2.2TB of data to write to the 5TB drive, and 16 hours later, it's still not finished.  Yesterday day I was able to fill the 8TB drive with 5TB data in a few hours (I didn't time it).  Near as I can tell, the 5TB drive has the "Seagate 5TB BarraCuda 5400 RPM 128MB Cache" inside and the 8TB has the "Seagate Barracuda 7200 RPM 256MB Cache" inside.  I'm guessing platter density is where most of the speed boost is coming from, heads are writting 2x more per pass, and the higher RPM and larger cache helps, but this 5x faster write speed on the 8TB drive is just amazing!

At $140 for 8TB, I don't see any other choice for cheap external storage, at that price I can buy three 8TB drives for the cost of one 14TB drive.  And I don't see any other choice for the next 5-10 years, the new drive head tech is coming later in 2019 with 20TB drives, but I expect it to be $1,000+, I doubt we'll see those 20TB drives hit the sub $200 price point in less than 5 years.

2020 - now the best value is the Seagate Expansion 10TB on Amazon, for around $180. (In 2021 I shelled out $270 for the 14TB) And when they say 10TB they mean 9TB, formatted:

 Number  Start (sector)    End (sector)  Size       Code  Name
   1              40     19532873693   9.1 TiB     8300  Linux filesystem

 lsusb -t - will tell you what speed (USB) the drive is attached at, sometimes if you had a USB 2 (480M) device plugged to that USB port you may not get USB 3 speed (5000M) until you reboot.

amber<534>%  lsusb -t

/:  Bus 04.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/4p, 5000M

    |__ Port 1: Dev 2, If 0, Class=Mass Storage, Driver=uas, 5000M

    |__ Port 3: Dev 3, If 0, Class=Mass Storage, Driver=uas, 5000M

    |__ Port 4: Dev 5, If 0, Class=Mass Storage, Driver=uas, 5000M

/:  Bus 03.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/4p, 480M

    |__ Port 2: Dev 3, If 0, Class=Mass Storage, Driver=uas, 480M

/:  Bus 02.Port 1: Dev 1, Class=root_hub, Driver=ehci-pci/3p, 480M

    |__ Port 1: Dev 2, If 0, Class=Hub, Driver=hub/8p, 480M

/:  Bus 01.Port 1: Dev 1, Class=root_hub, Driver=ehci-pci/3p, 480M

    |__ Port 1: Dev 2, If 0, Class=Hub, Driver=hub/6p, 480M

For backup of the most critical backups, like photographs, I'm using MDisc BD, said to last 1,000 years.

When will my internal SSD die?

smartctl --all /dev/sda tells me my Samsung EVO 850 has 32541 Power_On_Hours, nearly 4 years, and Total_LBAs_Written is 60265626638, 28.06 TB.  Only 33 Power-On cycles.  Should I worry?  No one knows for sure about SSD lifespan, with tests and opinions varying, except that the first generation SSDs died young and the next gen like the Samsung EVO's seems to live forever.  I don't expect any problems, I have a EVO 840 that's 6 years old and no sign of problems.  If and when I replace the EVO 850 likely I would go with the Seagate 2TB FireCuda, it's the best of the SSD+HD world.

In 2022 - NVMe pretty much owns the internal drive space, 2T for $200, and incredibly fast.

Formatting steps:

dmesg = drive letter "?" (or from mount, did it automount??)

0) umount !! ; dmesg -T

1) gdisk -l /dev/sdb (got the right drive???)

2) gdisk /dev/sdb

3) d = delete existing partion

4) n = create new, 8300 for Linux

5) w = write/exit

6) power off, or reread??

7) gdisk -l /dev/sdb ; umount  /run/media/charles/EFI

8) mkfs -V -t ext4 /dev/sdb1 = newfs

9) tune2fs -L seagate5 /dev/sdb1

10) tune2fs -m 0 /dev/sdb1

11) vi /etc/fstab, LABEL=seagate5  /seagate5       ext4    defaults 0 0 (use tabs!!)

12) mkdir /seagate5; mount /seagate5 ; df

Note - there are differences between umount and eject that can bite you!

May need a reboot to reset the USB port speed, if you had an optical in that port.

for i in 1 2 3; do hdparm -tT /dev/sdb; done

If dmesg is showing "Ring error" that may mean bad cable.

Mounting by label / file system busy error

Say you pull a drive from another system, in the old days one could mount the drive simply as /dev/sdb1. But if it's a LVM partition it's a bit more complicated and you likely will get the "file system busy" error when you try to mount it. Simple solution is to attach the drive via a USB enclosure and "ls /dev/mapper". Likely you'll see your drive there as the label it's know by and you can then mount that label:

locnar<76># ls /dev/mapper

control                       fedora_localhost--live-swap@  fedora_miner-swap@

fedora_localhost--live-home@  fedora_miner-home@

fedora_localhost--live-root@  fedora_miner-root@

add this line to fstab:

/dev/mapper/fedora_miner-home /mnt/d2    ext4    defaults 1 1              

mkdir /mnt/d2; mount /mnt/d2

External USB case, pop in 2.5 or 3.5 drives