May need to replace system drive

I rebooted my server tonight and was greeted with this message while watching the boot process via IPMI:

blk_update_request: critical medium error, dev sdi, sector 17774192

“smartctl -a /dev/sdi -d scsi” yields:

smartctl 7.0 2018-12-30 r4883 [x86_64-linux-4.12.4-1.el7.elrepo.x86_64] (local build)
Copyright © 2002-18, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Vendor: SanDisk
Product: Ultra Fit
Revision: 1.00
Compliance: SPC-4
User Capacity: 62,109,253,632 bytes [62.1 GB]
Logical block size: 512 bytes
Serial number: 4C530001251226102193
Device type: disk
Local Time is: Sat Jan 18 18:34:58 2020 CST
SMART support is: Available - device has SMART capability.
SMART support is: Enabled
Temperature Warning: Disabled or Not Supported

=== START OF READ SMART DATA SECTION ===
SMART Health Status: OK
Current Drive Temperature: 0 C
Drive Trip Temperature: 0 C

Error Counter logging not supported

Device does not support Self Test logging

So, two questions:

  1. Is there any reason I couldn’t just make an image of my system disk (64 GB, USB 3.0 flash drive), restore it to another drive, and boot from the new drive without any further configuration?
  2. Is there a “best” flash drive to use? I can find really expensive stuff, but I’m not convinced it is significantly better than the $12 Samsung flash drive.

@Noggin Hello again.
Re:

I don’t see why not. Assuming the the devices are compatible, ie size etc. It may be that you would need to edit the grub config and fstab if device names were different, but nothing that is not required for any other linux install undergoing this 'transition". And given both systems are likely to be ‘compatible’ ie both USB boot devices, this method should be good.

Only caveat I can think of is that during the copy process, or more exactly at the end of it, you don’t want to to mount both volumes, or to be safe, do any btrfs stuff while both instances of what is effectively the same pool are simultaneously connected to the same machine. Btrfs uses pool id to identify pool members and after the clone you will have 2 pools with identical uuid. So just don’t do btrfs stuff with both of them simultaneously connected as that would be bad. I think this has been addressed in much later kernels than we use in our CentOS variants but is a definite concern for our current CentOS offerings.

That’s a tricky question, and one I would like a clear answer to also. I used to recommend the SanDisk Extreme USB 3.0 32 GB and up as they have working S.M.A.R.T and are essentially a USB-to-SATA controller bridge with SSD attached and are super fast on read and write. But they have been discontinued and replaced with the “Pro” USB 3.1 variant (I have a 64 GB version) where there is, as yet, no working S.M.A.R.T. The newer Pro variant is faster on write but slower on read I think.

Which USB drive are you referencing in the following?

USB key performance came up recently and I posted a few basic speed tests. See the following forum post for these.

Incidentally, did you know that Rockstor supports custom smartctl options such as you have used in your post:

And the “-d scsi”, as well as “-d sat”, are supported. See the doc subsection: Disk Custom S.M.A.R.T Options

Hope that helps.

1 Like

@Noggin Just remembered a relevant caveate to system disk image transfere to another device re:

I had a stable subscriber report their experience of an issue after doing just this and documented it in the following GitHub issue:

So looks like you are going to run into this. But as the issue states it is currently though to be “mostly cosmetic”. I also think we may get this sorted soon in a future update as I’m scheduled to do some improvements in that area of the code for other reports of non removable ‘detached-’ disks.

Hope that helps.

It was a cheap Samsung for $12. There’s no reason to really link it. I only briefly considered it because it was Samsung and inexpensive.

I’m currently looking at the Kingston Digital HyperX Savage 64GB, trying to figure out if it support S.M.A.R.T. monitoring. From reading up on what I could find of the SanDisk Extreme, it looks like it was known for being ridiculously fast due to it having an actual sata SSD controller in it. This Kingston series seems to be in the same league, so I’m hoping that it performs well and supports SMART.

1 Like