0

This is not, in my opinion, a duplicate question. Please read it first:

Preamble: Most questions relating to this topic are outdated by 5 or more years. Much has changed since then so far as filesystems go.

Considering the following:

I have a partition >2TiB on an external HDD and run on Debian mainly, with a few experiments here and there (and the intent on joining the FreeBSD Master Race when I have enough time to actually dig into a whole new OS). EXT4 is a great FS, but when I format my partition to EXT4 it never seems to stop with activity. Strange since my internal HDD is only EXT4 partitions and there is not constant activity.

XFS looks good and I've taken it for a spin and it isn't quite as hyperactive as EXT4. This is good but I really don't know much about this FS other than NASA trusts it, it's 23 years of evolution and apparently very stable these days.

NTFS is trustworthy and equally mature with XFS. I'm familiar with it, having been a Windows user up to and including 7. The main flaw with using NTFS is that there's much lower performance (~10-20M/s) on Linux (at least on my PC, I don't have Windows any more).

The actual question is multiple parts:

  1. Regarding XFS, how does it compare, in terms of reliability and performance, on a Linux OS to others (not trying to illicit opinions, but rather first-hand experience, since benchmarks are just pretty graphs)? Is XFS a good, worthy fs for long-term storage?

  2. Does the HDD casing's firmware have anything to do with the odd behaviour of, most notably, EXT4?

  3. Regarding 2, how will the firmware impact on me (in all areas)?

  4. Also regarding 2, if not the firmware, why is the HDD behaving like this with EXT4?

I've never had this problem until now. I've always used internal storage and a 1TB NTFS external from ye olde days. Any links to more useful resources would be very appreciated.

2
  • Just a suggestion--people are more likely to read the question if you eliminate all of the commentary and focus on the actual question. Short, direct questions generally do better,
    – fixer1234
    Dec 6, 2016 at 21:21
  • I'm a writer by nature, verbosity is almost unavoidable to me... Actually the truth is I am terrible at being concise. One reason I decided not to become a teacher. Even this comment, as an example. And I'm worse IRL. Still, I'll try to improve on this as I know StackExchange discourages this behaviour.
    – nonzyro
    Dec 7, 2016 at 7:36

1 Answer 1

2

What you are experiencing is not a bug, but a feature: ext4 has the property of lazy initialization. In effect this means, that only the minimum of work to make the FS usable is done on mkfs.ext4, the major task of zeroing unused space is done in the background - and this is what you perceive as "it never seems to stop with activity" or "hyperactivity". Once this has been done for the complete partition, which may take days on 2TB+ volumes, the activity stops.

XFS does no free space initialization, so you don't see this behaviour.

As for reliability: I think both file systems are highly trustworthy - personally I prefer ext4 for plain vanilla volumes, but use XFS for big file server volumes.

1
  • Thanks, that information wasn't mentioned anywhere I looked until I looked up lazy initialisation. I think this feature is a good idea on paper, but I'd rather have taken the time to complete all tasks when calling mkfs, as this can be done while I'm sleeping/working on another machine/etc. So I think I'll go with XFS, it seems good/recommended for servers which makes me feel safe, especially as I store vhd images on it, too.
    – nonzyro
    Dec 7, 2016 at 7:33

You must log in to answer this question.

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