Ttyjfyk

What's a natural way to say that someone works somewhere (for a job)?

Your magic is very sketchy

What is the term when two people sing in harmony, but they aren't singing the same notes?

Generic lambda vs generic function give different behaviour

Irreducibility of a simple polynomial

There is only s̶i̶x̶t̶y one place he can be

How to combine multiple text files of different lengths and multiple columns by a column

Was the picture area of a CRT a parallelogram (instead of a true rectangle)?

Implement the Thanos sorting algorithm

Personal Teleportation as a Weapon

Ways to speed up user implemented RK4

Trouble understanding overseas colleagues

Increase performance creating Mandelbrot set in python

Teaching indefinite integrals that require special-casing

Where in the Bible does the greeting ("Dominus Vobiscum") used at Mass come from?

Mapping a list into a phase plot

What is difference between behavior and behaviour

I'm in charge of equipment buying but no one's ever happy with what I choose. How to fix this?

Why does John Bercow say “unlock” after reading out the results of a vote?

Greatest common substring

How do I keep an essay about "feeling flat" from feeling flat?

Why did Kant, Hegel, and Adorno leave some words and phrases in the Greek alphabet?

Can somebody explain Brexit in a few child-proof sentences?

Can I Retrieve Email Addresses from BCC?

Disk utilization on mdadm raid 5 is 100% even though none of the member disks is at 100%

Grow/resize RAID when upgrading visible size of disksHow to mount a disk from destroyed raid system?mdadm; previously working; after “failure”, cannot join array due to disk size“Spare” disk in a 2-disk mdadm RAID1 array?mdadm RAID 5 and parted unrecognized disk labelHow do I migrate a RAID system to a larger set of HDDS?mdadm raid filesystem differs from disk filesystemHow do I (re)build/create/assemble an IMSM RAID-0 array from disk images instead of disk drives using mdadm?mdadm 2x Raid 5 missing drivesUbuntu: RAID 10 degrades to RAID 0 after reboot?

1

1

I am currently confused about the disk utilisation of one of my machines.

The setup:

I have a machine containing 4 2TB HDDs.
For those HDDs I used mdadm to configure a RAID 5 (called md2)
On MD2, I am using a luks-volume containing a btrfs filesystem.

Besides other use cases, I also have samba installed and use it to share files and create a TimeMachine Backup.

When copying over large files, I am able to use the complete 1G connection of the machine to read and write files. So regarding this, I am happy with the speed I get.

When it comes to timemachine backups (I assume, this causes non-sequential reads/writes) the transfer starts to be very slow.

The machine is not swapping and the CPU utilization is fine as well.

When looking at netdata, I see that the utilization for "root-crypted" is constantly on 100% (root-crypted is the name of the mounted file system).

I would then expect at least one of the HDDs to also be utilized at 100%, but actually none of them is. (They all are between 40 and 60%)
Now I am wondering, why the raid does not utilize the disks by 100%.

I assume, that I might be able to improve the performance by maybe 20% if I would just manage to have the raid utilizing the disks by 100%.

Is my thinking here correct or am I missing / misunderstanding something here.
Sadly my knowledge about disks and filesystems is not good enough to understand this.
Would be cool, if you could give me some support here.

Thanks a lot :)

hard-disk performance raid btrfs mdadm

share|improve this question

asked yesterday

RacerRacer

315

• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  Isn't the encryption layer the bottleneck? Note that btrfs in my experience is slow with lots of metadata updates.
  
  – wurtel
  10 hours ago
  

  
  
  
  

add a comment | 

1

1

I am currently confused about the disk utilisation of one of my machines.

The setup:

I have a machine containing 4 2TB HDDs.
For those HDDs I used mdadm to configure a RAID 5 (called md2)
On MD2, I am using a luks-volume containing a btrfs filesystem.

Besides other use cases, I also have samba installed and use it to share files and create a TimeMachine Backup.

When copying over large files, I am able to use the complete 1G connection of the machine to read and write files. So regarding this, I am happy with the speed I get.

When it comes to timemachine backups (I assume, this causes non-sequential reads/writes) the transfer starts to be very slow.

The machine is not swapping and the CPU utilization is fine as well.

When looking at netdata, I see that the utilization for "root-crypted" is constantly on 100% (root-crypted is the name of the mounted file system).

I would then expect at least one of the HDDs to also be utilized at 100%, but actually none of them is. (They all are between 40 and 60%)
Now I am wondering, why the raid does not utilize the disks by 100%.

I assume, that I might be able to improve the performance by maybe 20% if I would just manage to have the raid utilizing the disks by 100%.

Is my thinking here correct or am I missing / misunderstanding something here.
Sadly my knowledge about disks and filesystems is not good enough to understand this.
Would be cool, if you could give me some support here.

Thanks a lot :)

hard-disk performance raid btrfs mdadm

share|improve this question

asked yesterday

RacerRacer

315

• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  Isn't the encryption layer the bottleneck? Note that btrfs in my experience is slow with lots of metadata updates.
  
  – wurtel
  10 hours ago
  

  
  
  
  

add a comment | 

I am currently confused about the disk utilisation of one of my machines.

The setup:

I have a machine containing 4 2TB HDDs.
For those HDDs I used mdadm to configure a RAID 5 (called md2)
On MD2, I am using a luks-volume containing a btrfs filesystem.

Besides other use cases, I also have samba installed and use it to share files and create a TimeMachine Backup.

When copying over large files, I am able to use the complete 1G connection of the machine to read and write files. So regarding this, I am happy with the speed I get.

When it comes to timemachine backups (I assume, this causes non-sequential reads/writes) the transfer starts to be very slow.

The machine is not swapping and the CPU utilization is fine as well.

When looking at netdata, I see that the utilization for "root-crypted" is constantly on 100% (root-crypted is the name of the mounted file system).

I would then expect at least one of the HDDs to also be utilized at 100%, but actually none of them is. (They all are between 40 and 60%)
Now I am wondering, why the raid does not utilize the disks by 100%.

I assume, that I might be able to improve the performance by maybe 20% if I would just manage to have the raid utilizing the disks by 100%.

Is my thinking here correct or am I missing / misunderstanding something here.
Sadly my knowledge about disks and filesystems is not good enough to understand this.
Would be cool, if you could give me some support here.

Thanks a lot :)

hard-disk performance raid btrfs mdadm

share|improve this question

asked yesterday

RacerRacer

315

share|improve this question

asked yesterday

RacerRacer

315

share|improve this question

asked yesterday

RacerRacer

315

asked yesterday

RacerRacer

315

asked yesterday

RacerRacer

315