New to Linode, rather disappointed with disk (Dreamhost / Linode) Comparison


I am new to Linode. Moved to Linode from Dreamhost Dreamcompute. The reasons for moving were for better IP Address management for mail servers (aka PTR records) to integrate with Barracude ESS.

Its doing everything I want, but see below the comparison of Dreamhost Disk speeds with Linode. (tests done on very similar spec VMs with clean fresh Ubuntu 14.04.5).

I would happily pay the 48 USD to get those speeds back. It certainly feels laggy to use at OS level.

VM Type & spec: gp1.warpspeed
Location USA East
OS: Ubuntu 14.04.5
Price: 48 USD / mo

Disk Speeds - Test 1:
root@groupware:~# hdparm -t /dev/vda
Timing buffered disk reads: 484 MB in 3.00 seconds = 161.30 MB/sec

Disk Speeds - Test 2:
root@groupware:~# hdparm -T /dev/vda
Timing cached reads: 15234 MB in 1.99 seconds = 7643.95 MB/sec

VM Type & spec: Linode 8192
Location Frankfurt DE
OS: Ubuntu 14.04.5
Price: 40 USD / mo

Disk Speeds - Test 1:
root@groupware:~# hdparm -t /dev/sda
Timing buffered disk reads: 170 MB in 3.00 seconds = 56.67 MB/sec

Disk Speeds - Test 2:
root@groupware:~# hdparm -T /dev/sda
Timing cached reads: 9820 MB in 2.00 seconds = 4920.16 MB/sec



5 Replies

Hi Alex,

I ran these same tests on one of my Linodes just now:

aj@nodejs:~$ hdparm -t /dev/sda
aj@nodejs:~$ sudo !!
sudo hdparm -t /dev/sda
[sudo] password for aj:

 Timing buffered disk reads: 2792 MB in  3.01 seconds = 927.84 MB/sec

aj@nodejs:~$ sudo hdparm -T /dev/sda

 Timing cached reads:   15624 MB in  1.99 seconds = 7840.82 MB/sec

Usually there is some discrepancy with these types of benchmarks, but clearly something else is going on here.

I would suggest opening up a support ticket so the team can make sure there are no issues from the infrastructure side of things.

There are also some common issues you can look for with the internals of your Linode:

Full Disk - If your disk is at or near full you will never get full performance. When creating your Linode it's possible you didn't size your disk to use all of the available disk space. If this is the case you can resize it and test your speeds again.

Aged Disk - Even if your disk is not full, it can become fragmented over time so it's recommended to rebuild your setup and then copy over any necessary files

The last thing that may play into this is that you are using an older version of Ubuntu. I'm not overly familiar with this but my tests were run on a Linode running 16.04. If you wanted to consider running a newer version we have another guide that walks you through the update process.


Thanks for the reply.

What region are you in? I am in Frankfurt.

Each of the VMs both Dreamhost and Linode are brand new, fresh Ubuntu 14.04.5 installs, with only apt-get update && apt-get upgrade performed.

Yes hdparm is not the best tool in the world, but I have run these multiple times, and the results are very consistent.

I will open a ticket.

Off Topic: How did you make your code look pretty?



Just another thing.

Its great to see Linode Staff being active in the forums.

Really appreciated, many thanks.


The tests I ran were on a Linode in Newark.

I see, I wasn't sure if you were copying over an existing image or booting from a fresh install.

The code block in markdown is done with 3 backticks on each side of the code. Like this:


My issue is solved.

I opened a ticket as advised, and the support team managed their systems, and now I have good performance as follows:

@groupware:~# hdparm -t /dev/sda

 Timing buffered disk reads: 3036 MB in  3.00 seconds = 1011.60 MB/sec
@groupware:~# hdparm -T /dev/sda

 Timing cached reads:   16764 MB in  1.99 seconds = 8405.33 MB/sec

Credit to the team for responding quickly and solving my issues on a Sunday.

Many thanks to all.



Please enter an answer

You can mention users to notify them: @username

You can use Markdown to format your question. For more examples see the Markdown Cheatsheet.

> I’m a blockquote.

I’m a blockquote.

[I'm a link] (

I'm a link

**I am bold** I am bold

*I am italicized* I am italicized

Community Code of Conduct