Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


RackNerd Review
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

RackNerd Review

archbeardarchbeard Member
edited April 2021 in Reviews

At first things were fantastic, great IO:

# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync;unlink test
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 3.56865 s, 301 MB/s

But then things got slow on March 31st.

# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 30.8056 s, 34.9 MB/s

RackNerd did some restarts and adjusted some things that got IO back up to 200MB/s.

# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync;unlink test
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 5.26892 s, 204 MB/s

But then today I checked my VPS to find it taking over 30 seconds to connect on SSH, finding that odd, I checked the IO to find:

# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync;unlink test
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 108.711 s, 9.9 MB/s

The response to RackNerd was not anything helpful other than to recreate the VPS with their template rather than using a Debian ISO (I want LUKS + LVM).

After a restart:

# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync;unlink test
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 80.995 s, 13.3 MB/s

However after a few minutes of support I noticed the IO improved.

# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync;unlink test
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 27.8868 s, 38.5 MB/s

At this point I'm at a loss. None of my VPS at Ramnode, Virmach, BuyVM have issues with IO like this. Is it really me using Debian with LUKS + LVM causing the poor IO? I was under the impression once I unlocked the LUKS it behaved the same as any other LVM EXT4 linux install. If I'm wrong I would like to learn how to improve things. The VPS has literally nothing on it, I had wireguard on March 20th and removed it recently. It's fully up to date with the latest Debian updates etc, no root login allowed on ssh, ssh keys only for my user to ssh with, fail2ban installed etc basics.

--Cody

Thanked by 1MagicApple

Comments

  • dustincdustinc Member, Patron Provider, Top Host

    Hello Cody,
    Thank you so much for being our customer and for providing us with the opportunity to work with you! I took a quick glance over your support ticket and it appears that your system has disk-encryption set up within the VM, and the VPS was installed from ISO.

    As your VPS was installed from a custom/uploaded ISO and not from one of our available OS templates, we would also recommend ensuring you are utilizing virtio drivers within your installation to take advantage of maximum Disk I/O benefits.

    Additionally, I performed a Disk I/O test on the host node your VPS is on just now and the results look normal (I'll be sharing these results with you by e-mail as well so we can compare the differences and identify what is causing the discrepancy).

    I will be reaching out to you by e-mail now to help troubleshoot the issue and will personally look into it. I am always here to help.

    Thank you again for working with us.

  • jcarlo9jcarlo9 Barred
    edited April 2021

    On which vps location are you on ?

  • @jcarlo9 said:
    On which vps location are you on ?

    Atlanta

  • archbeardarchbeard Member
    edited April 2021

    @dustinc said:
    Hello Cody,
    Thank you so much for being our customer and for providing us with the opportunity to work with you! I took a quick glance over your support ticket and it appears that your system has disk-encryption set up within the VM, and the VPS was installed from ISO.

    As your VPS was installed from a custom/uploaded ISO and not from one of our available OS templates, we would also recommend ensuring you are utilizing virtio drivers within your installation to take advantage of maximum Disk I/O benefits.

    Additionally, I performed a Disk I/O test on the host node your VPS is on just now and the results look normal (I'll be sharing these results with you by e-mail as well so we can compare the differences and identify what is causing the discrepancy).

    I will be reaching out to you by e-mail now to help troubleshoot the issue and will personally look into it. I am always here to help.

    Thank you again for working with us.

    Thank you Dustin. That being said, it's odd that I only have this issue at your hosts not with my other VPS which are also Debian 10 with LVM + LUKS.

    I'll check if the virtio drivers are being initialized.

  • dustincdustinc Member, Patron Provider, Top Host

    @archbeard said: Thank you Dustin. That being said, it's odd that I only have this issue at your hosts not with my other VPS which are also Debian 10 with LVM + LUKS.

    Thank You! Email sent.

    We are eager to help get to the bottom of it. :#

  • @dustinc said:

    @archbeard said: Thank you Dustin. That being said, it's odd that I only have this issue at your hosts not with my other VPS which are also Debian 10 with LVM + LUKS.

    Thank You! Email sent.

    We are eager to help get to the bottom of it. :#

    Replied to it, thank you Dustin.

  • This is a LET support request, but not a review.

  • Good thing you paid 3 years upfront for it

  • FalzoFalzo Member

    @corbpie said:
    Good thing you paid 3 years upfront for it

    and got the bandwidth doubled ;-) ;-) ;-)

    on a more serious note: maybe just ask for a migration to another node. I'd imagine over the pricing they raked in a lot of 'noisy' neighbours, so problems like this are to be expected. expectation management.

    Thanked by 1JasonM
  • Oh, I regularly check CPU performance and it degrades a lot(from GB5 ~1300 to ~700) since I ordered, but I forgot to check IO performance.

    So I just did a check:(debian 10 from their template):
    $ dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync;unlink test
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB, 1.0 GiB) copied, 2.03011 s, 529 MB/s

    Seems fine.

    Thanked by 1scooke
  • thanks for your nice review , your I/O has been doubled :#

    Thanked by 2Pilotseye WSWD
Sign In or Register to comment.