Scheduled Maintenance: We are aware of an issue with Google, AOL, and Yahoo services as email providers which are blocking new registrations. We are trying to fix the issue and we have several internal and external support tickets in process to resolve the issue. Please see: viewtopic.php?t=158230

 

 

 

Stretch odd bad disk performance in VMWare ESXi

Linux Kernel, Network, and Services configuration.
Post Reply
Message
Author
agpastore
Posts: 1
Joined: 2017-09-11 15:20

Stretch odd bad disk performance in VMWare ESXi

#1 Post by agpastore »

Hi.

I've run into a strange issue with a brand new installation of the latest Stretch (9.1) as a virtual machine on ESXi.

I have some HP blade servers running ESXi 6.0U3 with VMDKs stored on a IBM DS4700 storage unit.
So far so good with Wheezy & Jessie.

When it comes to Strech it looks like the virtual machine is performing very badly from a disk i/o point of view.
I've noticed it on MySQL INSERT tests.

For this reason, I've done a clean install of a Jessie x64 VM and a Stretch x64 VM with no specific services other than ssh.
ESXi VM Version is 11; both have 1 vCPU and 1 GB RAM, a single 8 GB VMDK with one root partition; SCSI controller is VMWare Paravirtual,
and both machines are running their open-vm-tools.

I tried to test small consecutive write operations using dd and I've found
a huge difference between Debian v7/v8 and Debian v9.

This is the command I've used:

Code: Select all

dd if=/dev/zero of=/root/test bs=512 count=1000 oflag=dsync
This is the output on any Debian but 9 (I had similar results on v7 and v8):

Code: Select all

1000+0 records in
1000+0 records out
512000 bytes (512 kB) copied, 1.77293 s, 289 kB/s
This is the output on Debian Stretch:

Code: Select all

1000+0 records in
1000+0 records out
512000 bytes (512 kB, 500 KiB) copied, 9.77881 s, 52.4 kB/s
so on Stretch it takes 6x times than older releases.

Any clue on that?

Thanks,
Alberto.

Post Reply