Solved

Debian Drive transfer speed on VMware ESX

Posted on 2011-03-09
7
638 Views
Last Modified: 2012-05-11
Hi all

I have an unusual problem with Disk transfer speeds on a virtual Linux environment.

This environment consists of an ESX host running 3i, 3.5.0, 153875 and 2 Debian 5.0 Linux virtual machines.

When booting the machines have a hdparm transfer rate output of:

 /dev/sda:
 Timing cached reads:   1418 MB in  2.00 seconds = 707.95 MB/sec
 Timing buffered disk reads:  132 MB in  3.01 seconds =  43.93 MB/sec

After a couple of days, the machines become sluggish and the same test produces:

/dev/sda:
 Timing cached reads:   162 MB in  2.00 seconds =  80.99 MB/sec
 Timing buffered disk reads:   12 MB in  3.78 seconds =   3.18 MB/sec


The command that is being run is hdparm -tT /dev/sda

Any ideas on why this is occurring, I cant see anything obvious in the logs that will cause this.
0
Comment
Question by:alexanderfoti
  • 4
  • 3
7 Comments
 
LVL 28

Expert Comment

by:bgoering
ID: 35083242
Is your vm configured with IDE drives?
0
 
LVL 1

Author Comment

by:alexanderfoti
ID: 35083317
No its an ESX server, you are only able to choose SCSI Devices.
0
 
LVL 28

Accepted Solution

by:
bgoering earned 500 total points
ID: 35083467
Ah, I see it is 3.5 (you can use IDE drives on 4.1). I am unable to recreate this issue on 4.1 but I suggest you take a look at http://communities.vmware.com/message/1077902#1077902, and in particular if you have a vSMP vm try it with a single virtual cpu (or disable the page sharing as in http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1004901) On 3.5 the overhead of manageing vSMP was a bit higher than on 4.x, and sometimes one could realize a huge performance increade by reducing to a single vCPU.

Here are the results from one of my 2 vCPU mail servers that has been up for a while now on ESX 4.1:
[root@prodvmail01 ~]# hdparm -tT /dev/sda

/dev/sda:
 Timing cached reads:   2396 MB in  2.00 seconds = 1197.89 MB/sec
 Timing buffered disk reads:  496 MB in  3.01 seconds = 164.95 MB/sec
[root@prodvmail01 ~]# uptime
 08:51:40 up 773 days, 11:37,  3 users,  load average: 0.14, 0.08, 0.04
[root@prodvmail01 ~]#


0
Netscaler Common Configuration How To guides

If you use NetScaler you will want to see these guides. The NetScaler How To Guides show administrators how to get NetScaler up and configured by providing instructions for common scenarios and some not so common ones.

 
LVL 1

Author Comment

by:alexanderfoti
ID: 35084455
Many Thanks.

I have turned off VSMP on both boxes and have restarted them. Now all we have to do is wait until they slow down again now!

I have also removed more than one core from the virtual machines, so at the moment, they only have 1 core each.
0
 
LVL 28

Expert Comment

by:bgoering
ID: 35084951
Sounds good - keep us posted
0
 
LVL 1

Author Comment

by:alexanderfoti
ID: 35093324
Bad news - HDparm results today:

/dev/sda:
 Timing cached reads:   1394 MB in  2.00 seconds = 695.98 MB/sec
 Timing buffered disk reads:    8 MB in  3.81 seconds =   2.10 MB/sec

Any ideas?
0
 
LVL 1

Author Comment

by:alexanderfoti
ID: 35201391
After restarting the host, it appears the drive speeds have now returned to normal. Many thanksk
0

Featured Post

Master Your Team's Linux and Cloud Stack

Come see why top tech companies like Mailchimp and Media Temple use Linux Academy to build their employee training programs.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Suggested Solutions

Title # Comments Views Activity
wifi not working on Raspberry Pi 3? 2 57
ESXi VM of Server 2003 Saving Slow. 7 60
ESXi 4.1.0 host unreachable every morning 1 47
VMWare & and Storage 10 73
HOW TO: Install and Configure VMware vSphere Hypervisor 6.5 (ESXi 6.5), Step by Step Tutorial with screenshots. From Download, Checking Media, to Completed Installation.
In this article we will learn how to backup a VMware farm using Nakivo Backup & Replication. In this tutorial we will install the software on a Windows 2012 R2 Server.
Teach the user how to rename, unmount, delete and upgrade VMFS datastores. Open vSphere Web Client: Rename VMFS and NFS datastores: Upgrade VMFS-3 volume to VMFS-5: Unmount VMFS datastore: Delete a VMFS datastore:
Get a first impression of how PRTG looks and learn how it works.   This video is a short introduction to PRTG, as an initial overview or as a quick start for new PRTG users.

803 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question