Salesforce

Thick or Thin provisioning for a VM hosting an OpenEdge Database?

« Go Back

Information

 
TitleThick or Thin provisioning for a VM hosting an OpenEdge Database?
URL NameIs-the-best-practice-thick-or-thin-provisioning-for-a-VM-hosting-an-OpenEdge-Database
Article Number000172580
EnvironmentProduct: OpenEdge
Version: All supported versions
OS: All supported platforms
Other: Virutal Machine thick/thin provisioning
Question/Problem Description
Is the best practice thick or thin provisioning for a VM hosting an OpenEdge Database?

What is the Progress recommendation for thick/thin provisioning on a virtual machine hosting an OpenEdge Database?
Steps to Reproduce
Clarifying Information
Error Message
Defect Number
Enhancement Number
Cause
Resolution

Progress does not recommend Thin Provisioning as this will cause performance problems since resources are not actually available until needed.

Progress's recommendation is to use Thick Provisioning:

Verify that the hard disks of the VM is using thick provisioned virtual hard drives (and not thin provisioning virtual disks). For example, the hard disk used for storing the database files should be using 'Thick eager zeroed' provisioning and not use 'Thin' provisioning 

The major problem with Thin Provisioning is when the time comes to extend the file system: 

thin provisioning == extra I/O while the virtual disk needs extending.

 
While the extra I/O overhead may be tolerable on an application VM (and save some disk space) as these are pretty much static with the exception of client-temp files and perhaps application related LOB objects or xml files that are imported. For the best database performance, avoid as much extra I/O as possible by storing database files on thick eager zeroed virtual disk and also behind thickly provisioned LUN.
 
Database servers need to be provisioned for peak loads, which is neither the average usage over 24 hours nor the "busy" time in the afternoon.  Peaks occur outside of usual daily production operations: e.g. month end processing, nightly data warehouse extracts, restore and roll-forward, index rebuild. During those times the system could run slow because the virtualization provider did not provide enough resources.

Workaround
Notes
Keyword Phrase
Last Modified Date5/15/2026 3:04 PM

Powered by