No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

High latency during vMotion process

Publication Date:  2018-03-21 Views:  131 Downloads:  2
Issue Description

There was an update of VMware up to 6.5 version on the host side, after that was made a decision to format datastores file system from VMFS5 to VMFS6.

After one of datastores was formatted, it was decided to test datastore with VMFS5 and datastore with VMFS6 inside, during that process high latency was mentioned on datastore with VMFS6.

Alarm Information

Handling Process

Feature difference between VMFS5 and VMFS6 was analyzed and compared.

It was found that in VMFS6 was added space automatically reclamation feature. It needed to manually reclaim space in VMFS5.

During the storage vMotion, every one block is migrated, this block will be reclaimed immediately by sending VMFS6 unmap command with multiple huge IOs to write zero to the storage.

These huge IO (16MB) will get higher latency than small normal R/W IO, therefore this will affect the other business performance by increasing latency.

Root Cause

Space automatically reclamation feature in VMFS6.

Solution

Disable space automatically reclamation feature of VMFS6 in VMware's vSphere and manually reclaim the space during peak-off hours.

To perform this operation check the attachment to this KB.

To find peak-off hours performance logs can be analyzed.

Suggestions

It should pay more attention to release notes/new features before perform any upgrade.

END