Data resync taking a longer time while performing the host reinstallation
search cancel

Data resync taking a longer time while performing the host reinstallation

book

Article ID: 446168

calendar_today

Updated On:

Products

VMware vSAN

Issue/Introduction

This article provides guidance on minimizing data resync volume and duration while performing host maintenance activities, such as ESXi reinstallation, by optimizing vSAN object repair timers.

  • vSAN maintenance mode using Full Data Migration takes several hours (4-5 hours) to complete.

  • Significant resync time and data movement observed during ESXi host reinstallation or boot device replacement.

  • Need to perform host maintenance without triggering full object rebuilds or long evacuation times.

Environment

VMware vSAN (All Versions)

Cause

By default, vSAN initiates a full rebuild of components if a host is absent for more than 60 minutes. Additionally, selecting the Full Data Migration option during Maintenance Mode forces all data to be evacuated to other hosts, which is time-consuming in clusters with large data sets.

Resolution

To minimize resync and maintenance time, follow the steps below to temporarily increase the repair delay and use the 'Ensure Accessibility' MM option.

  1. To avoid resync/data migration, please make sure the repair object timer is increased higher than the time in minutes expected for the host reinstallation (consider additional buffer time).  The Maximum value you can set in minutes for repair object timer is 4294967295 (minutes). Please find the below article for your reference.
    Changing the default repair delay time for a host failure in vSAN

  2. Select Ensure accessibility option for maintenance mode.

Note:

  • The object will stay in reduced availability during the time mentioned in the repair object timer or until the host is exited from maintenance mode. Hence, if there are more failures in other hosts in the cluster, that can lead to data unavailability or data loss.

  • Resync is expected to trigger after the host is added back to the cluster to update the delta changes.

  • Recommended to verify the impact of taking the host into maintenance mode using Data migration precheck.

  • Recommended to have backup for all the critical Virtual machines before the activity.

Additional Information

Changing the default repair delay time for a host failure in vSAN