Using the vSphere Storage Appliance 1.0 Manager GUI to enter cluster maintenance mode can result in clusters entering an inconsistent state
search cancel

Using the vSphere Storage Appliance 1.0 Manager GUI to enter cluster maintenance mode can result in clusters entering an inconsistent state

book

Article ID: 308054

calendar_today

Updated On:

Products

VMware vSphere ESXi

Issue/Introduction

VMware has become aware of an issue where clusters can enter an inconsistent state when placed into cluster maintenance mode using the vSphere Storage Appliance Manager. This inconsistency can lead to virtual machines becoming inaccessible.

This article provides work around steps to take an entire VMware vSphere Storage Appliance (VSA) 1.0 cluster offline for maintenance.
Typical use cases for this procedure include:
  • Preparing a VSA Cluster for a scheduled power down or generator test
  • Preparing a VSA Cluster for relocation to a different physical location


Environment

VMware vSphere Storage Appliance 1.0.x

Resolution

VMware recommends that you do not use the Cluster Maintenance functions in the VSA Manager GUI to take a VSA 1.0 cluster offline. Instead, use the steps in this article to take down a cluster and bring it back up after maintenance is performed. If you have already used the vSA Manager GUI to enter cluster maintenance mode, contact VMware Support.

Taking down the VSA cluster

To take down the VSA cluster for maintenance:
  1. Shut down all the virtual machines as gracefully as possible. For example, use RDP to log in and shut the machines down using the appropriate operating system procedure.
  2. Right-click on the HA Cluster (if enabled) and deselect HA.
  3. Enter appliance Maintenance mode for as many nodes as possible.

    Note: You will likely see an error alert for the last node, as the VSA tries to prevent all of the nodes from entering maintenance mode.

  4. After all the nodes except one are in Maintenance Mode (you may need to refresh the display several times to see this), launch the vSphere Client direct to the host on which the last remaining node is registered.
  5. Power off the last remaining VSA appliance node. Right-click the appliance node and click Power > Power off.

    Note: As outlined in step 4, this requires that you connect the vSphere Client directly to the host, since from the vCenter Server the Power Off function is greyed out.

  6. Shut down the hosts.

Bringing up the VSA Cluster

To bring up the VSA cluster after maintenance:
  1. Power on the hosts.
  2. After the hosts are up and connected to the vCenter Server, the VSAs datastores appear as inaccessible. This is expected. Power on each appliance. Right-click on each appliance (for exmaple, VSA-0 and VSA-1) and click Power > Power On.
  3. Wait a few minutes (approximately 10 minutes or less) for the automated processes of the VSA to determine if the storage can be brought online.

    If the storage remains inaccessible, open a console to each appliance, log in (root / svapass), and determine if the appliance node is in Maintenance Mode with the command:

    wscli 127.0.0.1 getSvaServerInfo

    If in maintenance mode, the output includes maintenance mode = true.

  4. If the appliance node shows in Maintenance Mode, run this wscli command to exit Maintenance Mode:

    wscli 127.0.0.1 exitMaintenanceMode

  5. Repeat steps 3 and 4 for each appliance node. After a few minutes, the storage should appear.
  6. Right-click on the HA Cluster (if present) and select HA.
  7. When these tasks are complete in the Recent Tasks window of the vCenter Server, power on the remaining virtual machines in an appropriate priority sequence for your environment (for example, domain controller, DNS server, email server).


Additional Information

For translated versions of this article, see:
vSphere Storage Appliance 1.0 Manager GUIを使用して、クラスタメンテナンスモードへ切り替えの際に、クラスタの状態が不整合となる