How to stop GemFire processes before rebooting a host

Products

• VMware Tanzu Gemfire

Issue/Introduction

In order to apply OS patches, update certificates, and perform many forms of system maintenance in GemFire, you need to stop and restart GemFire nodes.

This article provides guidance on how to best manage such situations and how to stop any GemFire processes before rebooting a host.

Environment

Product Version: 9.10

Resolution

Whenever doing maintenance on GemFire nodes where the host itself needs a reboot, GemFire could encounter the following issues leading to membership failure:

• Disk store corruption
• Network partitioning
• Split brain

In order to avoid such problems, it is recommended that you fully stop any GemFire processes before
rebooting a host. Follow these steps to fully stop any GemFire processes:

1. Stop any GemFire processes, which can include locators and servers, on the given host.

2. Apply any patches and do any required maintenance.

3. Restart the host.

4. Restart GemFire processes on that host (Locators or Servers).

5. Confirm GemFire member processes have successfully come back online by running the following GemFire gfsh commands from gfsh and verifying the output:

• status locator [--name=value] (the ones which went down as part of the host)
• status server [--name=value] (the ones which went down as part of the host)
• status cluster-config-service
• list members
• list regions
• show missing disk-stores (it should not show any missing disk-store at this point)

6. Repeat the above steps for other hosts in the cluster.