Upgrade cluster versions - Elasticsearch - Alibaba Cloud Documentation Center

This guide outlines how to upgrade the version or kernel of your Alibaba Cloud Elasticsearch cluster, ensuring your business leverages the latest features and improvements. It covers supported upgrade paths, cluster restart behaviors, and essential pre-upgrade considerations.

Pre-upgrade checklists

Before initiating any upgrade, carefully review and complete the following steps to minimize risks and ensure a smooth process.

Test Environment Validation:
Always perform the upgrade in a test environment or on a test cluster first before proceeding with a production cluster. This helps identify potential issues specific to your setup.
Data Backup:
Create a full snapshot of your cluster immediately before the upgrade to back up all data. In case of upgrade failure, you can restore data from this snapshot. (The upgrade check process will also verify this.)
Off-Peak Hours:
Schedule the upgrade during off-peak hours when your cluster experiences minimal traffic. Avoid creating new indices or performing heavy write/query operations during the upgrade process.
Compatibility Check:
- Breaking changes, minor changes, and deprecations may be involved in the target version. Consult the official Elastic migration guides: Migration guide.
  Note
  Versions 8.9 and 8.15 include breaking changes. Other 8.x versions involve only minor changes. For more information, see Migrating to 8.15 and Migrating to 8.9.
- Plugin compatibility:
  If you use Elasticsearch plugins, verify that all plugin versions are compatible with the destination cluster version. Incompatible plugins may fail to work or cause cluster instability.
- Associated service compatibility:
  If your cluster integrates with other Elastic Stack components (e.g., Logstash, APM, Beats, Fleet/Elastic Agent, Elastic Security, Enterprise Search), confirm their compatibility with the target Elasticsearch version.
Kibana Unavailability:
The upgrade process involves a Kibana version change, so Kibana may be temporarily unavailable during the upgrade.
Client Configuration:
Be aware that IP addresses of cluster nodes might change depending on the restart mode (see below). If your client applications or other services hardcode node IP addresses, you will need to update their configurations post-upgrade.

Version compatibility

Alibaba Cloud Elasticsearch supports specific upgrade paths. If your cluster's current version does not meet these requirements, you cannot directly upgrade. In such cases, consider migrating to a new cluster with the desired version.

Source version	Target version	Instructions
5.5.3	5.6.16	V5.x and V6.x version or kernel upgrade
5.6.16	6.3.2
6.3.2	6.7.0
7.x	Not directly upgradable	Migrate to a new 8.x instance.
8.x	Latest available 8.X version	V8.x to the latest version (phased upgrade)

Understand cluster restart modes

After an upgrade (version or kernel), your cluster will undergo a restart. A restart mode is automatically decided based on your scenario. The restart duration typically ranges from a few minutes to about an hour, depending on cluster specifications, data volume, and data structure.

Restart mode	Mechanism	Impact	Typical scenarios
Blue/green deployment	Adds new nodes (with the target version/kernel) to the cluster, migrates data to these new nodes, and removes the original nodes.	Cannot be canceled once started. Read/Write operations are supported during the upgrade, but other operations are not. Node IP addresses change. If client configurations hardcode node IPs, update them post-upgrade. Recommendation: Perform the upgrade during off-peak hours.	Version upgrades: 5.5.3 to 5.6.16, 6.3.2 to 6.7.0, and 8.x to the latest version Kernel upgrades
In-place change	All nodes are shut down, upgraded, and then restarted.	Service interruption will occur. Data loss is not expected, but data will be unavailable during the restart. Upgrade duration depends on cluster specifications and data volume. Node IP addresses do not change. No client configuration updates are needed. Recommendation: Plan the upgrade in advance and schedule during a dedicated maintenance window.	Version upgrades: 5.6.16 to 6.3.2

Procedure

V5.x and V6.x version or kernel upgrade

Log on to the Alibaba Cloud Elasticsearch console.
In the left navigation menu, choose Elasticsearch Clusters.
Navigate to the target cluster.
1. In the top navigation bar, select the resource group to which the cluster belongs and the region where the cluster resides.
2. On the Elasticsearch Clusters page, find the cluster and click its ID.
On the Basic Information page, in the upper-right corner, click Update and Upgrade.
In the Update and Upgrade dialog box, set Operation Type as needed.
- Kernel Patch Update: Upgrades the cluster's kernel without changing the Elasticsearch version. (Supported only for Kernel-enhanced Edition clusters). For more information about the features of each kernel version, see Kernel version release notes.
  Note
  If a new kernel patch is available, a message like "A new kernel patch Is available" will appear next to Version field on the Basic Information page. Click this message to initiate the kernel upgrade.
- Elasticsearch Version Upgrade: Upgrades the Elasticsearch version without changing the kernel.
Click Upgrade Check.
Alibaba Cloud Elasticsearch will perform upgrade check on configuration compatibility, resource sufficiency, cluster status, and recent snapshots.
Note
- If no snapshot backup has been completed in the last hour, click Manual Backup in the dialog to trigger a manual snapshot operation. For more information, see Data backup.
- The time required for a snapshot backup depends on the data volume of the cluster. If automatic snapshot backup is not enabled for the cluster and the data volume is large, the first snapshot backup may take a long time.
(Optional) Resolve any exceptions as prompted on the page, and click Re-check.
Once all checks pass, click Upgrade.
Monitor the upgrade progress in Tasks. After completion, the new version will be reflected on the Basic Information page.

V8.x to the latest version (phased upgrade)

Log on to the Alibaba Cloud Elasticsearch console.
In the left navigation menu, choose Elasticsearch Clusters.
Navigate to the target cluster.
1. In the top navigation bar, select the resource group to which the cluster belongs and the region where the cluster resides.
2. On the Elasticsearch Clusters page, find the cluster and click its ID.
In the upper-right corner of the Basic Information area, click Update and Upgrade.
In the Version Selection step of the Update and Upgrade panel, configure the following parameters:
1. Set Operation Type to Elasticsearch Version Upgrade.
2. Set Destination Version.
3. Review the risks and select the confirmation checkbox.
4. Click Next.
(Optional) In the Upgrade Check step, perform pre-upgrade checks:
1. If no cluster snapshot is taken within the previous hour, click Manual Backup to create a snapshot.
2. If the cluster health status or basic resource checks fail, see Upgrade check for troubleshooting.
3. Click Re-check.
After all checks pass, click Next.
In the Plugin Configuration step, configure plugins as needed.
If Check Passed is displayed in the Plugin Configuration area, the plugin is successfully configured.
Click Upgrade. In the dialog box, click OK.
During the upgrade, service jitter may occur due to node changes. Make sure that you perform this operation during off-peak hours. The version upgrade consists of two stages:
1. Phased change: The initial stage involves a partial update. After this phase, immediately confirm cluster stability and check your business functionality. If an exception occurs, consult technical support to request a rollback.
2. Full change: If you confirm your cluster is stable and operating normally after the phased change, click Full Change to complete the upgrade.
  Warning
  After a full change is initiated, rollback is no longer supported.

Elasticsearch:Upgrade cluster versions