Upgrade WEKA versions
Upgrade your WEKA system with the latest version.
Last updated
Upgrade your WEKA system with the latest version.
Last updated
The WEKA upgrade process supports non-disruptive upgrades (NDUs) to ensure minimal impact on system operations. When planning an upgrade, adhere to the following guidelines:
Supported source versions: To upgrade to version 4.4.1, the minimum supported source version is 4.2.1. The maximum supported source versions are 4.2.12.92, 4.3.5.105, and 4.4.0.
Upgrade path: Upgrades must progress from an older version to a newer version. For example, upgrading from version 4.2.X to 4.4.Y is permitted only if version 4.4.Y was released after the 4.2.X Long-Term Support (LTS) release.
Major version compatibility: Upgrades must remain within the same major version or transition from an older major version to its direct successor.
Version-specific exceptions: Exceptions to upgrade paths may exist. Always confirm the supported minimum and maximum versions for the desired upgrade. For detailed information on release dates and compatible versions, visit get.weka.io.
Examples:
Permitted upgrade: Upgrading from version 4.2.12.92 LTS (released August 1, 2024) to version 4.4.1 (released November 5, 2024) is allowed.
Restricted upgrade: Upgrading from versions 4.2.14, 4.2.15, and 4.2.16 to version 4.4.1 is not allowed.
Always review release notes and guidelines to confirm upgrade eligibility for your specific version.
The source system must be set up in MCB architecture. If not, contact the Customer Success Team to convert the cluster architecture to MCB. See Convert cluster to multi-container backend. This workflow is only intended for professional services.
Customers running WEKA clusters on AWS with auto-scaling groups must contact the Customer Success Team before converting the cluster to MCB.
In MCB architecture, each container serves a single type of process, drive, frontend, or compute function. Therefore, upgrading one container at a time (rolling upgrade) is possible while the remaining containers continue serving the clients.
Some background tasks, such as snapshot uploads or downloads, must be postponed or aborted. See the prerequisites in the upgrade workflow for details.
Once you run the upgrade command in ndu
mode, the following occurs:
Downloading the version and preparing all backend servers.
Rolling upgrade of the drive containers.
Rolling upgrade of the compute containers.
Rolling upgrade of the frontend and protocol containers and the protocol gateways.
Related topics
WEKA containers architecture overview
Adhere to the following considerations:
Protocol separation: Upgrading a WEKA cluster with a server used for more than one of the following protocols, NFS, SMB, or S3, is not permitted. If such a case arises, the upgrade process does not initiate and indicates the servers that require protocol separation. Contact the Customer Success Team to ensure only one additional protocol is installed on each server.
Legacy NFS protocol: If a legacy NFS protocol is implemented, contact the Customer Success Team. In this case, the upgrade is blocked.
NFS file-locking prerequisite before upgrade: Ensure the rpc.statd
and rpc-statd-notifiy
services are stopped on the WEKA servers. If not, run the following commands:
systemctl disable rpc-statd.service
systemctl disable srpc-statd-notify-service
S3 Cluster Creation: If you plan to create an S3 cluster, it’s crucial to ensure the upgrade process is complete and all containers are up before initiating the creation.
Ensure the environment meets the necessary prerequisites before proceeding with any system upgrade. The WEKA Upgrade Checker Tool automates these essential checks, comprehensively assessing the system’s readiness. Whether performing a single-version upgrade or a multi-hop upgrade, following this procedure is mandatory.
Passed checks (Green): The system meets all prerequisites for the upgrade.
Warnings (Yellow): Address promptly to resolve potential issues.
Failures (Red): Do not proceed; they may lead to data loss.
Multi-hop version upgrades:
After completing an upgrade, a background process initiates the conversion of metadata to a new format (in specific versions). This conversion may take several minutes before another upgrade can commence. To monitor the progress, use the weka status
CLI command and check if a data upgrade task is RUNNING.
By diligently following this system readiness validation procedure, you can confidently proceed with system upgrades, minimizing risks and ensuring a smooth upgrade.
Prioritize running the WEKA Upgrade Checker 24 hours before any scheduled upgrades. This step is critical to identify and address any potential issues proactively.
Ensure passwordless SSH access is set up on all backend servers. This is crucial for the seamless execution of the Python script while running the WEKA Upgrade Checker.
Log in to one of the backend servers as a root user:
Access the server using the appropriate credentials.
Obtain the WEKA Upgrade Checker: Choose one of the following methods:
Method A: Direct download
Clone the WEKA Upgrade Checker GIT repository with the command:
git clone https://github.com/weka/tools.git
Method B: Update from existing tools repository
If you have previously downloaded the tools repository, navigate to the tools directory.
Run git pull
to update the tools repository with the latest enhancements. (The WEKA tools, including the WEKA Upgrade Checker, continuously evolve.)
Run the WEKA Upgrade Checker: Navigate to the weka_upgrade_checker directory. It includes a binary version and a Python script of the tool. A minimum of Python 3.8 is required if you run the Python script.
Run the Python script: python3.8 ./weka_upgrade_checker.py
Or
Run the Python precompiled script ./weka_upgrade_checker
The tool scans the backend servers and verifies the upgrade prerequisites.
Review the results:
Pay attention to the following indicators:
Green: Passed checks. Ensure the tool's version is the latest.
Yellow: Warnings that require attention and remedy.
Red: Failed checks. If any exist, do not proceed. Contact the Customer Success Team.
Send the log file to the Customer Success Team:
The weka_upgrade_checker.log
is located in the same directory where you ran the tool. Share the latest log file with the Customer Success Team for further analysis.
Download the new WEKA version to one of the backend servers using one of the following methods depending on the cluster deployment:
Method A: Using a distribution server
Method B: Direct download and install from get.weka.io
Method C: If the connectivity to get.weka.io is limited
For details, select the relevant tab.
Use this method if the cluster environment includes a distribution server from which the target WEKA version can be downloaded.
If the distribution server contains the target WEKA version, run the following commands from the cluster backend server:
Where: <version> is the target WEKA version, for example: 4.4.0
.
If the distribution server does not contain the target WEKA version, add the option --from
to the command, and specify the get.weka.io distribution site, along with the token.
Example:
When working with many backend servers, preparing them separately from the upgrade process in advance is possible to minimize the total upgrade time. For a small number of backend servers, this step is not required.
The preparation phase prepares all the connected backend servers for the upgrade, which includes downloading the new version and getting it ready to be applied.
Once the new version is downloaded to one of the backend servers, run the following CLI command:
weka local run --container drives0 --in <new-version> upgrade --prepare-only
Where:
<new-version>
: Specify the new version. For example,4.4.1
.
Once a new software version is installed on one of the backend servers, upgrade the entire cluster backend servers to the new version by running the following command on the backend server.
If you already ran the preparation step, the upgrade command skips the download and preparation operations.
weka local run --container drives0 --in <new-version> upgrade
Consider the following guidelines:
Before switching the cluster to the new software release, the upgrade command distributes the new release to all cluster servers. It makes the necessary preparations, such as compiling the new wekafs
driver.
If a failure occurs during the preparation, such as a disconnection of a server or failure to build a driver, the upgrade process stops, and a summary message indicates the problematic server.
If cleanup issues occur during a specific upgrade phase, rerun it with the relevant option:
In a successful process, the upgrade stops the cluster IO service, switches all servers to the new release, and then turns the IO service back on. This process takes about 1 minute, depending on the cluster size.
If the container running the upgrade process uses a port other than the default (14000), include the option --mgmt-port <existing-port>
to the command.
Enabling the Low Latency Queue (LLQ) improves data processing efficiency in AWS by reducing I/O operation delays. LLQ is enabled by default after an upgrade, but if Write Combining (WC) is not activated in the igb_uio
driver, the LLQ driver option does not function. After upgrading the backends, verify that WC is enabled.
Procedure
Check for upgrade events:
Review the upgrade events on the backends.
If NetDevDriverReloadFailed
appears, restart the WEKA service by running the following commands on each backend server:
Verify WC activation:
Check if WC is activated by running:
If the output is #1
, WC is activated, which enables the LLQ driver option.
Once all backends are upgraded, the clients remain with the existing version and continue working with the upgraded backends. The client's version can only be one major version behind the version of the backends. Therefore, clients must be upgraded before the next cluster software version upgrade. The minimum source version for client upgrades is 4.3.X.
If a stateless client is mounted on a single cluster, it is automatically upgraded to the backend version after rebooting, or a complete umount
and mount
is performed.
If a stateless client is mounted on multiple clusters, the client container version is the same as the client-target-version
in the cluster (see Mount filesystems from multiple clusters on a single client).
Stateless clients can also be upgraded manually.
Use the --client-only
flag in the weka version get
command to ensure that only the essential components relevant to the stateless client operation are downloaded, excluding non-relevant packages.
To limit the display of versions unless the complete set of components is present, use the --full
flag with the weka version
command This provides you with finer control over version information visibility.
You can manually upgrade the clients locally (one by one) or remotely (in batches), usually during a maintenance window.
Clients can be upgraded manually. This can be done either locally on each client individually or remotely in batches. This process typically occurs during a scheduled maintenance window.
An upgrade is performed on a gateway, which is a persistent client that runs a specific protocol. This gateway is associated with containers with the allow_protocols
parameter set to true. The upgrade process involves interaction with backend servers.
To upgrade a stateless or persistent client locally, connect to the client and run the following command line:
Run: weka version get <target-version> --from <backend name or IP>:<port>
Upgrade the agent by running the following (replace the x with the latest minor version):
weka version set --agent-only 4.4.x
Upgrade the client containers. Do one the following following:
For clients connected to a single cluster, run weka local upgrade
For clients connected to a multiple clusters, upgrade all containers simultaneously by running weka local upgrade --all
An alert is raised if there is a mismatch between the clients' and the cluster versions.
Add the --from <backend name or IP>
option to download the client package only from the backend, thus avoiding downloading from get.weka.io. The default port is 14000.
Once the upgrade is complete, verify that the cluster is in the new version by running the weka status
command.
Example: The following is returned when the system is upgraded to version 4.4.0:
# weka status
Weka v4.4.0
...