Upgrading your API Connect solution from v50810-ifix2 to v50811

Apply a product image upgrade for each server in the Management service of your IBM® API Connect cloud environment. Then, refresh servers in the Gateway services to apply the corresponding maintenance level.

Before you begin

Note: For any service, the servers must be at the same version level.

To upgrade to v50811, the management servers must be upgraded to v50810-ifix2 first. For more information, see Upgrading your API Connect solution.

Downgrading a Management server to an earlier fix pack level is not supported.

Before you upgrade, complete the following steps:

  • Back up your API Connect cloud and data, see Preserve your cloud data.
  • If the Cloud Manager on-premises cloud contains only one Management server, ensure that the API Connect cloud is not being used while you upgrade the server because the Cloud Manager is inactive during parts of the upgrade process. If you have more than one Management server, you can continue to use the API Connect cloud while you upgrade each server.
  • You must have 16 GB of RAM to ensure that an upgrade to IBM API Connect Version 5.0 or later, is possible.
  • Login to vcenter and take a snapshot of the virtual machine.
  • SSH to the vm and run the cli command config autodiskadd disable.
  • Add a 10 GB free hard disk to the VM.
    Note: There must be only 1 free hard disk of exactly 10 GB for the upgrade to proceed.

About this task

To successfully upgrade the API Connect solution, you must complete operations on both the Management services and Gateway services. For more information, see Upgrading your API Connect solution.

Server upgrade path for API Connect
In the Upgrading Servers in the IBM API Connect diagram, there are four Management servers in the Management service and two Gateway servers in the Gateway service.

Steps 1-8 illustrate the upgrade sequence for the Management service. Upgrade the Management servers one at a time. As each Management server is upgraded, the server is temporarily removed from the cloud and remains inactive until the last server is upgraded. You can begin with any non-primary Management server in the Management service, and upgrade the primary Management server last. Verify that the upgrade is successful before you start the upgrade for the next Management server. After the successful upgrade of the last Management server, which should be the primary Management server, the primary Management server automatically activates and reactivates all the other upgraded servers in the Management service. The role of the server, for example, primary or RSS, can be viewed under server details for each server in the Services tab of the Cloud Manager.

The steps that follow illustrate the sequence in which the Gateway service should be refreshed after you upgrade the Management services. You might also need to upgrade your DataPower® product. For more information about upgrading DataPower, see Upgrading your IBM DataPower Gateways in API Connect.

To refresh the IBM API Connect Gateway service, delete and then add each Gateway server in the service, one at a time.

Lastly, verify that all the servers in the Management and Gateway services are running the same version, see Verifying maintenance level.
Note: When you compare versions, verify that the entire build number matches. For example:
5.0.0.0.20160321-1354_H9_64
5.0.0.0.20160321-1354_H9_64

not

5.0.0.0.20160321-1354_H9_64
5.0.0.0.20160209-1615_H9_64

When you apply maintenance, the new level of the product overwrites the existing level. Your user configuration, APIs, and Products are retained.

Procedure

This procedure provides the steps to upgrade an individual Management server. Repeat this procedure for each Management server in your Management service.

  1. Add new certificates to your DataPower Gateway servers.

    You must complete this step once before you upgrade to API Connect 5.0.8.7 iFix 4 or later, to prevent the loss of analytics events data during the upgrade. For instructions, see Add certificates to gateways before upgrading API Connect.

    If you skip this task, the upgrade will be successful but you lose analytics event records that span the time when the management servers start at the upgraded level until each Gateway server is removed and readded after the upgrade.

    Attention: This is a one-time task and does not need to be repeated with subsequent upgrades.
  2. Download the appropriate firmware image from IBM Fix Central. See IBM Fix Central and Downloading the images to upgrade IBM API Connect servers.
    The Management server is a virtual appliance. The firmware image file name includes ManagementAppliance and the file extension is .vcrypt2.
    For example,
    version-APIConnect-ManagementAppliance-timestamp_H9_64-CUMUIFIX-003.vcrypt2
    Host the downloaded firmware on an FTP server that is accessible from the Management server that you want to upgrade. "Manually manipulating the load balancer is not necessary because upgraded servers are down until the upgrade process is complete"
  3. In your virtual machine hosting software, for all of your Management servers that you are upgrading, increase the disk size to 10 GB.
    A newly deployed IBM API Connect Version 5.0 appliance has a disk size of 10 GB. Upgrading an IBM API Management Version 4.0 appliance requires a disk size increase. The only disk that must be increased is the code disk. The new disk size comes into effect after the upgrade to IBM API Connect Version 5.0 is complete.

    [V5.0.1 or later]For more information, see Increasing code disk size for appliances

  4. Log on to the CLI on the Management server through a Secure Shell (SSH).
  5. Using the firmware image that is hosted on an FTP server, update the Management server by entering the following system update osfirmware command:
    Note: FTP is shown here as an example. If you prefer, you can select another protocol (SFTP, HTTP, or HTTPS) that is more convenient for your environment.
    Syntax : 
        	
        	system update osfirmware from ftp <hostname|ip> user <username> file <filename> 
    Where:
    • hostname|ip is the FTP server where the firmware image is stored
    • username is the user name that is used to log in to the FTP server
    • filename is the absolute path to the firmware image .vcrypt2 file that is stored on the FTP server
    When prompted, enter the password to connect to the FTP server.
    The firmware upgrade is applied to the Management server. After the successful completion of the firmware installation, the Management server remains inactive until the primary Management server in the service is upgraded. To monitor the upgrade status, use system show status.
  6. After the upgrade is complete, verify that the correct version is installed by logging in to the CLI and issuing the system show version command. Check that is shows the version 50811, and does not report any problems that are related to the upgrade.
    Note: The status of the Management server that you are upgrading shows as Inactive in the Cloud Manager until all of the other servers in the cloud are upgraded to the same version.
  7. Repeat the previous steps until every Management server has been successfully upgraded and you verify that all the Management servers in the Management service are running the same exact version, by using the CLI.
  8. After all of the Management servers are upgraded, verify that all of the servers resume normal operations by starting the Cloud Manager and completing the following steps:
    Note: It might take several minutes for all of the servers to come back online.
    1. Log in to the Cloud Manager of the upgraded Management server and select the Services page.
    2. Refresh the browser to ensure that the latest information about the services and servers is displayed.
    3. Click the Server Details icon (i) of each server to view the server details and verify that the upgraded version is installed.
    4. Verify that each Management server is in Active status.
      Note: After a management server upgrade, the Gateway servers are displayed as out of sync on the Services tab in the Cloud Manager. For more information, see Gateway resynchronization.

After maintenance to the Management server, or the Gateway server, each Gateway server in a service must be refreshed.

Removing the Gateway server from the Cloud Manager removes it from the load balancer group, or groups, and also ensures that when you re-add the Gateway server, it obtains the latest DataPower configuration.

Do not delete the Gateway services, instead, refresh (remove and add) the Gateway servers within them.

  1. Remove the Gateway server.
    When you remove a DataPower Gateway server, historical monitoring information that relates to CPU, memory and disk usage, and server load is lost. However, analytics data that relates to API usage is retained.
    1. If you have two or more Gateway servers and you are using an external load balancer, remove the Gateway server from the load balancer and allow time for any in-flight API transactions to complete.
    2. Remove the Gateway server by using the Cloud Manager, see Removing servers.
  2. Add the Gateway server to a service.
    1. Add the Gateway server by using the Cloud Manager. For more information, see Adding a Gateway server.
    2. If you have two or more Gateway servers and you are using an external load balancer, add the Gateway server to the load balancer again.
  3. Verify that the Configuration version displayed for each Gateway server matches the version of the Management servers, see Verifying maintenance level.
  4. Upgrade your Developer Portal appliance by following the procedure in Applying an IBM fix pack and upgrading all sites to use the new distribution.

Results

The upgrade from v50810-ifix2 to v50811 is successfully applied to the Management service of your cloud environment.