Installing CTERA Portal Instances

  • 4 minute(s) read

Use the following workflow to install CTERA Portal.

  1. Creating a Portal Instance using a CTERA Portal image obtainable from CTERA support.
  2. Configuring Network Settings.
  3. Optionally, configure a default gateway
  4. Additional Installation Instructions for Customers Without Internet Access.
  5. For the first server you install, follow the steps in Configuring the Primary Server.
  6. For any additional servers beside the primary server, install the server as described below and configure it as an additional server, as described in Installing and Configuring Additional CTERA Portal Servers.
  7. Make sure that you replicate the database, as described in Configuring the CTERA Portal Database for Backup.
  8. Backup the server as described in Backing Up the CTERA Portal Servers and Storage.
Note

You can use block-storage-level snapshots for backup, but snapshots are periodical in nature, configured to run every few hours. Therefore, you cannot recover the metadata to any point-in-time, and can lose a significant amount of data on failure. Also, many storage systems do not support block-level snapshots and replication, or do not do so efficiently.

Creating a Portal Instance

You can install the CTERA Portal on KVM via the OpenStack console or another console. The following instructions describe how to install a CTERA Portal directly in KVM.

To install the CTERA Portal Server:

  1. Log in to the KVM machine.
  2. Copy the KVM image file, Portal_Image_<version>.qcow2 to a folder on the KVM machine where version is the CTERA Portal version.
  3. In the folder where you copied the KVM image file, create the data disk for the portal server using the following command: qemu-img create -f qcow2 data.qcow2 <diskSize> where diskSize is the size of the data disk, based on the server requirements. For details, see Requirements.
  4. Change the permissions for the data disk using the following command: chmod 777
  5. Create the portal server using the following command:
    virt-install --name <name> --vcpus <cpus> --memory <memory> --disk path=Portal_Image_<version>.qcow2 --disk path=data.qcow2 --disk path=archive.qcow2 --network bridge=br0 --os-variant centos-stream9 --import --noautoconsole
    where:
    name is the name assigned to the portal server VM.
    cpu is the number of vCPUs as specified in Requirements.
    memory is the memory requirement in MB as specified in Requirements.
    version is the CTERA Portal version.

    The server is created and started up.
    Note

    If the script to create the data pool does not successfully run, it will start on every boot until it completes. The script has a timeout which means it will exit if the data pool is not created within the timeout after boot time. If the data pool is not created, dockers required by the portal are not loaded to the data pool.


    To make sure that the script completed successfully, before continuing, run docker images to check that the docker images are available, including zookeeper, which is the last docker to load to the data pool.


    If all the dockers do not load you need to run the script /usr/bin/ctera_firstboot.sh


    Also, refer to Troubleshooting the Installation if the script does not complete successfully.

  6. Continue with Logging in to the Server to Create the Storage.
  7. Start CTERA Portal services, by running the following command: portal-manage.sh start
Note

Do not start the portal until both the sdconv and envoy dockers have been loaded to the data pool. You can check that these dockers have loaded in /var/log/ctera_firstboot.log or by running docker images

Logging in to the Server to Create the Archive Storage

You need a data pool on every server. The data pool is created automatically when you first start the virtual machine and dockers loaded on to it.

You need to create the archive pool for the primary database server and when PostgreSQL streaming replication is required, for the replication database server. See Using PostgreSQL Streaming Replication for details about PostgreSQL streaming replication.

Access the new portal machine, using SSH.

To log in to the portal server:

  • Log in as the root user.
    The default password is ctera321

You are prompted to change the password on your first login.

To create the archive pool:

  1. Log in as the root user using SSH.
  2. Add a second disk to use as the archive pool using the following command: qemu-img create -f qcow2 archive.qcow2 <diskSize> where diskSize is the size of the data disk, based on the server requirements.
    Note

    The minimum archive pool should be 200GB but it should be sized around 2% of the expected global file system size. For more details, see Requirements.

  3. Change the permissions for the archive disk using the following command: chmod 777
  4. Attach the disk to the portal server using the following command: virsh attach-disk <vm_name> /<path>/archive.qcow2 vdc --targetbus virtio --persistent --driver qemu --subdriver qcow2
    where vm_name is the name of the portal server specified in the virt-install command when you created the virtual machine and path is the full path to the added disk.
  5. Run lsblk or fdisk -l to identify the disk to use for the archive pool.
  6. Run the following command to create the archive pool: portal-storage-util.sh create_db_archive_pool Device
    Where Device is the device name of the disk to use for the archive pool.

Troubleshooting the Installation

You can check on the progress of the docker loads in one of the following ways to ensure that all the dockers are loaded: The last docker to load is called zookeeper:

  • In /var/log/ctera_firstboot.log
  • By running docker images to check that the docker images are available.
  • By checking if /var/log/ctera_firstboot_completed is present with the date and time when the installation was performed.

If all the dockers do not load you need to run the script /usr/bin/ctera_firstboot.sh