Creating an NFS Mount

For backup and restore it is recommended that you set up an NFS mount to a remote server for the purpose of storing backups safely on a remote system should the accelerator suffer a critical failure.

In addition, replication may require a NFS mount to the source database for some replication strategies.

How to Create an NFS mount to a Remote NFS Server on the Accelerator

VM or Physical Accelerator

This section outlines the steps to add an NFS mount to the Hubble Accelerator. In this example the NFS mount is intended to be used to access Archive redo logs from Qlik using the Direct access to logs option. The Hubble Tech Mirroring Overview details the replication options document.

Open /config/config/config.xml nano /config/config/config.xml

Replace the values in{} and paste the code block right after </root-password> or right after </network> if </root-password> does not exist.

<remote-files><nfs name="{local_folder_name}">
<host>{remote_host_ip}</host>
<path>{remote_folder_path}</path>
</nfs>
</remote-files>
For example:
<remote-files>
<nfs name="qlik logs">
<host>l0.13.0.16</host>
<path>/logs</path>
</nfs>
</remote-files>

Save the file, exit the editor.
To create the NFS mount you can either reload the autofs command or reboot the server.
```
/etc/init.d/S60autofs  reload
```
After the reload command has been run or the reboot of the server, there should be a folder, with the same name given in "" in the config, under /import.

Given the example, there should be a folder called qlik_logs and it should mirror whatever is in /logs in the remote host.
```
/import/qlik_logs
```
There will be different permissions depending on how the shared folder is set up, but root should be able to write in there.

Dockerized Accelerator

This section outlines the steps to add an NFS mount to a Hubble Accelerator running in a Docker container. In this example we assume that you have already set up a NFS server and you want to configure your dockerized accelerator to be able to view and read files from the NFS server. After these steps are completed, the dockerized accelerator will have access to the files that are in the export directory on the NFS server. For this guide we will assume you are sharing the /export/accelerator directory. The files in that directory will appear in the containerized accelerator, in the sub-directory of / imports defined by ACCL YB NFS NAME 1, for this guide we will assume /imports/nfsl.

On the NFS Server

Edit the /etc/exports file to add an additional line with the docker accelerator host IP.

If your dockerized accelerator is deployed in AWS then you are using the docker network. Add the following line:

/export/accelerator your_host ip(rw,sync,no subtree check)

Then restart the NFS server to apply the new configuration:

systemctl restart nfs-kernel-server

On the Docker Accelerator Host

On the accelerator host edit the / etc/hubble/ accelerator/. env and enter the following details:

ACCL YB IDENTITY=hubble-docker-accelerator

ACCL YB NFS NAME l=nfsl

ACCL YB NFS HOST l=your nfs server ip

ACCL YB_NFS PATH_l=/export/accelerator

ACCL_YB_NFS_NAME_lwill show in the accelerator as a sub-directory of /imports if the NFS is working correctly.
ACCL_YB_NFS_HOST_l is the NFS server IP address.
ACCL_YB_NFS_PATH_l is the path on the NFS server where the files will be copied from.

Restart the docker container, making sure that replication has been stopped first. On the host run the following commands:

cd /etc/hubble/accelerator

docker-compose -p accelerator down && docker-compose -p accelerator up -d

Now check to see if the files on the NFS server appear in your Accelerator Docker container:

docker exec -ti accelerator hubble accelerator 1 sh -1 -c "ls /import/ nfsl"