Moving images on storage nodes

[sonic136]

Good Afternoon Community

We have 2 nodes (a nas with slower HDDS and a node in a DC with SSDs)

We use the nas storage for building and testing images, after this are forced to re-deploy our golden image and then re-capture the image on Fog.

Is there a way to move the captured image from one storage node to another.

Both the nodes are set to master so im not sure if the replication would work (as read on a different thread).

Is there a feature in the console that would allow the moving of images from dev (nas) to production (DC storage)

[Sebastian Roth]

@sonic136 said in Moving images on storage nodes:

Both the nodes are set to master so im not sure if the replication would work (as read on a different thread).
Is there a feature in the console that would allow the moving of images from dev (nas) to production (DC storage)

In this scenario your only choise would be using external tools like rsync to replicate the image from dev to production. That shouldn’t be too hard to setup using a cron job.

Image definitions need to be created manually though. There might be a way to query the information and add to the other node using the FOG API but I don’t think it’s worth the time to implement this unless you are really keen.

[george1421]

@sonic136 There is no automated way to do this but it is possible.

Just so you know a FOG “image” is made of 2 parts.

1. The raw image files saved in /images/<image_name> directory
2. The metadata stored in the FOG server database.

For option 1 you can either use rsync/scp to copy the raw data files between the two servers. OR you can, on your development server, add the deployment server as a storage node in your development server’s storage group. (FWIW you can only have one master node in each storage group, but a master node in storage group A can be a slave node in storage group B). We can leverage this feature to allow the raw file replication between the development server and the deployment server. To control the file copy just be sure to disable replication in the development server’s image definition. Once replication is enabled the raw data files will be sync’d between the two servers.

For option 2 its less clear. The easiest solution is to just manually create the image definition on the deployment server to match the settings on the development server. The issue is that the metadata copy function doesn’t exist in the FOG replicator because in a normal master node, storage node setup there is only one database and that is on the root FOG server.

[sonic136]

@sebastian-roth Thank you. Will look into it!