Describe the issue: Creating BZ tracker for 5x user story - Disconnected installation
We need to ask users to provide ip address of the host to be added along with hostname if host dnf is not reachable over private IP (In regular case where nodes are reachable over public IP with dns reach,just host name is sufficient, example - ceph orch host add <hostname> <host address> ceph orch host add pluto009 172.20.20.9
Typically there is a proxy host that has access to our registry. So a user would typically install a local container registry on that proxy host. Otherwise someone would need to physically copy the images to the local registry, but that should not differ to RHCS 4
Hi Karen, I've shared my thoughts offline (Mail - Sub: "5.0 User story : Disconnected installation : Brainstorm thread") I'm moving back the BZ to ASSIGNED state. Please let me know if there are any concerns. Regards, Vasishta shastry QE, ceph
Hi Karen, in 3.5.1. Performing a disconnected installation 1) We have mentioned >> A disconnected installation uses custom images and configuration files and local hosts, instead of downloading files from the network. It could be changed as >> A disconnected installation uses custom container registry. As i) All users (connected installation) use custom spec files (aka config file). ii) I'm not sure which files are to be downloaded. If we are talking about 'packages', I'm not sure how we recommend users to install packages (satellite of iso) 2) Step 1 can be removed We are asking ame thing to be done in step 4 - Adding hosts 3) As I had requested in mail thread, If user doesn't provide this argument with bootstrap (i,e step 3) >> --image <custom_container_registry_name_with_port>:<custom-image-name>:<image-tag>' the default image with be considered for which nodes will not be having access to ------------------------------------------------------------------------- in 3.5.2. Creating custom container images for disconnected installations 1) The procedure of this chapter is explaining "Changing configs of monitoring daemon container images" Steps are not to create, but to change configs. 2) We need to ask users to follow this section after the bootstrap in previous section 3) Step 1 and 2 are redundant and I think they need to be be removed. 4) Since https://bugzilla.redhat.com/show_bug.cgi?id=1935044 is targetted to 5.1 we need to ask users to redeploy alertmanager as mentioned in https://bugzilla.redhat.com/show_bug.cgi?id=1935044#c4 ------------------------------------------------------------------------- As mentioned in mail thread and given references to 4.x installation guide, In 4.x installation guide we have mentioned steps to configure custom/imternal container registry which will be accessible for disconnected nodes. I'm not sure whether we would like to provide same steps in 5.x. If we are not providing the steps to configure the internal/custom registry, at least we should add a point in prerequisites in 3.5.1 saying `as nodes might not have access to redhat network, please be ready with container images of ceph and monitoring daemons uploaded to your local/custom container registry. Please feel free to ask me anything or to let me know if there are any concerns. Regards, Vasishta shastry QE, Ceph