Description of problem: Currently it is not possible to sync all the hosts which are out of sync with a single click. Version-Release number of selected component (if applicable): How reproducible: 100% Steps to Reproduce: 1. 2. 3. Actual results: Expected results: After implementation, a button with 'Sync all hosts' will enable this action Additional info:
A simple Ansible playbook can be used perhaps?
Eitan, please fill in the Doc Text with a draft documentation. requires_doc_text should be set to + only by the Docs team, after they have properly addressed our require_doc_text? request. Specifically, I think that we should modify https://access.redhat.com/documentation/en-us/red_hat_virtualization/4.2/html/administration_guide/chap-logical_networks#Editing_a_Logical_Network to mention your new button.
Can also be tested with any nightly build virt-engine-0:4.2.6.3_SNAPSHOT-*
Verified on - Few notes: 1. It is still missing indication in the UI for the cluster level that networks are out of sync in the cluster. The only way to know that networks are out-of-sync are on the host level. BZ 1613702 should provide some indication(a warning icon next to the cluster) that networks are out-of-sync in the cluster. There is column for it in the Clusters main tab. 2. "Sync all hosts in cluster" invokes the pre-existing "sync all host networks" operation per each host in the cluster. "sync all host networks" in turn invokes "setup networks" without checking if the host is in\out of sync. This means that even for hosts that are in good state in the cluster we will see 2 audit logs about changes applied on the host and changes succeeded on the host, even if the state on the host is good and all networks are in sync. "Sync all hosts in cluster" only writes a single additional audit log message that the operation has started ("Started sync of all cluster ${ClusterName} networks.") Followed by "Managed to sync all host $[{hostname} networks." when done.
Verified on - 4.2.6.3_SNAPSHOT-94.gbbcd5cb.0.scratch.master.el7ev Few notes: 1. It is still missing indication in the UI for the cluster level that networks are out of sync in the cluster. The only way to know that networks are out-of-sync are on the host level. Or inside the cluster entity in the logical networks sub tab, the 'sync all networks' button will be enabled. BZ 1613702 should provide some indication(a warning icon next to the cluster) that networks are out-of-sync in the cluster. There is column for it in the Clusters main tab. 2. "Sync all hosts in cluster" invokes the pre-existing "sync all host networks" operation per each host in the cluster. "sync all host networks" in turn invokes "setup networks" without checking if the host is in\out of sync. This means that even for hosts that are in good state in the cluster we will see 2 audit logs about changes applied on the host and changes succeeded on the host, even if the state on the host is good and all networks are in sync. "Sync all hosts in cluster" only writes a single additional audit log message that the operation has started ("Started sync of all cluster ${ClusterName} networks.") Followed by "Managed to sync all host $[{hostname} networks." when done. 3. If there is a problematic host in the cluster while invoking sync all host in cluster, there will be 2 audit messages that changes has failed to applied and sync has failed.
Worth to mention that indication for problematic out of sync hosts are available in the events log when the host is active and the events drawer is open.
Eitan/Michael copying over comments from Jira: "the indications will be as follows (pending final approval from UX team): cluster hosts list view: textual indication in host status column hosts list view: textual indication in host status column host network interfaces view: graphical + textual indication on interfaces network hosts list view: graphical + tooltip indication in 'out-of-sync' column" Are these UI changes part of a separate bug? This bug has already been verified so how are there still issues that need to be approved?
Dan, can you answer Emma?
Eitan, you are the one who listed these indications; you are in best position to answer if they are included in this current code. I don't understand what you need from me. To the best of my knowledge the only change included with this bug is the added button, which becomes enabled when the cluster is somehow out-of-sync. If you think that something should be added, you should file a new bug, much like what Burman did with bug 1613702.
I do not know why something has automatically set requires_doc_text to plus. I believe it should happen only after an intervention of a human technical writer. Can you tell where is the bug, Derek? Red Hat Bugzilla 2018-07-25 14:03:33 IDT Flags: requires_doc_text? → requires_doc_text+
Emma, would you also mention in the docs that this new capability will also be available via the REST API? Thanks