Sandeep Kumar created AMBARI-26644:
--------------------------------------
Summary: React UI bug fixes in cluster wizard, config
recommendation completeness, alerts/sidebar refresh, host
maintenance/decommission handling, and Observer NameNode HA transitions.
Key: AMBARI-26644
URL: https://issues.apache.org/jira/browse/AMBARI-26644
Project: Ambari
Issue Type: Story
Reporter: Sandeep Kumar
Assignee: Sandeep Kumar
Fixes several issues across the cluster install wizard, host/service
configuration, alerts, and host management screens.
- ClusterWizard: allow hyphens and special characters in DB credential
usernames; fix repo lookup keying off the selected stack's id instead of
the selected version's id, which broke the baseurl lookup during manual
installs; complete configuration recommendation properties so
dependent-config recommendations apply consistently across the
cluster-install wizard, the Add-Service wizard, and in-place service
config saves.
- Alerts/Sidebar: align alert data refresh with the expected polling
pattern so the sidebar and alert screens react immediately to socket
events instead of lagging behind the periodic poll cycle; fix the
sidebar restart-required icon and general responsiveness after a config
save.
- Stack/Versions: fix the multi-patch upgrade UI showing the first patch's
info while a second patch is actually in progress, by tracking the
upgrade's associated version explicitly instead of matching on display
name.
- Configs: render HOSTS-type config properties (e.g. DataNode/NameNode
hosts) as a clickable link/modal instead of plain text; stop the "add to
config group" modal from snapping back to the first group when
overriding a default config; restore the dfs.journalnode.edits.dir
initializer, which had been commented out and was causing NameNode
startup failures during HA setup.
- Observer NameNode: add the missing standby<->observer reverse
transition (only one direction existed) and switch off the stale
MAKEOBSERVER custom command in favor of TRANSITION_NAMENODE with an
explicit target_ha_state.
- Hosts: fix the DataNode status icon not updating after
recommission/decommission without a manual refresh, by reconciling live
NameNode LiveNodes metrics with desired_admin_state and removing a
stale-closure bug in the polling loop; fix Start/Stop/Restart being
unable to run after enabling host Maintenance Mode, caused by
cloneDeep() stripping the Host/HostComponent class prototypes needed by
isActive(); fix bulk Start/Stop from the Hosts list triggering unwanted
Install Client tasks, caused by a forEach with an inner return (silently
discarded) that always fell through to a hardcoded false, plus reading
the wrong response field for stack component data.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
