Kernel / Ramdisk not set when registering nodes with the UI
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tripleo |
Fix Released
|
High
|
Jiri Tomasek |
Bug Description
Using Newton, UI version is 20160930145215.
{u'ipmi_address': u'192.168.101.60', u'ipmi_username': u'admin', u'ipmi_password': u'******'}
While when registering via the CLI using the same nodes.json file I get:
{u'ipmi_password': u'******', u'ipmi_address': u'192.168.101.60', u'deploy_ramdisk': u'49418005-
I think this is happening because the workflow doesn't set the kernel name and ramdisk name by default:
https:/
This was to allow the CLI import commands to still work with the '--no-deploy-image' parameter when using Mistral. The CLI otherwise hardcodes the kernel and ramdisk names:
https:/
https:/
I think the UI needs to do the same thing, probably around:
https:/
I suspect not having this set may cause issues down the line, and since we're trying to keep the UI and CLI in sync we probably want to make sure nodes are registered with the same defaults. (I can't test an actual deployment in that environment at the moment due to other issues I'm debugging - if this affects deployments the bug importance likely needs to be raised)
Changed in tripleo: | |
assignee: | nobody → Jiri Tomasek (jtomasek) |
milestone: | none → newton-rc3 |
Changed in tripleo: | |
importance: | Medium → High |
Fix proposed to branch: master /review. openstack. org/381855
Review: https:/