1. Create new environment by default
2. Open "Networks" tab
3. Add 5-6 new node network groups
4. Click "Cross" icon for last created node network group and click "Delete" button in confirmation dialogue
6. Group disappeared from UI
7. Move to last node group and delete it too
8. This group also disappeared from UI but there is error in Firebug Console (see screen)
9. Refresh page. Group from step 7 not deleted, it appears in UI
I reproduced this bug many times.
I think cause is we need some time to delete node group. And only after one node network group deleted, user can delete another node group.
VERSION:
feature_groups:
- mirantis
production: "docker"
release: "8.0"
openstack_version: "2015.1.0-8.0"
api: "1.0"
build_number: "197"
build_id: "197"
fuel-nailgun_sha: "8d6ef41f1bef84378a61ee98b46aeb2b925fd10f"
python-fuelclient_sha: "e685d68c1c0d0fa0491a250f07d9c3a8d0f9608c"
fuel-agent_sha: "1d98edb0468aa70b9b3a43b8422804e9095e7d9d"
fuel-nailgun-agent_sha: "b56f832abc18aee9a8c603fd6cc2055c5f4287bc"
astute_sha: "c8400f51b0b92254da206de55ef89d17fdf35393"
fuel-library_sha: "33c0fa3aada734dc9e6f315197ce0e4a16f5987c"
fuel-ostf_sha: "11afd5743a12b1006317d3ca7000d1ede77bdae2"
fuel-createmirror_sha: "994fed9b1ed889718b61a59733275c08c2dd4c64"
fuelmenu_sha: "d12061b1aee82f81b3d074de74ea27a6e962a686"
shotgun_sha: "c377d163519f6d10b69a654019d6086ba5f14edc"
network-checker_sha: "2c62cd52655ea6456ff6294fd63f18d6ea54fe38"
fuel-upgrade_sha: "1e894e26d4e1423a9b0d66abd6a79505f4175ff6"
fuelmain_sha: "22fe551f5525d11a1854fd87dbc8c77fae8fec08"
When we delete node group Nailgun triggers "update_dnsmasq" task to remove unneeded DHCP ranges from cobbler. This task consists of 3 "subtasks" so it can take few seconds. And if you try to remove another nodegroup while "update_dnsmasq" task is running, then you'll get "409 Client Error: Conflict (A task is already running)" response from API:
# fuel --env 2 nodegroup --delete --group 18
Node group with id=18 was deleted!
# fuel --env 2 nodegroup --delete --group 17
409 Client Error: Conflict (A task is already running)
It looks like this should be fixed on the UI side: UI should check response on delete nodegroup request and alert user in case of error reply from API.