"No nodes removed" after successful node deletion

[Vladimir Khlyunev]

ISO #262

Steps:
1) Build & install detach-service plugins:
https://github.com/stackforge/fuel-plugin-detach-database
https://github.com/stackforge/fuel-plugin-detach-rabbitmq
https://github.com/stackforge/fuel-plugin-detach-keystone
2) Create cluster:
Ubuntu, Ceph
3) Add:
3 controllers
2 computes
2 ceph
3 database+keystone+rabbitmq (I assume that 1 node with 1 role will be enough)
4) Deploy cluster
5) After deployment mark 1 database+keystone+rabbitmq node for deletion and deploy changes

Result: node successfully removed, but "No nodes were removed." is shown in dashboard tab:

Snapshot is still generating, I will add it as soon as it will be available (also I can provide my env)
UPD: snapshot https://drive.google.com/a/mirantis.com/file/d/0B_bjwT_xlxy1NERYak43Y0JDRHM/view?usp=sharing

[Vladimir Khlyunev (vkhlyunev)]

wrong assignment

[Vladimir Khlyunev (vkhlyunev)]

Deleted node can not get IP => can not boot via pxe.

[Vladimir Khlyunev (vkhlyunev)]

snapshot https://drive.google.com/a/mirantis.com/file/d/0B_bjwT_xlxy1NERYak43Y0JDRHM/view?usp=sharing

[Dima Shulyak (dshulyak)]

I see that machine with mac 64:fd:2d:37:89:ef eventually received ipaddress and registered in nailgun with new id which is 15, and it is currently online and i am able to use it for provisioning

>> fuel node
15 | discover | Untitled (89:ef) | None | 10.109.0.17 | 64:fd:2d:37:89:ef

Additionally, astute finished removal of the node at

   2015-09-02T11:47:49 info: [678] 16308432-8231-4203-9053-ca888137484b: Finished removing of nodes: ["9"]

And node with mac 64:fd:2d:37:89:ef talks to dnsmasq and according to dnsmasq it received ip short after removal (/var/log/dnsmasq.conf)

Sep 2 11:47:56 dnsmasq-dhcp[3139]: DHCPDISCOVER(eth0) 64:fd:2d:37:89:ef
Sep 2 11:47:56 dnsmasq-dhcp[3139]: DHCPOFFER(eth0) 10.109.0.17 64:fd:2d:37:89:ef
Sep 2 11:48:03 dnsmasq-dhcp[3139]: DHCPREQUEST(eth0) 10.109.0.17 64:fd:2d:37:89:ef
Sep 2 11:48:03 dnsmasq-dhcp[3139]: DHCPACK(eth0) 10.109.0.17 64:fd:2d:37:89:ef

So, i think that this error in comment #4 is sporadic and wont be reproducible, it leads to an issue [0], which says that such problem can appear due to DHCP timeout, and in [1] there is several suggestion how to avoid it when you are using VMs. So, please consider taking a look at [1] and maybe it makes sense to use this configuration in fuel-devops.

As for the bug with message - astute returned
   "nodes"=>[], "inaccessible_nodes"=>[{"uid"=>"9", "error"=>"Node not answered by RPC."}]
And in [2] you can see that we arent taking into account *inaccessible_nodes* array, this needs to be fixed and message should take this array into account

0. http://ipxe.org/err/040ee1
1. http://ipxe.org/err/4c1060
2. https://github.com/stackforge/fuel-web/blob/master/nailgun/nailgun/rpc/receiver.py#L129-L135

It doesnt look like this bug is critical, because the only issue is that message is not relevant in the specific case, so my suggestion is to mark it as Medium and move to 8.0

[Dmitry Pyzhov (dpyzhov)]

Looks like bugs #1491373 and #1527571 are duplicates, we should check it

[Bug Checker Bot (bug-checker)]

(This check performed automatically)
Please, make sure that bug description contains the following sections filled in with the appropriate data related to the bug you are describing:

actual result

expected result

For more detailed information on the contents of each of the listed sections see https://wiki.openstack.org/wiki/Fuel/How_to_contribute#Here_is_how_you_file_a_bug

[Ivan Ponomarev (ivanzipfer)]

I can't reproduce this bug. The messages that I've got from dashboard was correct:
Plugin detach-database is deployed. Creates separate Database role for deploying
Plugin detach-keystone is deployed. Creates separate Keystone role for deploying
Plugin detach-rabbitmq is deployed. Creates separate RabbitMQ role for deploying
Successfully removed 1 node(s). No errors occurred

[Oleksiy Molchanov (omolchanov)]

Marking as Invalid, because of no update for more than a month