Ask Your Question
2

Sahara cluster stuck in waiting state

asked 2014-11-12 04:01:22 -0500

i_like_pie gravatar image

Hello,

Whenever I try to launch a Sahara cluster it gets stuck in the waiting state.

The installation is the following:

  • 2 nodes: compute and controller
  • Nova networking
  • Used image - Sahara-vanilla-plugin-1.2.1-Ubuntu-13.10

The cluster is starts and creates one instance. After some time the instance goes into active state. There is connectivity between the controller and the instance (can manually ssh to it). However, it seems that Sahara does not see that the node is in active state.

I can see the following debug messages being spammed ("controller" is the hostname if the controller node). There are no warning/error messages at all:

2014-11-12 11:36:49.753 3364 INFO urllib3.connectionpool [-] Starting new HTTP connection (1): controller
2014-11-12 11:36:49.863 3364 DEBUG urllib3.connectionpool [-] "GET /v2/86125242c42e4bd5a3018b572a847c9f/servers/f3ba0692-58c0-4dc6-8e49-62753841b1c5 HTTP/1.1" 200 1711 _make_request /usr/lib/python2.7/site-packages/urllib3/connectionpool.py:295
2014-11-12 11:36:50.912 3364 INFO urllib3.connectionpool [-] Starting new HTTP connection (1): controller
2014-11-12 11:36:51.008 3364 DEBUG urllib3.connectionpool [-] "GET /v2/86125242c42e4bd5a3018b572a847c9f/servers/f3ba0692-58c0-4dc6-8e49-62753841b1c5 HTTP/1.1" 200 1711 _make_request /usr/lib/python2.7/site-packages/urllib3/connectionpool.py:295
edit retag flag offensive close merge delete

4 answers

Sort by ยป oldest newest most voted
3

answered 2014-12-08 04:39:15 -0500

i_like_pie gravatar image

The issue was that floating IP's were disasbled under nova but enabled in sahara configuration. This made Sahara get stuck in an infinite loop waiting for those addresses. While this was an issue on our side it would have been nice to see some debug messages from sahara, as finding the cause of the problem involved digging in the source code quite a bit.

edit flag offensive delete link more

Comments

thanks for the update, i wonder if there is a way for Sahara to determine if Nova has floating ips available?

elmiko gravatar imageelmiko ( 2014-12-10 09:07:42 -0500 )edit
0

answered 2015-04-03 04:09:15 -0500

BiskrianO gravatar image

how do you resolved it !!!

edit flag offensive delete link more

Comments

Check the selected answer, it was a configuration conflict issue.

i_like_pie gravatar imagei_like_pie ( 2015-04-05 13:18:27 -0500 )edit

can you help me in my ( https://ask.openstack.org/en/question/64481/sahara-cluster-launching-stay-on-waiting/ (https://ask.openstack.org/en/question...) )

BiskrianO gravatar imageBiskrianO ( 2015-04-06 05:45:38 -0500 )edit
0

answered 2015-09-04 11:48:22 -0500

saurav_purdue gravatar image

I am facing the exact same issue but I am using "Neutron" network, please let me know if any ideas.

Thanks in advance.

edit flag offensive delete link more
0

answered 2014-11-16 00:35:50 -0500

9lives gravatar image

=write here to override the comments charater limit=

First check if any proxy used , second check the status of sahara cluster instance using sahara client sahara cluster-show cluster_id makes sure you can see the instances in cluster and can see all the instances in project.

Cluster is in waiting state probably means the sahara is setting up the hadoop cluster via ssh command and it is stuck by some reason, check the sahara logs for details.

Hope that helps!

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2014-11-12 04:01:22 -0500

Seen: 1,296 times

Last updated: Sep 04 '15