[FUEL][7.0]Timeout of deployment is exceeded

[Samer Machara]

I'm trying to deploy the basic three nodes architecture to learn OpenStack: 1 controller and 2 compute nodes. After several hours of deployment, I got this error 'Timeout of deployment is exceeded.'. So I tried to redeploy it but without success.
I have installed Fuel 7.0 with launch_8GB.sh script, and I have 130GB HHD-free and my Download speed is 6.41 Mbit/s. It should work well.

To see what is happening, I'm tried to deploy one node at the time I start with the controller node. However, I still with the same problem 'Timeout of deployment is exceeded.'

Another thing, I remove a node by error, and now Fuel does not autodetect it. Even I tried to add another node by cloning a VM and changing its MAC, but Fuel does not detect it as well. All VMs are booting from the PXE.
There is a command to execute the autodetection?

        Here is the diagnostic Snapshot.

[Vladimir Kozhukalov (kozhukalov)]

Preliminary diagnosis: your test server is too slow.

There are plenty of messages like "warning: [579] TCP connection to AMQP failed: Could not establish TCP connection to 10.20.0.2:5672. Retry 5 sec later..." but at the same time Rabbitmq is running and sometimes astute is able to connect to it. Also slave nodes are often not available by RPC.

Could you please expose the system configuration of the hardware that you use for your lab?

[Vladimir Sharshov (vsharshov)]

I agree with Vladimir, deploy is going well without error, but get timeout error.

What can you do in this case?

- use more simple cluster configuration;
- check our new Fuel version - 8.0. From release to release we works hard to increase performance;
- improve your test hardware environment.

[Vladimir Sharshov (vsharshov)]

You can also try to use launch.sh instead of launch_8GB.sh. Also you can decrease number of creating slave nodes from 3 to 2 by editing config.sh. Please find cluster_size in else section and change it:

if [ "$CONFIG_FOR" = "16GB" ]; then
  cluster_size=5
elif [ "$CONFIG_FOR" = "8GB" ]; then
  cluster_size=3
else
  # Section for custom configuration
  cluster_size=3
fi

and change it

if [ "$CONFIG_FOR" = "16GB" ]; then
  cluster_size=5
elif [ "$CONFIG_FOR" = "8GB" ]; then
  cluster_size=3
else
  # Section for custom configuration
  cluster_size=2
fi

After it just use launch.sh

[Samer Machara (samer-machara)]

 Last week I had a "time out error", deploying only one controller node on one virtual machine, searching for the error I found that my hard disk doesn't have enough space (110GB), So I mounted another HHD linked VBox folder to this new disk, then I re-started with a clean installation of Fuel.
  Now, I found a different error "Deployment has failed. Method granular_deploy. Deployment failed on nodes 1. Inspect Astute logs for the details."

Here is the "Diagnostic Snapshot"

https://bugs.launchpad.net/fuel/+bug/1565672

[Oleksiy Molchanov (omolchanov)]

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