Announcing improvements to how NDS Labs Environments are launched and managed
Now easier than ever to start and use NDS Labs Environments!
I am happy to announce the completion of a series of changes that makes
it significantly easier for anyone to create and use their own NDS Labs
Environment. The new step-by-step instructions can be found here (and
copied inline below for your convenience):
https://bitbucket.org/nds-org/nds-labs/src/2428a7e3cee839afe50fa0aa4868471e459c2c15/LAUNCHING?at=dev
This information will be incorporated in a more formal way with the
existing documentation later, but I wanted to get this out to everyone
quickly to get some early feedback.
We are also working on a web interface that will further simplify the
process.
Please let me know of any questions, problems, concerns.
Thank you.
- Mike
***
This document describes the steps needed to start up a new NDS Labs
Environment.
NOTES
* The defaults are set to work on the NCSA VLAD OpenStack cluster.
* All of these steps should be performed from a shell on the
vlad-mgmt.ncsa.illinois.edu node. This is not a requirement,
but does make some things a bit more convenient.
PREREQUISITES
* OpenStack credentials (userid and password)
PROCEDURE
1. Generate new key pair (or use an existing one if you prefer)
ssh-keygen
2. Start an ssh agent and load the key
eval $(ssh-agent -s)
ssh-add
3. Clone the nds-labs source code repository
hg clone https://bitbucket.org/nds-org/nds-labs
cd nds-labs
4. Update OS_USERNAME in script NDS-openrc.sh
vim NDS-openrc.sh
5. Copy your public key created in Step 1 to ssh-key.pub
cp ~/.ssh/id_rsa.pub ssh-key.pub
This is the key that will be dropped into the authorized_keys files.
6. Review and update as desired the values in the script run_startup.sh
Note: The "--name" parameter should definitely be changed to something
uniquely identifiable, such as your username.
7. Execute the startup script
./run_startup.sh
You will be asked for your OpenStack password. This is your NCSA
Kerberos password if you are on the VLAD system.
8. Check for the existence of the fleetctl binary. If "which fleetctl"
comes up empty, just copy the binary (the single file) from any other
machine to _this_ machine.
scp wherever:/usr/bin/fleetctl ~/
alias fleetctl='~/fleetctl' <-- optional, but convenient
9. Edit setfleet.sh and set the FLEETCTL_TUNNEL to the public IP address
assigned by the startup script in Step 7.
vim setfleet.sh
10. Source setfleet.sh to set the FLEETCTL environment variables.
source setfleet.sh
11. Launch the master service. This starts various services inside the
NDS Labs Environment to bring it to a fully functional state.
fleetctl start master.service
fleetctl status master.service
12. Your NDS Labs Environment is now up and running.
You can ssh to any of your OpenStack instances:
fleetctl list-machines
fleetctl ssh <machine-id>
You can see the status of all the services in your environment:
fleetctl list-units
You can ssh into the master service:
ssh -p 80 root@xxxxxxxxxxxxxxx <-- the public IP address that was
assigned
You can start and stop additional services as desired, from either your
local machine or from the NDS Labs master service. For example,
fleetctl start rstudio.service
<eom>
Other Mailing lists |
Author Index |
Date Index |
Subject Index |
Thread Index