cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
kdoherty
Community Participant

Question about Canvas Data CLI implementation

Jump to solution

Hello data team.  We plan to implement Canvas Data CLI on an Enterprise Linux V6 server.  Our systems people have a couple of questions that I did not find answer to on the community site.  I hope you can help.    (

1) Node.js:  CLI documentation says "any version newer than 0.12.0 should work".  Node.js appears to have several version branches with different numbering systems (0.10 0.12, 4.x, 5.x, 6.x). Should we install the latest release within the 0.12 branch?  Or should we install some other version branch with a higher number? 

(2) Can we use Canvas Data CLI through a proxy server?  Our servers are not allowed to directly access the internet. 

We appreciate any guidance that you can provide.  Thanks.

1 Solution

Accepted Solutions
ccoan
Instructure
Instructure

Hey Kevin,

I'd be happy to answer your questions.

1. We recommend using anything above the 0.12 branch. Although it should work with anything newer than 0.12 we do our automated testing on 4.x, 5.x, and 6.x. (We may eventually even bump up to 6.x requirement once it becomes more widely used, and easier to get). So any 4.x, 5.x, and 6.x version is what we highly recommend.

2. The CLI doesn't have proxy functionality built in. For now our current solutions have been two fold.

     A. Download on a separate box, and then transfer.
      B. Use a transparent proxy setup (such as Squid on linux).

I'll bring this up with the CLI team though, and make sure we have a conversation on the proxy functionality.

View solution in original post

5 Replies
ccoan
Instructure
Instructure

Hey Kevin,

I'd be happy to answer your questions.

1. We recommend using anything above the 0.12 branch. Although it should work with anything newer than 0.12 we do our automated testing on 4.x, 5.x, and 6.x. (We may eventually even bump up to 6.x requirement once it becomes more widely used, and easier to get). So any 4.x, 5.x, and 6.x version is what we highly recommend.

2. The CLI doesn't have proxy functionality built in. For now our current solutions have been two fold.

     A. Download on a separate box, and then transfer.
      B. Use a transparent proxy setup (such as Squid on linux).

I'll bring this up with the CLI team though, and make sure we have a conversation on the proxy functionality.

kdoherty
Community Participant

Thank you Eric.   Can you answer a follow-up?

We do have squid proxy, can you pose another question on how the CLI will access it? Such as, will setting the system environment variable for http_proxy be adequate enough?

Hey Kevin,

You'd want to set squid proxy up as a transparent proxy currently. Which is basically a way of performing a proxy without the application knowing it. (Basically set up IP Tables to direct HTTPS traffice to SQUID). There's a few tutorials out there if you search for "Squid Transparent Proxy".

However after talking with the team we will be implementing the HTTP Proxy feature so it should appear as just a config option. I'll try to remember to update this thread when the feature gets published, but you can also watch the repo for yourself: instructure/canvas-data-cli​.

Hey Kevin,

The change for https proxies has now been published. You can update your CLI, and use the configuration options `httpsProxy`, `proxyUsername`, and `proxyPassword` in order to connect to your proxy. Any version greater than 0.5 should support this.

Thanks

kdoherty
Community Participant

Thank you very much for the update Eric.

Kevin Doherty