TM1 Forum

Hello Community,

i am searching "RedBook" or "BestPractise" or any deepdive-samples how we can use GIT (BitBucket) and TM1 (Planning Analystics)
Is there any experiance around?

Chris

You could do a lot worse than start here: https://ykud.com/blog/cognos/tm1-cognos ... tm1-part-1

After several testing with Bitbucket, I don't think PA is mature enough to integrated with Git at this stage (We are on PA 2.0.8).
With those documents from IBM, you will find yourself struggling with the syntax of tm1project file eventually. Yes, it can do a few things with Git but very hard to use.

macsir wrote: ↑Tue May 12, 2020 10:26 pm With those documents from IBM, you will find yourself struggling with the syntax of tm1project file eventually. Yes, it can do a few things with Git but very hard to use.

Do you mind sharing what in tm1project file was confusing?

I agree it's very far from intuitive, it's more of an API to build future tools on, but I still think it's quite useful even in the current shape.
Maybe it's the case when of having more examples out there might help, especially for tm1project?

Y

Hi, ykud
Yes, I agree with the point that it is an API which some tools can build on. But the questions I have so far are
1. How to use Objects section in tm1project file rather than use Ignore section all the time to include/exclude objects?
2. How to use Pre/PostPull, Pre/PostPush, Dependencies/Preconditions and Links? The examples in IBM document are very raw and hard to understand.
3. Is it possible to hot promote cube data files via this method? I can push/pull the data files but looks like I have to reboot the tm1 server to get data file loaded.

Overall, I don't see any detailed examples from IBM apart from some detailed articles from other developers.

macsir wrote: ↑Wed May 13, 2020 4:49 am 3. Is it possible to hot promote cube data files via this method? I can push/pull the data files but looks like I have to reboot the tm1 server to get data file loaded.

Overall, I don't see any detailed examples from IBM apart from some detailed articles from other developers.

I agree with you that the documentation is very lacking and those examples leave you more confused than they explain how to use it.

And answering #3 the official answer from IBM support is NO, manually including the .cub files in the "files" section in tm1project and then doing a server reboot after a GitPull is not recomended by IBM, that's what they told me when I opened a support ticket. Also If we have .cub files over 100MB it won't upload to free git servers like github for example.
So we have to create a process that loads the attribute values from a file somewhere. One more thing to version and ensure that the file is updated with the latest values. Very disappointing.

Also, github is deactivating the basic auth that the tm1git uses to connect to their server in august 2021. Now they'll only accept tokens or SSH Key authentication, but tm1git doesn't have support for that!

luizg2019 wrote: ↑Fri May 21, 2021 1:40 pm Also, github is deactivating the basic auth that the tm1git uses to connect to their server in august 2021. Now they'll only accept tokens or SSH Key authentication, but tm1git doesn't have support for that!

SSH is supported : https://www.ibm.com/docs/en/planning-an ... ntegration

ykud wrote: ↑Fri May 21, 2021 7:37 pm

luizg2019 wrote: ↑Fri May 21, 2021 1:40 pm Also, github is deactivating the basic auth that the tm1git uses to connect to their server in august 2021. Now they'll only accept tokens or SSH Key authentication, but tm1git doesn't have support for that!

SSH is supported : https://www.ibm.com/docs/en/planning-an ... ntegration

oops my bad

Thank you!

Hi,
did you manage to connect to Github via SSH?

I get response:
"error": {
"code": "278",
"message": "Invalid parameter in payload \"PublicKey\". GitInit failed."
}
And I followed instructions from IBM pages:

{
"URL":"git@github.com:username/repository.git",
"Deployment":"main",
"PublicKey" : "ssh-ed25519 xxx",
"PrivateKey" : "-----BEGIN OPENSSH PRIVATE KEY-----xxx-----END OPENSSH PRIVATE KEY-----",
"Passphrase" : ""
}

marko_p wrote: ↑Tue Oct 05, 2021 9:55 am Hi,
did you manage to connect to Github via SSH?

Yep, works for me, but you need the 2.9.0.4 or later ( https://www.ibm.com/docs/en/planning-an ... er-17-2020 )

Works for us as well. Unfortunately it is only possible to use ssh for all users or for none. That is a bit annoying.

I've been messing around with GitHub and PA Cloud and it seems to be working well. I did a video to help share what I did to make it work.

https://www.youtube.com/watch?v=9XT9ALcKf4k

Hi, folks!

I'm diving into tm1 git integration and implementation of ci cd concept to tm1 development.
Just have read discussions here, but they seems to be a bit outdated. Since then some things happened, e.g. PAW 2.0.92 released.
I want to revive this discussion and aggregate here more useful links and tm1git related info.

I've got familiar with materials from links below and tested it on my environment (PAW 2.0.92 on linux, PA 2.0.9.19 on windows server 2019).
And want to say big THANKS to all who made this instructions and demos!

Yuri Kudryavcev's topics on doing this via REST API calls:
https://www.ykud.com/blog/cognos/tm1-co ... m1-part-1/
https://www.ykud.com/blog/cognos/tm1-co ... m1-part-2/
https://www.ykud.com/blog/tm1-git-challenges/

Bob Gillvan's videos on REST API and new PAW feature:
https://www.youtube.com/@robsonmedia9499

Stuart King's explanation of pre/post push/pull feature and its usage.
https://www.youtube.com/watch?v=TjDKiPS ... x=2&t=316s

Cubewise technote about TM1Project and overall approach:
https://cubewise-code.github.io/tm1py-t ... oject.html

But for me there is still no clear and precise picture how to manage branches and organise continuous ci cd.

It would be great if you share your experience of using git with tm1, your approaches to ci cd workflow, branches managing, prod deployment, etc...

Examples of your own git implementations (without REST API Git feature) would be great too!

Or everybody just give up this and use stop\copy\start approach or hot promotion via reading the code & patching it?