Bedrock on PAaaS

[markhackett]

Hi.

I'm currently using a test instance of PAaaS with PAW 2.0.95. I haven't used Bedrock before, but it looks like it would be very useful for a task that I am currently working on, and I am keen to get more familiar with Bedrock anyway - I hear a lot of people singing it's praises, so it's probably about time I gave it a proper look.

Anyway, my issue is this: I can't seem to figure out how to get Bedrock to work on PAaaS. Maybe I'm overlooking something obvious, so hoping someone can point me in the right direction.

Any content that I have in PAaaS is either:

1) As a result of the migration process from my old on-premise server.
2) Developed directly on the PAaaS instance.

I have tried uploading some of the Bedrock .pro files into PAaaS using File manager - however, this doesn't appear to work (even after restarting the instance); from what I can judge, this facility is really just intended for loading data files rather than TM1 objects.

So I'm guessing that the correct way to do this is by using Git integration, but while I can successfully use Git to transfer objects in and out of my PAaaS instance, the format of the objects is different to what's in the Bedrock Git folders, i.e. Bedrock has "regular" tm1 file extensions like ".pro" whereas my respository has ".ti" and ".json" files, so I don't think this will work.

In a nutshell - can I use Bedrock with PAaaS and, if so, what is the recommended approach?

Thanks and Regards,

Mark

[lotsaram]

Hi Mark,

There isn't yet a new release branch of bedrock for v12. As many TI functions have been deprecated in PA Engine (basically anything to do with modifying transaction logging or saving data) the library as is won't work any many processes will refuse to compile or will throw errors in v12. The changes required are easy and minor, but they haven't been done is all.

You can think of "PAaaS", "v12" and "PA Engine" all being synonyms in this case.

In terms of the other part of your question (how to deploy), currently the text files in the bedrock repository are the offline files in the legacy .pro format which doesn't exist in v12 where processes are stored as odata json files. Whether the new bedrock release branch for v12 will use this exact format is undecided as it could make diffing even more painful than it already is with the existing format (which at least preserves white space and line breaks.) So no easy or definitive answer here just jet however I would definitely recommend having a separate instance for a bedrock pull and from there update other instances via rest command (e.g. with tm1py.)

[markhackett]

Thanks very much lotsaram. Appreciate the quick response. That makes sense - look forward to seeing the v12 version at some stage.

With regard to your suggestion, I presume this would be a non-v12 instance where I would keep a copy of all the Bedrock objects, and then use this as the "source" to copy processes, etc. to v12 environment?

Regards,

Mark

[lotsaram]

No I meant the clean "bedrock deploy instance" would also be v12. It's just a staging area to push into the real model(s). (Obviously this won't work until such time as there is a v12 compatible branch of bedrock TI.)

[markhackett]

Understood!

Thanks - Mark