From cschum at suse.de  Mon May 28 01:19:17 2018
From: cschum at suse.de (Cornelius Schumacher)
Date: Mon, 28 May 2018 09:19:17 +0200
Subject: [Machinery] workload_mapper
In-Reply-To: <MW2PR18MB2268FE06C298E54AE8FCDA53896B0@MW2PR18MB2268.namprd18.prod.outlook.com>
References: <MW2PR18MB2268FE06C298E54AE8FCDA53896B0@MW2PR18MB2268.namprd18.prod.outlook.com>
Message-ID: <4688562.37sumaeSKS@e82>

On Mittwoch, 23. Mai 2018 13:05:19 CEST Sanjeet Kumar wrote:
> 
> I am trying to work on a couple of use cases with machinery.
> 
>   1.  export to kickstart
>   2.  RedHat/Centos to SUSE/openSUSE migration
> 
> 
> For first use case, I wanted to understand where(in which file)
> export-{autoyast/kiwi etc} are called. Please help me understand the flow.
> I tried to trace using byebug unsuccessfully, I am looking for some quick
> help. I find that referring lib/autoyast.rb to create a lib/kickstart.rb
> should not be a challenge.

It's happening in the Cli class:

https://github.com/SUSE/machinery/blob/
5e1d017b1c7463677d46e371ca2ee9b444481a0f/lib/cli.rb#L580

There you can also find the call for the kiwi export. A Kickstart export could 
be modeled in the same way.

> For second use case, I wanted to understand the use of workload_mapper. I
> was wildly guessing if workload-mapper is a model of migration-engine and
> for some reason its contribution was stopped?

This came from an experiment to implement containerization of workloads. We 
didn't move ahead with that so the code which is there is quite limited. It 
could probably be the base for a more sophisticated migration engine but this 
most probably would be quite a bit of work. If you have a quite specific use 
case where the environment is known to a good degree it might be less work.

-- 
Cornelius Schumacher <cschum at suse.de>

From cschum at suse.de  Mon May 28 01:20:00 2018
From: cschum at suse.de (Cornelius Schumacher)
Date: Mon, 28 May 2018 09:20:00 +0200
Subject: [Machinery] Partition information
In-Reply-To: <BL0PR18MB225949D9D17246E37BEE04BD89680@BL0PR18MB2259.namprd18.prod.outlook.com>
References: <BL0PR18MB225949D9D17246E37BEE04BD89680@BL0PR18MB2259.namprd18.prod.outlook.com>
Message-ID: <4522211.5Ql4LcKTxN@e82>

On Samstag, 26. Mai 2018 12:56:03 CEST Sanjeet Kumar wrote:
> 
> I wanted to understand if there any reason why we are not capturing
> partition using inspect? My reason for exploring about this is for
> migration use cases where it would help to know the partition information
> of the source system. An application or service might be needing a special
> partition setup.

That's a very fair point. There is no particular reason why partition 
information is not extracted yet. This would just need another scope to be 
added.

-- 
Cornelius Schumacher <cschum at suse.de>