[Pulp-list] Generic Files and PULP_MANIFEST Syncing

[Kodiak Firesmith]

What you desire is actually a common refrain.  I thought there was already
an RFE in for this for Pulp 3 but I could be mistaken.  IIRC, I was told a
lot of the complexity of doing a sync with any more elegance than a scrape
is that without a manifest file it's hard to know you've gotten everything.
  But I can't find the discussion where this came up - it was probably on
Freenode #pulp which isn't logged.

I did however file an RFE that I encourage you to pile onto with your use
case:  https://pulp.plan.io/issues/1999

 - Kodiak

On Thu, Jul 28, 2016 at 9:03 AM, Eric Helms <ehelms at redhat.com> wrote:

> We (Katello) have been working to expose the "iso" type to users as a
> "file" type repository for exposing generic content support. In this way,
> users can use Katello as a single gateway for any content and Pulp to store
> and manage the underlying content for them. I've created a feature request
> [1] to make this a first-class plugin in Pulp.
>
> On to my questions. As I understand it, users can either upload content
> directly or sync a "file" location in the same way that say an RPM
> repository is synced. The difference being that Pulp requires a
> PULP_MANIFEST file that indicates:
>
>  * filename
>  * filesize
>  * checksum
>
> And while we can provide a script to generate this recursively for users
> [2], this requires them to have access to the content to place the
> PULP_MANIFEST file or to mirror the content locally, generate the manifest
> file and then sync. However, the latter seems both redundant and the intent
> of Pulp (to mirror content for me!). Now for my questions:
>
> 1) Is there a work around or way to configure Pulp to not require the
> manifest file and simply slurp up whatever it finds?
> 2) Is there a strategy we could implement to do #1 if it does not already
> exist? Something that perhaps scans the structure, builds some base level
> knowledge of what is present and allows Pulp to sync it?
>
> I think this would also play into the situation where I want to slurp up
> larger files without having to upload (as I feel Pulp syncing them is more
> robust than me uploading, this could however be a misconception on my
> part). I'm thinking of the situation where I want to use Pulp or Katello as
> a repository for my internal Vagrant boxes for example.
>
>
> Thanks,
> Eric
>
> [1] https://pulp.plan.io/issues/2110
> [2] https://gist.github.com/ehelms/3fd956ee887db3d7bac20b29efa3dd51
>
> _______________________________________________
> Pulp-list mailing list
> Pulp-list at redhat.com
> https://www.redhat.com/mailman/listinfo/pulp-list
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/pulp-list/attachments/20160728/ffba37b4/attachment.htm>