We are thinking about use "Mirroring" on our new website. EPiServer CMS R2.
We want to use a tree channel from the startpage, to make a replica of the source site. We are thinking about using a PageProvider to provide information from another system on the source site.
Will the PageProvider-pages on the source-site be replicated to the destination-site as well? Or do we have to configure the PageProvider on the destination-site. We would prefer not, to avoid opening up to the inside.
All pages below your mirroring root will be mirrored regardless of serving page provider, that is your PageProvider-pages will be mirrored as well.
On the destination though they will stored as "ordinary" EPiServer pages, that is stored in EPiServer db (unless your destination root is an entrypoint or an page served by some other PageProvider).
If your destination root is an entrypoint or a page served by another page provider than the "local" page provider then all mirrored pages will be created by that provider.
So an usage of PageProviders can actually be to move data from one storage to another. So you could use PageProviders to import/export data to/from EpiServer.
One thing to remember using this though is that mirroring "large" tree structure takes a lot of the system resources. And if you use a custom provider here you should make sure it can handle that amount of data.
Sorry for stealing this thread but, if I have a node with a custom page provider. Then I mirror that node. Will the mirrored node be searchable with a searchdatasource?
So if I understand you correctly this is very powerful from a security point of view. I mean that the destination-site can be isolated with no connections to the inside. Still we can do integrations with other systems on the source-site using custom pageproviders. And in the end all pages will be mirrored/replicated to the destination-site. If we use only the EPiServer-pageprovider at the destination-site all pages will be stored in the EPiServer destination DB.
Have I understood you correctly?
If your destination node is served by EPiServer (meaning pages are stored in EPiServer db) then yes the pages should be searchable. A note here as explanation is that the SearchDataSource uses db search meaning it only works with EPiServer db stored pages.
Yes you are correct.
One thing to have in mind though is that mirroring large structures can be a problem (you can reduce the problem by dividing the content in several mirroring packages). We are currently looking into increasing the scalability/performance of mirroring but you should not expect to have anything released before R3 regarding that.
I mean, thanks Johan