Documents Migrations

Hello,

You mean documents uploaded by the end-user in a test environment to put it in another?

It really depends what is your set-up. For instance, do you have a folder per record? Do you save the documentId or a folder id? are the documents named the same?

Probably you will need to do a script to migrate. You need to consider that

- Every document uploaded has a new document id, meaning you will need to re-do all references

- Every folder in each environment has a unique folderid, meaning if you use it, you need to re-do all the references.

Uploading the documents can be easy, you download from one environment and upload on the other, however most of the times that will be useless as you lose the references.

It depends on your set-up, but probably the easiest is to do a CDT that contains metadata and that i saved in the DB and overwrite the names of the files from the source environment. Then, on the target environment use the same CDT import the DB data with the metadata. Use an script to upload multiple files and map your file name with your record.

Cheers,

Manuel

Yeah we definitely need more information here. If they are documents uploaded by a developer into Appian, those objects behave like any other design object. You can import, export, and create deployment packages with those documents.

If you are talking about objects uploaded by users, I would be interested in understanding your use case in more detail. Usually objects uploaded by users would only make sense in the current environment as Manuel mentioned above.

We have an onprem environment where user upload documents for each request ( around 5 Lakh documents in total). If we go with the below approach it take more time then expected for the migration , approximately it takes 4 hours for migrating 1000 documents from onprem to cloud environment)

Onprem Process:

1. Identify the folder and docs in onprem , store the metadata(Folderid,DocId,Doc name....) in Temp table.

2. Move the documents to a File server location .

Cloud Process:

1. Iterate trough each record in the temp table

2. Download the file from file server and create a folder if required and move the document to folder.

3. Update the old folder id reference in the table with new folderid and document id

Do we have any built in solution provided by Appian on this document migration approach. any better approach to do run this process faster considering the document count is huge ?!?

I am considering other approaches like

1. Adding the metadata in the description of the document and use it in the cloud for retrieving the document from FTP and update the docid in the reference tables.

2. Any posibility of moving the document from onprem to cloud by creating an webapi service at both end and store the doc in folder and update the metadata in tables.

To clarify - is this a one time migration of your environment from on-premise to cloud? If so, I would strongly suggest reaching out to Appian Support. They can provide guidance on the best way to perform the migration that will meet your specific needs.

Yes, this is one time migration. I am also reaching out to appian team on this. Just want to check if any has came across better solution for this in the meantime.

It seems as if you could speed up the processing time somewhat by taking your approach and making it more multi-threaded.  That is, have some parent process launch multiple instances of your transfer process, where each one is set to handle some specified subset of your entire data.  If this is in an environment with multiple engines, etc, then this approach could make good use of those.