Forum Discussion
second migration to Fabric - what is the best approach
hello
we have d365fo which was running on export to data lake; this was migrated to fabric at the time that the default collation was case sensitive; now that there is possibility to have case insensitive collation id like to migrate to a case insensistive;
we have 2 enviornments, uat and prd; on our dataverse environements we have also plenty of other applications running;
data from dataverse are used via fabric link in our warehouse
i see here two options:
1. configure synapse link in current dataverse uat and switch the warehouse to use the synpase link; unlink the fabric link from current dataverse uat and link back to case insensitive workspace; switch back warehouse to use fabric; do the same for prd later on
2.create new dataverse environemnt , and create fabric link to case insensitive workspace; migrate all the current apps from dataverse to new dataverse environemnt; switch the warehouse to use the new fabric link; do the same for prd later on
which option will be more safe, efficient, less prone to errors, and also pricey? are there any other options?
any tips /advise is appreciated
thanks very much!
Hi mama_Hani,
The advantages of option one is that you do not need to rebuild dataverse and migrate all apps. This would definitally be faster to implement. But the cons are that you do not have an environment to do a test migration and you have the risk of temporary not having data available during the transition. Make sure you reprocess all data to prevent collation conflicts.
Option two is more work, you have to migrate your apps, data, security roles, etc.. And you might also have extra costs for licensing an extra environment. But this option is easier to test before you transition.
There could be an option three, where you clone your current dataverse environment to a new one. You can link this environment to a case-insensitive workspace and test everything. Then you can gradually migrate you apps one by one. This option combines both options.
Hope this helps. If so, please give a Kudos 👍 and mark as Accepted Solution ✔️.
3 Replies
- nielsvdc
Super User
Hi mama_Hani,
The advantages of option one is that you do not need to rebuild dataverse and migrate all apps. This would definitally be faster to implement. But the cons are that you do not have an environment to do a test migration and you have the risk of temporary not having data available during the transition. Make sure you reprocess all data to prevent collation conflicts.
Option two is more work, you have to migrate your apps, data, security roles, etc.. And you might also have extra costs for licensing an extra environment. But this option is easier to test before you transition.
There could be an option three, where you clone your current dataverse environment to a new one. You can link this environment to a case-insensitive workspace and test everything. Then you can gradually migrate you apps one by one. This option combines both options.
Hope this helps. If so, please give a Kudos 👍 and mark as Accepted Solution ✔️.
hi, thanks very much for your reply
why do you think i would not have an environmnet to do a test migration?
- enable synapse link on uat -> dump data to dwh dev storage account
- source my dwh dev from dwh dev storage account
- my current fabric uat is used as source for dwh acc
-> then i can compare these two: dwh dev vs dwh acc so i know they are identical;then if they are identical
- i unlink the fabric link from uat and relink to case insensitive workspace;
- then i source the dwh acc again from the fabric uat
-> then i can compare these two: dwh dev vs dwh accthen if they are identical
- i switch the dwh dev to use the fabric uat as well,
-> then i can compare these two : dwh dev vs dwh accthen if they are identical i move to prd:
-i enable synapse link on prd - > dump data to dwh acc storage account
-source my dwh acc from dwh acc storage account
-my current fabric prd is used as source for dwh prd
-> then i can compare these two: dwh acc vs dwh prd so i know they are identical
etc-> like with uat migration
am i missing sth?what about the costs, which will be cheaper and by what order of magnitude, by 100 usd, by 1000 usd, etc , any idea?
in synapse link we will have spark costs, but in dataverse, new environemnt with the full set of data, seems to me it will cost a lot, just dont know which will be cheaper- nielsvdc
I think your plan sounds solid. You already have a test environment simular to production, with which you can test the migration scenario before migrating the production environment.
The cost might not be an issue, depending on the type of licensing you have for your users and the amount of data you have in Dataverse. Check the Power App pricing details and look at your licensing situation. But if you already have a uat environment with the same amount of data as in production, then you should not experience any extra costs for licensing or storage for Dataverse.
Hope this helps. If so, please give a Kudos 👍 and mark as Accepted Solution ✔️.
