07/06/2021, 9:02 PM
Hi everybody! I was wondering the best approach to handle a
point in time recovery
with dynamoDB. I created a script to recreate a table using the aws-cli, but pulumi doesn't know the creation of the table. If I added into the code the new table, it will try to create the table and the only thing I need to do is to enable the
point in time recovery
. Can anybody help me with this?


07/06/2021, 9:47 PM
So are you looking to create the table in the beginning and enable point in time or if you lose it, refill the table then enable point in time?


07/06/2021, 10:39 PM
I have a table already with pitr (I created with pulumi) . If the table for some reason (a bad fix) mess up the data I need restore the table at certain time. I saw pitr, but aws says you need to create a new one.
aws dynamodb restore-table-to-point-in-time \
    --source-table-name Music \
    --target-table-name MusicEarliestRestorableDateTime \
    --no-use-latest-restorable-time \
    --restore-date-time 1519257118.0
So the new table is created (outside pulumi)
And I need a way to tell pulumi that there’s a new table
So I can activate the pitr in the new table. I saw the pulumi import, but not sure if the right command to solve the problem. And Idk how to use it to Dynamo (there is only a s3 example)
I'm facing this issue right now. I'm creating a new table in dyanmodb (part of the
point in time recovery
) The table
(for example) has two properties in the creation (Partition Key
and a Sort Key
). I'm doing a
pulumi import aws:dynamodb/table:Table basic-table MusicTableMinutesAgo
and I'm getting this:
all attributes must be indexed. Unused attributes ["sk"]
Is there a status regarding this situation? I saw an issue on github with the same problem