feat: delete events #4
Conversation
* following same structure as r2dbc plugin, which also has the grouping of deletes to a max number of events per statement * using TransactWriteItems, deletes must be on individual pk * delete marker to keep highest sequence number if all events have been deleted
| @@ -41,27 +42,35 @@ import software.amazon.awssdk.services.dynamodb.model.QueryRequest | |||
| import JournalAttributes._ | |||
| val req = QueryRequest.builder | |||
| .tableName(settings.journalTable) | |||
| .consistentRead(true) | |||
There was a problem hiding this comment.
I saw some flaky test fails in ci. Thought it was this, but maybe there is something more because another fail after changing this.
However, we need consistentReads for replay and seqNr queries. Seems default is false.
There was a problem hiding this comment.
Yes, we'll need consistent reads. I wonder if the local in-memory DynamoDB creates replicas anyway?
We should probably start collecting test failures that need follow up. I'll try a few re-runs as well.
There was a problem hiding this comment.
Ran a few times and several similar errors. Always looks like eventually consistent reads. Eventually discovered this:
Read operations are eventually consistent. However, due to the speed of DynamoDB running on your computer, most reads appear to be strongly consistent.
So looks like dynamodb-local is always eventually consistent reads, ignoring the consistent reads flag? Not sure on its behaviour, but maybe we can just wait for writes to be seen first in the tests. And we'll probably want some regular testing against actual DynamoDB as well.
DynamoDBJournalSpec is now successful