Fork me on GitHub

I have no experience with XTDB but while reading up on it I started wondering if there was a delay between having written some piece of data and it showing up in reads. In something like PG I know when a transaction commits that it can be read immediately. In case it's the same with XTDB: does anybody know how this is solved? How would the transactor know that the caches (or at least the one on the same node) are up to date?


If you need to read from your transaction, you can elect to await for it with await-tx.


Ah ok that's great to know! What kind of latency would you expect when you await? And do you know how XTDB solves this problem? It seems rather challenging if you write to a log but read from caches

Jacob O'Bryant18:02:11

The caches/indexes are updated eagerly from the log, so it's fairly straightforward (at least from my perspective as a user!). I don't remember what I'm getting for latency off the top of my head, but I use await-tx regularly in my web app for rendering the next page after POST requests

👍 1

but, you need to be careful with integrations... if you have a bulk import endpoint that does "fire and forget" you could DoS yourself with that

✔️ 1