xtdb 2022-09-09 | Slack Archive

Hukka07:09:41

How do you handle fixtures and migrations with xtdb? In SQL I would have SQL files to seed testing environments with some data, and go with migratus

Hukka07:09:39

The only discussion / mention of the topic I found was https://clojurians-log.clojureverse.org/xtdb/2021-11-14 but I was wondering if there are new developments or ideas

jarohen10:09:14

hey @U8ZQ1J1RR 👋 not dissimilar for XT, in our experience - in the repo, for example, we have an https://github.com/xtdb/xtdb/blob/master/test/resources/data/james-bond.edn of sample data, and then a https://github.com/xtdb/xtdb/blob/master/core/test/xtdb/pull_test.clj#L8-L13 to pull it into an in-memory node

plexus10:09:20

Hiii 👋 I'm looking at the docs for Checkpointing and I have some questions 😄 https://docs.xtdb.com/administration/1.21.0/checkpointing/ Does this only checkpoint the indices, or does it also include the whole history? Can I use this to restore a database, or only to add nodes to an existing one?

nivekuil10:09:32

it's only the indices, a checkpoint is just a full copy of the rocksdb directory along with some edn metadata

jarohen10:09:57

Hey @U07FP7QJ0 👋 Checkpointing's only intended for the indices, as a means of getting nodes joining the cluster to start faster.

plexus10:09:57

ok, that's what I thought. Seems there was some confusion about this in the team. To do actual backups people normally would use kafka-based backup solutions?

👍 1

jarohen10:09:00

for the golden stores (the tx-log and the document store) it's recommended to use the standard backup/restore practices for whatever store you've picked (e.g. Kafka, S3, JDBC, etc)

👍 1

2022-09-09

Channels