Ben Johnson
Ben Johnson
Document the minimum policy for `replicate` (read-write) and for `restore` (read-only). Currently there is a policy specified in JSON (via https://github.com/benbjohnson/litestream.io/pull/12) but another proposal exists in YAML (https://github.com/benbjohnson/litestream/issues/76#issuecomment-783926359). Need to...
https://sqlite.org/src/doc/begin-concurrent/doc/begin_concurrent.md
Currently, Litestream sends cleartext bytes to a replica such as S3. This works well enough for many use cases as S3 buckets are generally secured through IAM. However, other use...
Currently, Litestream requires users to explicitly to restore a database using the `litestream restore` command. However, it would be ideal to automatically restore a database when Litestream starts up and...
## Overview Per @kellabyte's suggestion, the `DB.Batch()` method needs better documentation about the guarantees it provides -- specifically that it is not serializable because goroutines do not have an ordering...
## Problem As mentioned by @br0xen on the #boltdb channel on Gopher Slack, opening a file that is not a Bolt database will cause the size to increase. If the...
## Overview As discovered by @evanphx, if you write data that exactly fills a page and you set `FillPercent` to `1` then a empty overflow page will be created for...
Currently, LiteFS supports a single primary node that performs all the writes. However, there are situations where it would be useful to have multiple nodes that can write—even if it...
LiteFS provides some redundancy by running in a cluster, however, losing all nodes would cause all data to be lost. Replicating to S3 in manner similar to [Litestream](https://litestream.io/) would provide...