Hacker News Reader

HNHacker News

Top New Best Ask Show Jobs

Comment by rapsey | Hacker News Reader

Parent Full thread

rapsey·

Because memcache and redis are in-memory. Writing to fdb will be complete once it is fsync'ed to disk.

Memcache is a cache. Fdb is a an ordered kv store.

View on HN

Redis can be configured to persist and fsync every operation.

Redis is not meant as a primary database and should not be used as such. FoundationDB is meant as a reliable source of truth.

Redis began as a caching database, but it has since evolved into a primary database. Many applications built today use Redis as a primary database.

https://redis.com/blog/redis-cache-vs-redis-primary-database...

Very much seems like an acceptable use now

Redis can't have a working set larger than memory. It has no mechanism to page data to disk. If your data set grows too large, you're hosed unless you add more hardware.

You can configure redis to flush to disk on write operations though you lose on performance.

That is not comparable to writes completing after they are flushed to disk.

>appendfsync always: fsync every time new commands are appended to the AOF. Very very slow, very safe. Note that the commands are appended to the AOF after a batch of commands from multiple clients or a pipeline are executed, so it means a single write and a single fsync (before sending the replies).

https://redis.io/docs/management/persistence/

It's very slow, but if you really want to wait for fsync before replying, it can do that.

I was unaware they could make that guarantee.

Thanks for the correction.

Reply on news.ycombinator.com