database - In Apache Derby, for large data churn, do you need to compress? - Stack Overflow

admin管理员组

文章数量:1391934

The TLDR: In a high churn Apache Derby database, if we never compress the database and only do an update statistics, will that affect performance long term?

We have an Apache Derby database running as a network server on a linux box. This is a sort of microsystem with a GUI app running on a laptop, hitting its own processing box where this "server" resides. It's a one to one: client to server environment that way. This database goes through high data churn: up to 20,000 new records a day followed by up to 20,000 deletes of old data daily totaling up to 4 GB coming and going essentially. Also this system does not run all the time, so we can't schedule something to happen during off hours.

If I do a compress on startup, this naturally can take a long time (I've seen up to 10 minutes). Naturally it's not great to make the user to wait 10 minutes before they can do anything. Compress also updates statistics which is important for performance for us. So can we simply skip the compress completely, and only ever do an update statistics on startup? Will this slow down query performance overtime if we never compress?

本文标签： databaseIn Apache Derbyfor large data churndo you need to compressStack Overflow