Simple memcached server.
###Usage
go get github.com/Skipor/memcached/cmd/memcached
memcached --help for all options
memcached to start server on default port
memcached -aof-name ./memcached.aof to start server with AOF persistence on
memcached -aof-name ./memcached.aof -sync 2s sync log not on every command, but every passed period
- No third party packages for app, server and cache logic.
- That was tack constraint.
- I have used only small github.com/facebookgo/stackerr package for errors with stack traces.
- Parallel reads, serialized writes.
- Items data sync.Pool recycle.
- Low allocation text protocol parse.
- AOF persistence with configurable sync options.
- Every command can be synced, or sync can be done by ticker.
- Low latency fast log rotation.
- When log grows large, background goroutine takes fast snapshot under read lock.
- Snapshot is writen without log lock.
- Many unit and integration tests.
- LRU crawler
- goroutine waking up on ticker; take RLock; start goroutine per queue; find expired; save them; take write lock and remove them.
- Refactor integration load tester. Do more load tests and comparation with original memcached.
- Test CLI helper functions
- RDB persistence
- point-in-time snapshots of your dataset at specified intervals
- current version already use snapshots to archive large AOF, so it is easy to implement such feature
- Profiling with pprof and optimisation
- Snapshot consistency check
- Now check happen only in build with debug tag
- Optional
- replace state snapshot usage as AOF rotation