memcached 1.4.31 发布了，主要更新如下：
- More fixes for defaults related to large item support.
- Several improvements to how the LRU crawler’s default background job is launched. Should be less aggressive.
- Fix LRU crawler rate limiting sleep never actually being used.
-o modern is in use (or at least
-o lru_crawler), a new command is usable:
lru_crawler metadump [all|1|2|3]. You may dump all of the metadata for a particular slab class, a list of slab classes, or all slab classes.
This is useful for inspection, pulling cache data from one server to another, etc.
This is not a way to replicate a cache: it is an inconsistent non-multiversioned grab of the key storage. It makes a best effort to not include duplicates but a small number can happen on busy servers.
Many thanks for Netflix, for sponsoring and patiently testing the work done in this release!
More details and code can be seen in the PR:https://github.com/memcached/memcached/pull/193