- summary Release Notes for Release 1.3.3
Date: 2009-04-03 Fri
Download link:
http://memcached.googlecode.com/files/memcached-1.3.3.tar.gz
Prevent connection refused during connection storms at the cost of kernel memory.
Show all current server settings (useful for troubleshooting as well as internal verification).
* Alignment bug in binary stats (bug26) * Occasional buffer overflow in stats (bug27) * Try to recycle memory more aggressively. (bug14) * incr validation (bug31) * 64-bit incr/decr delta fixes (bug21) * ascii UDP set (bug36) * stats slabs' used chunks (bug29) * stats reset should reset item stats, eviction counters, etc... (bug22) * Fix all stat buffer management
* More tests * More/better documentation * Code cleanup
1 or 0 to indicate whether the server is currently accepting connections or not.
The server will stop accepting connections when it has as many as it's configured to take.
The number of times socket listeners were disabled due to hitting the connection limit.
The number of times the flush command was issued.
With verbosity enabled, you can see *why* objects were not found. In many cases, an item exists under a given key, but is considered invalid due to lazy expiration or flush.
There is a rare, unidentified reference leak that causes a slab to be full of invalid objects that cannot be evicted via the LRU nor will they expire on their own.
Tail repair is a strategy by which we forcefully evict objects that are marked as ``in-use (that is, in-flight or otherwise being used), but haven't been accessed in a long time (currently three hours).
There is an additional stat that comes along with this (tailrepairs on a slab) that will allow you to detect that this condition has occurred on one of your slabs.
There were some issues listening to sockets on machines with different network interface configurations (i.e. no network, only ipv4, only ipv6, etc...).
The following people contributed to this release since 1.3.2. Please refer to the 1.3.2 release notes for more info:
ReleaseNotes133