Permalink
Commits on Jun 12, 2009
Commits on May 15, 2009
  1. Don't abort when creating a socket with an unknown binding protocol.

    dustin authored and Trond Norbye committed May 14, 2009
    This changes the abort to an assert and tests for such failures.
Commits on May 13, 2009
  1. When receiving bad magic in the binary protocol, hang up.

    dustin authored and Trond Norbye committed May 13, 2009
    try_read_command's return value of 0 indicates to the caller that more
    data is needed, thus it changes the connection state to conn_waiting
    which blows away the closing state and the connection stays open
    forever accumulating bad input.
  2. Allow the binding protocol to be specified.

    dustin authored and Trond Norbye committed May 13, 2009
    Instead of always binding as autonegotiate, allow the user to specify
    which protocol the server will run.
  3. Moved struct definitions below enum declarations.

    dustin authored and Trond Norbye committed May 13, 2009
    enums end up in structs, so they should come first.
Commits on May 3, 2009
  1. fix and test for issue 38 (server does not respond to binary requests)

    Eric Lambert Trond Norbye
    Eric Lambert authored and Trond Norbye committed Apr 9, 2009
Commits on Apr 17, 2009
  1. More docs.

    dustin committed Apr 17, 2009
  2. Doxygenating.

    dustin committed Apr 17, 2009
  3. Added documentation on the cache interface

    Trond Norbye authored and dustin committed Apr 17, 2009
  4. Create a generic cache for objects of same size

    Trond Norbye authored and dustin committed Mar 27, 2009
    The suffix pool could be thread-local and use the generic cache
Commits on Apr 13, 2009
  1. Ignore SIGHUP when running as a daemon

    Trond Norbye Trond Norbye
    Trond Norbye authored and Trond Norbye committed Apr 11, 2009
Commits on Apr 12, 2009
  1. Changelog update from 1.2.8 fixes.

    dormando authored and dustin committed Apr 11, 2009
  2. fix resumption of accept loop under multithread

    dormando authored and dustin committed Apr 11, 2009
    critical fix:
    under multithreaded mode, in version 1.2.7, memcached would not resume
    accepting connections after hitting the max connections limit.
    
    this is due to a pre-existing bug in the accept_new_conns code, which
    was hit when the "fix" was rolled in to ensure listening threads never
    did real work.  Previously, at least one closing connection would
    (randomly) be on the accepting thread, so the test for (are we the
    listening thread?) would still work.
Commits on Apr 10, 2009
  1. repeated s option in getopt

    steveyen committed Mar 27, 2009
Commits on Apr 6, 2009
  1. Stat growth assertion.

    dustin committed Apr 6, 2009
Commits on Apr 3, 2009
  1. Changelog update.

    dustin committed Apr 3, 2009
Commits on Apr 2, 2009
  1. Don't expose the protocol used to the client api of the stats

    Trond Norbye authored and dustin committed Apr 2, 2009
    (dustin) I made some changes to the original growth code to pass in
    the required size.
Commits on Apr 1, 2009
  1. 'stats reset' resets new main stats.

    dormando authored and dustin committed Apr 1, 2009
Commits on Mar 31, 2009
  1. Changelog updates for 1.3.3.

    dustin committed Mar 31, 2009
Commits on Mar 30, 2009
  1. two new troubleshooting stats

    dormando authored and dustin committed Mar 27, 2009
    accepting_conns for completeness, and listen_disabled_num to see how many
    times you've hit maxconns and disabled incoming connections. probably a good
    stat to monitor and flip out on.
  2. add a cmd_flush stat

    dormando authored and dustin committed Mar 25, 2009
    shouldn't add much lock contention for just this.
    I want to add this one stat (mayb a few more?) since it's happened more than
    once that folks think memcached is broken when a cron or something is calling
    'flush_all' once a minute.
Commits on Mar 29, 2009
  1. print why a key was not found when extra verbose

    dormando authored and dustin committed Mar 25, 2009
    simple logs for simple people. Patch inspired by a bug hunting session with
    evan weaver. It's been useful a few times since.
  2. dumb hack to self-repair stuck slabs

    dormando authored and dustin committed Mar 28, 2009
    since 1.2.6, most of the refcount leaks have been quashed.
    I still get them in production, extremely rarely.
    It's possibly we'll have refcount leaks on and off even in the future.
    
    This hack acknowledges this and exists since we want to guarantee, as much as
    possible, that memcached is a stable service. Having to monitor for and
    restart the service on account of "rare bugs" isn't acceptable.
  3. fix a handful of socket listen bugs.

    dormando authored and dustin committed Mar 26, 2009
    AF_UNSPEC is still necessary for UDP sometimes.
    We guarantee that at least one address returned from getaddrinfo binds
    successfully, and in cases of lacking network or ipv6 addresses some of those
    socket() calls might fail. That's normal. We were bailing on them.
    This change also removes the need to pass AI_ADDRCONFIG on machines with ipv6
    stacks disabled.
Commits on Mar 27, 2009
  1. Added a HACKING file.

    dustin committed Mar 27, 2009
Commits on Mar 26, 2009
  1. Use git's version number for releases.

    dustin committed Mar 9, 2009
    This will allow more specific version numbers, while simplifying a
    proper release down to a tag and make dist.
    
    During development, ./version.sh needs to run periodically to update
    the version number.  I'd recommend just adding a call to version.sh as
    a git post commit hook:
    
    % cat .git/hooks/post-commit
    
    echo "Updating version."
    ./version.sh
    
    (and make sure the file is executable)