Switch branches/tags
Commits on Jan 8, 2014
Commits on Jan 7, 2014
Commits on Jan 6, 2014
Commits on Jan 4, 2014
  1. Add logging with rotation and pretty printer.

    * Fix race condition in EQC test that arose from queries happening outside of gen_server.Adjust
    * Implement disk_log logging of events based on log level
    * log level = on | off | debug
    * normal (on) logs get_token and get_lock events
    * status is included in log message so we can see when processes get denied resources
    * included a term to ascii pretty printer that can be executed from the shell
    * it's possible (and probable!) to generate messages sequence charts from this data
    * Use extended date log file format in pretty printing
    buddhisthead committed Jan 4, 2014
Commits on Dec 26, 2013
  1. Merge pull request #491 from basho/jrw-remove-bad-forwarding

    remove legacy forwarding code that snuck in as part of 632af2b
    jrwest committed Dec 26, 2013
Commits on Dec 25, 2013
  1. remove legacy forwarding code that snuck in as part of 632af2b

    Handling of forwarding code was changed in #301. Changes made as part
    of the commit mentioned above were most likely made prior to that PRs merge.
    Remove the forwarding code as it is redundant (the decision to forward
    is handled further up the call stack). It caused vnode when resizing to
    crash, however, since the check was not as comprehensive as its
    caller.  The vnode would crash when attempting to forward to a
    partition identified by `{Index, [{Index, Owner}]}` which is invalid.
    jrwest committed Dec 25, 2013
Commits on Dec 24, 2013
  1. Look at eleveldb develop branch instead of master

    Jared Morrow committed Dec 24, 2013
  2. Merge pull request #486 from basho/fsm-refactor-support

    riak_core support for kv#734
    evanmcc committed Dec 24, 2013
  3. Merge pull request #437 from basho/adt-security-enable

    Implement riak security enable/disable/status
    Vagabond committed Dec 24, 2013
  4. Implement riak security enable/disable/status

    Implemented in terms of a cluster_metadata modulated capability.
    Vagabond committed Oct 23, 2013
  5. Merge pull request #415 from basho/feature/change_net_ticktime

    Add infrastructure to change the OTP net_kernel's net_ticktime
    slfritchie committed Dec 24, 2013
  6. Merge branch 'develop' of into develop

    Too many merges happening this evening, and I was too slow.  Re-merge.
    slfritchie committed Dec 24, 2013
  7. Merge pull request #451 from basho/adt-security-admin

    Implement alter-user, del-user and del-source
    Vagabond committed Dec 24, 2013
Commits on Dec 23, 2013
  1. Address review comments

    Vagabond committed Dec 23, 2013
  2. Initial version of del_user

    Vagabond committed Nov 9, 2013
  3. Merge pull request #482 from basho/bugfix/ensure-stats-progress

    Ensure stats progress and tag individual stale stats
    evanmcc committed Dec 23, 2013
  4. refactor staleness tagging

    evanmcc committed Dec 23, 2013
Commits on Dec 22, 2013
  1. Add infrastructure to change the OTP net_kernel's net_ticktime

    Changing the net_ticktime on a single node is easy: it's a single
    function call.  However, in a cluster environment, that function
    must be run on all nodes simultaneously, including any new nodes
    that connect to the cluster during the time change ... and that
    includes any hidden nodes.
    In Riak's case, those hidden nodes are usually `nodetool`, which
    is an escript.  Those scripts don't have access to `riak_core_util.erl`
    or to any other module module that a fun sent over-the-network
    might refer to.  So there's this icky dance of checking the remote
    node's code server path and appending to the code path if the
    riak_core dir is missing.
    There is additional icky spawns and timeout checking and exit()
    use to avoid: 1. being blocked by the local net_kernel when trying
    to use `rpc:call/4` or `spawn/2` to a remote node that's down/slow,
    and 2. leaking zillions of processes during such net_kernel instability.
    Yes, there's an accompanying riak_test test.
    slfritchie committed Oct 9, 2013
Commits on Dec 21, 2013
  1. Merge pull request #468 from bbrietzke/from-pr-464

    Changing how pathing is determined
    jrwest committed Dec 21, 2013
  2. do you even compile

    evanmcc committed Dec 21, 2013
Commits on Dec 20, 2013
  1. - remove a sneaky call to now()

    - add chash_key/2
    evanmcc committed Dec 20, 2013
  2. Merge pull request #469 from basho/adt-cipher-suites

    Make cipher suites configurable via the command line
    Dave Parfitt committed Dec 20, 2013
  3. Merge pull request #463 from basho/bug/rdb/sibling-explosion

    Add get dot and dot type for dvv style causality
    russelldb committed Dec 20, 2013
  4. Merge pull request #484 from basho/cet-bg-mgr-vnode-lock-rebased

    Background Manager Integration with Handoff
    buddhisthead committed Dec 20, 2013
  5. Address final PR comments.

    buddhisthead committed Dec 20, 2013
  6. Address remaining issues in PR

    * Remove lock defs file; it was moved to riak_kv_vnode.hrl since all these locks are KV-specific
    * The optional call to handoff_started/2 allows vnodes one last chance to abort the handoff process. the function is passed the source vnode's partition number and node name (this node) because the callback does not have access to the full vnode state at this time. In addition the worker pid is passed so the vnode may use that information in its decision to cancel the handoff or not (e.g. get a lock on behalf of the process)
    * Move code that calls handooff_started/2 into separate function and put the call inside the try/catch.
    buddhisthead committed Dec 19, 2013