Skip to content
Commits on Oct 18, 2010
  1. @steveyen

    implemented gdb-callable connections_diag

    This provides relatively easily accessible high-level diagnostics of
    what is going on inside moxi
    
    Change-Id: Iff04d63be694891bf0d329a89bc83ad5ffbd22b1
    Reviewed-on: http://review.membase.org/3302
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Aliaksey Kandratsenka committed with steveyen Oct 18, 2010
Commits on Oct 14, 2010
  1. @steveyen

    bug 2759 - two is not enough verbosity

    Change-Id: Idf6a3812f19d6a82f485434f99dce3bfb28acf2f
    Reviewed-on: http://review.membase.org/3236
    Reviewed-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    steveyen committed Oct 12, 2010
Commits on Oct 12, 2010
  1. @trondn @dustin

    Don't loose the cppflags added by automake

    This includes an -I directive to allow compilation of moxi from
    another directory (eg. ../moxi/configure && make )
    
    Change-Id: Ie84242bc12cfaee08e65ac1217fa161473131e9b
    Reviewed-on: http://review.membase.org/3222
    Reviewed-by: Dustin Sallings <dustin@spy.net>
    Tested-by: Dustin Sallings <dustin@spy.net>
    trondn committed with dustin Oct 11, 2010
Commits on Oct 5, 2010
  1. @steveyen

    bug 2600 - fcntl() F_GETFL was lying on windows

    The bad effect was that all sockets on windows ended up being
    incorrectly in blocking mode instead of non-blocking mode,
    after doing a sasl-auth.
    
    Change-Id: I06d8a1447b33bbeef4aa35bf119a9c448e635a55
    Reviewed-on: http://review.northscale.com/3047
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Oct 4, 2010
Commits on Oct 4, 2010
  1. @steveyen

    more asserts around read()

    Change-Id: Iea765fa712daad66c24176468c19b5e0c3e34f55
    Reviewed-on: http://review.northscale.com/3035
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Oct 4, 2010
Commits on Oct 3, 2010
  1. @steveyen

    bug 2570 - downstream_conn_max unlimited again

    Change-Id: If48ad1ab11023e2ecf7d681d14a5f32fc4cc3876
    steveyen committed Oct 2, 2010
  2. @steveyen

    use appropriate lock before accessing proxy->name

    Change-Id: I4b38326385bb0ed0565cde998a0e23d3f1e6cfea
    steveyen committed Oct 2, 2010
  3. @steveyen

    bug 2605 - propagate usr/pwd correctly to mcs layer

    A helpful fix for empty-password buckets like the default bucket.
    The host_ident hash string caching in mcs, in particular, was
    sensitive to formatting NULL vs empty ("") passwords.
    
    Change-Id: I17650074e2a962881f9904c7c6ede218f58ae9d5
    Reviewed-on: http://review.northscale.com/2998
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Oct 2, 2010
Commits on Oct 2, 2010
  1. @steveyen

    logging had 1 too many ampersands

    Change-Id: I69af882a1cda688f3c9b4f533be662fd482115ad
    Reviewed-on: http://review.northscale.com/2992
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Oct 2, 2010
  2. @steveyen

    bug 2600 - more logging for diagnosis

    Change-Id: I695b3bc0d3a911e8dc33cb6e04be214f858edbc8
    Reviewed-on: http://review.northscale.com/2981
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Oct 1, 2010
  3. @steveyen

    bug 2602 - fixed multitenancy default bucket test

    In the new regime of saslBucketsStreaming, moxi will sasl auth for the
    default bucket, using username 'default', password '', so this commit
    updates the mock tests to reflect that.
    
    Change-Id: I79f0af838377f155b2d8c2a62af13549bc120987
    Reviewed-on: http://review.northscale.com/2975
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Oct 1, 2010
  4. @steveyen

    bug 2602 - update test data to match cleaned up JSON format

    Change-Id: I7e19ced9e17bb09a38ea753b2e620e20bbe45ac7
    Reviewed-on: http://review.northscale.com/2974
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Oct 1, 2010
  5. @steveyen @chiyoung

    moxi logging thread and buffer safety

    The scratch buffer wasn't thread-safe, as we saw an assert when going
    past the MAX_LOGBUF_LEN.  Added some extra byte slop to avoid going
    past that, too.
    
    Change-Id: Ie1b3135e2b1b41c5952e2d224fc614b3fb194054
    Reviewed-on: http://review.northscale.com/2965
    Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
    Reviewed-by: Chiyoung Seo <chiyoung.seo@gmail.com>
    steveyen committed with chiyoung Oct 1, 2010
Commits on Oct 1, 2010
  1. @steveyen @seanlynch

    bug 2602 - use name/saslPassword, not vBucketServerMap.user/password

    Change-Id: I365d61ef3b6c29454a943851115b85fb91db2052
    Reviewed-on: http://review.northscale.com/2933
    Tested-by: Sean Lynch <seanl@literati.org>
    Reviewed-by: Sean Lynch <seanl@literati.org>
    steveyen committed with seanlynch Sep 30, 2010
  2. @steveyen @seanlynch

    bug 2608 - use json-provided bucket name rather than "default"

    Change-Id: Ifb5bc73e39fa0925e5bd7a9f756ca87029752575
    Reviewed-on: http://review.northscale.com/2931
    Reviewed-by: Sean Lynch <seanl@literati.org>
    Tested-by: Sean Lynch <seanl@literati.org>
    steveyen committed with seanlynch Sep 30, 2010
  3. @steveyen @seanlynch

    bug 2608 - moxi should ignore unhealthy nodes for ketama

    Change-Id: Ic7b8654781d08bc4f0ada1d538b3a7c15e63ddf9
    Reviewed-on: http://review.northscale.com/2925
    Reviewed-by: Sean Lynch <seanl@literati.org>
    Tested-by: Sean Lynch <seanl@literati.org>
    steveyen committed with seanlynch Sep 30, 2010
Commits on Sep 30, 2010
  1. @steveyen

    bug 2592 - extra case of reading bucket pswd from JSON

    Missed one of the many places to get a user/pswd from the JSON.
    
    Change-Id: I2eff20a487f61c74001e6b1b627d1f6a095c252e
    Reviewed-on: http://review.northscale.com/2917
    Tested-by: Steve Yen <steve.yen@gmail.com>
    Reviewed-by: Dustin Sallings <dustin@spy.net>
    steveyen committed Sep 30, 2010
  2. @steveyen

    bug 2576 - relevant err msg when bad uds socket

    Change-Id: I714feaf92d0596c8f81e6c60ccb8848345ecc585
    Reviewed-on: http://review.northscale.com/2884
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Sep 29, 2010
  3. @steveyen

    bug 2541 - allow heterogeneous buckets (memcached & membase)

    Change-Id: I3fb9b361ef0124a53c27ed56b6a771b25aafaba9
    Reviewed-on: http://review.northscale.com/2883
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Sep 29, 2010
Commits on Sep 29, 2010
  1. @steveyen

    bug 1998 - better error msg when client tries to use NULL bucket

    Change-Id: Ia434f701c76e0174535894491a4329244c21d82f
    Reviewed-on: http://review.northscale.com/2861
    Tested-by: Steve Yen <steve.yen@gmail.com>
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Sep 29, 2010
  2. @steveyen

    bug 1998 - handle missing default bucket correctly

    This change introduces or corrects a few related concepts...
    
    - default_bucket_name
      - This is a -Z cmd-line key-value parameter.  Example:
        "-Z default_bucket_name=default,port_listen=11211"
      - By default, the default_bucket_name parameter is,
        confusingly, FIRST_BUCKET.
      - moxi assign new client/upstream connections to the bucket named
        $(default_bucket_name), if it exists.  If that named bucket
        doesn't exist, new client/upstream connections are assigned
        to the NULL_BUCKET.
    
    - FIRST_BUCKET
      - A special bucket name, used like a sentinel value.
      - Have moxi assign new connections to the first configured bucket.
        Confusingly, this is the default for the default_bucket_name.
    
    - NULL_BUCKET
      - A special bucket name, used like a sentinel value.
      - This refers to a special black-hole bucket, where any mutations
        result in SERVER_ERROR responses, and get's result in misses/END.
    
    When moxi is used standalone on a per-bucket basis, the
    default_bucket_name is left unspecifiedm, so FIRST_BUCKET.
    
    When moxi is used as a gateway proxy, default_bucket_name=default.
    
    Change-Id: Idfbdf68a75964fe84dee0c52e8d54fc90834cca8
    Reviewed-on: http://review.northscale.com/2824
    Reviewed-by: Dustin Sallings <dustin@spy.net>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Sep 28, 2010
  3. @steveyen

    bug 1998 - emitting test cmd-lines

    Change-Id: I291a1a4256eac0941bd34d28429b904640743297
    Reviewed-on: http://review.northscale.com/2823
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Sep 28, 2010
Commits on Sep 28, 2010
  1. @steveyen

    bug 2539 - use protocol correctly while caching host_ident

    In mcmux, the client can dynamically switch between ascii & binary
    protocols ("[A|B]:host:port" prefix).  That protocol was incorrectly
    getting ignored by the recent performance fix that was caching the
    host_ident strings.
    
    mcs_server_st_ident() API and implementation was changed to
    switch correctly to the right ident_a or ident_b buffer in
    this change.
    
    Change-Id: If0d6b7d643bd9611a48dd590454bc681adace610
    Reviewed-on: http://review.northscale.com/2810
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    Tested-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed with Matt Ingenthron Sep 28, 2010
  2. @steveyen

    bug 2529 - port 8080 to 8091

    Change-Id: Ic1a1d0ed6be0c8afc7c91b05487eefaf09dfe500
    Reviewed-on: http://review.northscale.com/2787
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    Tested-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed with Matt Ingenthron Sep 27, 2010
Commits on Sep 27, 2010
  1. @steveyen

    bug 2521 - avoid locks around front cache & set optimization

    Check the config/behavior flags before accessing locks, as an
    optimization when the front cache & set optimization are configured
    to be off.
    
    Change-Id: If6333fc63f08bca927dedfa17bc4eaa8eb3f06bd
    Reviewed-on: http://review.northscale.com/2771
    Tested-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Matt Ingenthron <matt@northscale.com>
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed with Matt Ingenthron Sep 27, 2010
  2. @ingenthr @steveyen

    Updated README to look after users first.

    Change-Id: I5f8ece48dd5b5b5872a8b2fee27ea6823c7f72a7
    Reviewed-on: http://review.northscale.com/2758
    Tested-by: Steve Yen <steve.yen@gmail.com>
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    ingenthr committed with steveyen Sep 27, 2010
  3. @steveyen

    added usr/pwd to usage() printf

    Change-Id: Idc25ed99c302c4de73f7a57851f575dacb85025e
    Reviewed-on: http://review.northscale.com/2741
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    Tested-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed Sep 26, 2010
  4. @steveyen

    bug 2499 - allow empty password

    Change-Id: Id3c8e1a264a2791a9adb3ca7530b31d4e3973c9b
    Reviewed-on: http://review.northscale.com/2743
    Tested-by: Matt Ingenthron <matt@northscale.com>
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed Sep 26, 2010
Commits on Sep 26, 2010
  1. @steveyen

    bug 1953 - cache host_ident string

    profiling showed sprintf of the host_ident_buf to be a hotspot, so
    caching it for reuse.
    
    Change-Id: I8860f5bc21298bdbe01fed6c101c9da3d9bc8e30
    Reviewed-on: http://review.northscale.com/2724
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    Tested-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed with Matt Ingenthron Sep 25, 2010
  2. @steveyen

    mcmux peer_protocol should take precedence over default protocol

    If the peer_protocol is set, it should take precedence
    over default behavior downstream protocol.
    
    Also, fixed some whitespace/coding conventions in this change.
    
    Change-Id: Ic185d8e2e3902298cfcb5e6107cd480e5d929652
    Reviewed-on: http://review.northscale.com/2715
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    Tested-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed with Matt Ingenthron Sep 25, 2010
  3. @steveyen

    unused param

    Change-Id: I6c3ef4971528548b25bd507dd76c6808b7db1ce8
    Reviewed-on: http://review.northscale.com/2713
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    Tested-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed with Matt Ingenthron Sep 25, 2010
  4. @steveyen

    bug 2316, clear timing histograms on stats proxy reset

    Change-Id: If444df59e25dcdc994d075cdec5291c4f58895ec
    Reviewed-on: http://review.northscale.com/2712
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    Tested-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed with Matt Ingenthron Sep 24, 2010
  5. @steveyen

    bug 2316, stats proxy reset

    Use the per-thread work inboxes to request worker threads
    to asynchronously clear their stats.
    
    Change-Id: Ia021bc516d9fef87fa89c4911627527c9cde33fb
    Reviewed-on: http://review.northscale.com/2711
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    Tested-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed with Matt Ingenthron Sep 20, 2010
Commits on Sep 25, 2010
  1. @steveyen

    tests pass when downstream_conn_max=0 (unlimited)

    Change-Id: If9b921d40fc35303de9d26db57e0d5c94ab705a7
    Reviewed-on: http://review.northscale.com/2710
    Reviewed-by: Steve Yen <steve.yen@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
    steveyen committed Sep 24, 2010
Commits on Sep 24, 2010
  1. @steveyen

    downstream_max / downstream_conn_max defaults

    200 is too optimistic, given the NUM_THREADS factor.
    
    Change-Id: Iea9e767090d7f486efc17afdf4b712280c6fc21c
    Reviewed-on: http://review.northscale.com/2696
    Tested-by: Matt Ingenthron <matt@northscale.com>
    Reviewed-by: Matt Ingenthron <matt@northscale.com>
    steveyen committed with Matt Ingenthron Sep 24, 2010
Something went wrong with that request. Please try again.