Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with
or
.
Download ZIP
Commits on Aug 11, 2012
  1. @bcui6611

    MB-100:Merge branch 'branch-181' into branch-18

    bcui6611 authored
    * branch-181:
      MB-5895: consider upgrading vb map history at rebalance start
      MB-5678: cbcollect_info invokes cbstats with its changed params
    
    Change-Id: I292cedb61a9bf00676b73ab864c6e6c29647ff41
Commits on Aug 10, 2012
  1. @alk

    MB-5895: consider upgrading vb map history at rebalance start

    alk authored Farshid Ghods committed
    When 1.8.1 is added to cluster of 1.8.0 nodes there's initially no
    vbucket map history. This happens because 1.8.1 currently considers
    upgrading config only when offline upgrading. But for vbucket map
    history which is global config element, we need to do that with actual
    vbucket maps (after joining cluster) but before rebalance, so that
    swap rebalance works.
    
    Change-Id: I1757f869b5542b06ccf0951d8827e21963f1c0cc
    Reviewed-on: http://review.couchbase.org/19387
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
Commits on Jun 25, 2012
  1. @steveyen

    MB-5678: cbcollect_info invokes cbstats with its changed params

    steveyen authored Farshid Ghods committed
    Change-Id: I2639efc0604386ad9c66c35e58f2d082bf0d9a7e
    Reviewed-on: http://review.couchbase.org/17595
    Tested-by: Farshid Ghods <farshid@couchbase.com>
    Reviewed-by: Farshid Ghods <farshid@couchbase.com>
Commits on Jun 23, 2012
  1. @steveyen

    Merge remote branch 'origin/branch-181' into branch-18

    steveyen authored
    * origin/branch-181:
      MB-5657: create cbcollect_info tasks after main() has fixed PATH
      MB-5650: try harder to find escript tool and config.dat
      MB-5645: cbbrowse_logs uses .., not double dirname
      MB-5646: cbcollect_info fixes, including cbstats raw allocator
      MB-5307: improved ns_config logging around clear call
      CBD-364: updated enterprise product registration
      MB-5602: consider buckets' servers list when computing down nodes
      MB-5308: don't set 'for' attr. of label for checkbox inside it
      MB-5579: made phone home send it's stuff via POST
      MB-4703: Make cbbrowse_logs command work on Mac OS
      MB-5565 Check cpu_utilization_rate sanity correctly.
      MB-5307 Check moxi_port for conflicts with dedicated_port.
    
    Change-Id: I51f3264e199b9197827c3e49b4eed9d9a82e82bf
  2. @steveyen

    MB-5657: create cbcollect_info tasks after main() has fixed PATH

    steveyen authored
    Change-Id: I153f624e199b9197827c3e49b4eed9d9a82e82bf
    Reviewed-on: http://review.couchbase.org/17562
    Reviewed-by: Bin Cui <bin.cui@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
  3. @steveyen

    MB-5650: try harder to find escript tool and config.dat

    steveyen authored
    Try the usual places for linux, windows and OSX.
    
    Windows is special because "escript" would be automatically
    incorrectly executed as "escript.escript".
    
    Change-Id: Ieaa019ea0aad1dd57e30eb78907207ab5dabd0ba
    Reviewed-on: http://review.couchbase.org/17557
    Reviewed-by: Bin Cui <bin.cui@gmail.com>
    Tested-by: Bin Cui <bin.cui@gmail.com>
Commits on Jun 22, 2012
  1. @steveyen

    MB-5645: cbbrowse_logs uses .., not double dirname

    steveyen authored
    OSX needs absolute path, and its logs live in a different place.
    
    Change-Id: I4c342cfefa98449237c4c2710f6f1ad306ae0a3d
    Reviewed-on: http://review.couchbase.org/17527
    Tested-by: Jens Alfke <jens@couchbase.com>
    Reviewed-by: Jens Alfke <jens@couchbase.com>
  2. @steveyen

    MB-5646: cbcollect_info fixes, including cbstats raw allocator

    steveyen authored
    Besides adding 'cbstats raw allocator', cbstats might not also accept
    '-a' flag anymore, so we call it twice, with and without, just in
    case.
    
    Additionally, fixed trying to get the dbdir on linux.
    
    Change-Id: Icdf657425b82f2e78ee270c26c2f6689532f9002
    Reviewed-on: http://review.couchbase.org/17523
    Reviewed-by: Bin Cui <bin.cui@gmail.com>
    Tested-by: Steve Yen <steve.yen@gmail.com>
  3. @alk

    MB-5307: improved ns_config logging around clear call

    alk authored Farshid Ghods committed
    Change-Id: I356ea4b55d1d56ccdfbdf399720a0aadd8b7d3d9
    Reviewed-on: http://review.couchbase.org/17520
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Reviewed-by: Farshid Ghods <farshid@couchbase.com>
Commits on Jun 19, 2012
  1. @alk

    CBD-364: updated enterprise product registration

    alk authored Farshid Ghods committed
    As requested by Dipti.
    
    Change-Id: I708302d354169bed84d7e000b0c60988236627bf
    Reviewed-on: http://review.couchbase.org/17374
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
  2. @alk

    MB-5602: consider buckets' servers list when computing down nodes

    alk authored Farshid Ghods committed
    Previous code assumed that all nodes need to have all buckets, but
    that's not the case. Particularly, for partially rebalanced in or
    partially rebalanced out nodes.
    
    Fix is, naturally, to look exactly which buckets are expected to be up
    and running on each node.
    
    Change-Id: Iec3630a86fc92f213c32cf84f461388e4f5004f1
    Reviewed-on: http://review.couchbase.org/17372
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
  3. @alk

    MB-5308: don't set 'for' attr. of label for checkbox inside it

    alk authored Farshid Ghods committed
    Apparently that's not working with Chrome 19. Removing "for" attribute
    makes clicking on label work, i.e. it'll toggle checkbox as expected.
    
    Change-Id: I676e69c150f9f1fe891fbf547702ea4b1395d302
    Reviewed-on: http://review.couchbase.org/17339
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
  4. @alk

    MB-5579: made phone home send it's stuff via POST

    alk authored Farshid Ghods committed
    JSONP cannot be used to deliver large amounts of stuff. Instead we're
    POSTing stuff via hidden iframe's form submit, and we're getting
    information about new versions by doing short JSONP GET.
    
    Update notifications UI is also refactored a bit to do be more
    cellfull. Particularly it now draws nice spinner when it's fetching
    updates data.
    
    As can be seen in fake-phone-home script. POSTed data is still json,
    that's sent as "stats" field that's traditional form
    url-encoding. That should be same as before, just uses POST instead of
    GET.
    
    It's possible to match POST & GET by looking at query string parameter
    launchID that's unique identifier of phone home call. _Both_ GET &
    POST will have this query string parameter and it will be
    same. launchID as can be seen in code is built as concatenation of
    cluster's uuid, current epoch milliseconds and random 16-bit value. So
    not only it's unique, but even if we lost POST we can at least count
    distinct clusters in the wild (who agreed to send phone home stats).
    
    GET now only sends launchID as described above and version. I assume
    that should be enough to send back something meaningful. In worst case
    server-side implementation of GET can wait until POST is received and
    extract any additional info it may need from there, then wake up GET
    request and make it reply something.
    
    Change-Id: I34de54a705dda3495fe64e8da86d12e2798392aa
    Reviewed-on: http://review.couchbase.org/17338
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Dustin Sallings <dustin@spy.net>
Commits on Jun 18, 2012
  1. @snej

    MB-4703: Make cbbrowse_logs command work on Mac OS

    snej authored Farshid Ghods committed
    Find path to Couchbase root relative to script,
    then use that as the base for search paths.
    
    Change-Id: Iecb59df358afa3ef0f81b37c9c2001a5a5a798d7
    [alk@tut.by: rebased for 1.8.1]
    [alk@tut.by: renamed to cbbrowse_logs and dont use sed anymore]
    Signed-off-by: Aliaksey Kandratsenka <alk@tut.by>
    Reviewed-on: http://review.couchbase.org/17327
    Reviewed-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Jens Alfke <jens@couchbase.com>
    Tested-by: Farshid Ghods <farshid@couchbase.com>
    Reviewed-by: Farshid Ghods <farshid@couchbase.com>
Commits on Jun 14, 2012
  1. @aartamonau

    MB-5565 Check cpu_utilization_rate sanity correctly.

    aartamonau authored Farshid Ghods committed
    Change-Id: Idde0054fa82315793cb446a5f8b0747948bdfa1a
    Reviewed-on: http://review.couchbase.org/17268
    Reviewed-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Tested-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
Commits on Jun 13, 2012
  1. @aartamonau @alk

    MB-5307 Check moxi_port for conflicts with dedicated_port.

    aartamonau authored alk committed
    Change-Id: I136eb9acb71e581f5422d206436b0318e233a0a6
    Reviewed-on: http://review.couchbase.org/17236
    Tested-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Reviewed-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
  2. @alk

    Merge remote-tracking branch 'origin/branch-181' into branch-18

    alk authored
    * origin/branch-181:
      MB-5546: raised some timeouts to cope with some paging
      MB-5535:bp:use gen_udp instead for address changed alert testing
    
    Change-Id: Ie048f2bef1e580ee645991df211874e3f1852481
  3. @alk @aartamonau

    MB-5546: raised some timeouts to cope with some paging

    alk authored aartamonau committed
    When system is on the limits some ops take longer. Apparently our
    experiments told us making some timeouts twice longer helps.
    
    NOTE: ticket also requests raising ns_memcached_connected timeout
    which I'd like to avoid. Thus that's the only timeout that's not
    raised out of timeout raises requested.
    
    Change-Id: If3b45c6fbbb566315602101d677dae90d2e642ba
    Reviewed-on: http://review.couchbase.org/17230
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
  4. @alk @aartamonau

    MB-5535:bp:use gen_udp instead for address changed alert testing

    alk authored aartamonau committed
    Hopefully it'll fix our rare false positives from this code caused by
    some system limits.
    
    gen_tcp:listen is doing full bind + listen and apparently is not
    passing SO_REUSEADDR. That's our theory at least.
    
    With UDP this is not needed and is much lighter. I.e. it's not
    creating those TIME_WAIT tcp sockets.
    
    Change-Id: I1157062f002a51bbef4cf41d231b8b7853ccc920
    Originally-Reviewed-on: http://review.couchbase.org/15472
    Reviewed-on: http://review.couchbase.org/17189
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
Commits on Jun 12, 2012
  1. @alk

    Merge remote-tracking branch 'origin/branch-181' into branch-18

    alk authored
    * origin/branch-181:
      MB-5307: display avg disk commit time in seconds
    
    Change-Id: I653d26a4b644d6e7a949a9fbca9e1ea7bbe8af66
  2. @alk

    MB-5307: display avg disk commit time in seconds

    alk authored alk committed
    Otherwise it shows quite inconvenient millions.
    
    Change-Id: Ib7ad183ec21cfe76130ba3d40ef1f93fac7a1138
    Reviewed-on: http://review.couchbase.org/17163
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
  3. @alk

    Merge remote-tracking branch 'origin/branch-181' into branch-18

    alk authored
    * origin/branch-181:
      CBD-172: hide bg_wait_time and disk_commit_time in UI
      CBD-172: gather and display average disk update time
      CBD-364: got rid of "plant tree" option for enterprise edition
      MB-5513: wait isasl.pw refresh before flagging warmup completion
      MB-5484: updated disk creates/updates per sec descriptions
      MB-5500 Show all the gauges as N/A if anything goes wrong.
      MB-5308 Bound unallocated memory by zero.
      MB-5307: fixed dialyzer complaints
      MB-5307: fixed getting of not ready to replicate from vbuckets
      MB-5307: only consider stopping rebalance for normal janitor runs
      MB-5307: increased size of master activity events history
      MB-5307: [bp] Fix cluster overview graphs showing blank data
      MB-5307 Pass global port to memcached via -p.
      MB-5307: parallelize sync_shutdown_many for less latency
      MB-5307: try to wait for all built replicas has same checkpoint
      MB-5307: configure-ized max count of concurrently moved vbuckets
      MB-5307: correctly account for very heavy memcached requests
    
    Conflicts:
    	src/mc_client_binary.erl
    	src/ns_janitor.erl
    
    Change-Id: I7c8badb8a238c83ed144112e8686d23561ab254a
  4. @alk @aartamonau

    CBD-172: hide bg_wait_time and disk_commit_time in UI

    alk authored aartamonau committed
    But allow access via editing URL
    
    Change-Id: Id82619b00cd425f695c88a6fa7fde83edeb6ac5b
    Reviewed-on: http://review.couchbase.org/17110
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
  5. @alk @aartamonau

    CBD-172: gather and display average disk update time

    alk authored aartamonau committed
    Instead of write_seeks_distance which ended up being less useful than
    initially thought.
    
    Change-Id: Ic3c3bdbd1ef99c6b4e31bb51c98780de0d04b313
    Reviewed-on: http://review.couchbase.org/17109
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
  6. @alk @aartamonau

    CBD-364: got rid of "plant tree" option for enterprise edition

    alk authored aartamonau committed
    I was told we can't do that anymore.
    
    Change-Id: If13f721ed32c8a092ec47d2e01f7662db6d6d5d1
    Reviewed-on: http://review.couchbase.org/17108
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
  7. @alk @aartamonau

    MB-5513: wait isasl.pw refresh before flagging warmup completion

    alk authored aartamonau committed
    With this change bucket will only be marked as warmed up after bucket
    config is reflected in isasl.pw. So that ebucketmigrators and other
    clients can actually auth against bucket after janitor completed
    bringing it up.
    
    Change-Id: I9d58ab1d66e1b480c934896b5bba1755bcc6171d
    Reviewed-on: http://review.couchbase.org/17104
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
  8. @alk @aartamonau

    MB-5484: updated disk creates/updates per sec descriptions

    alk authored aartamonau committed
    As pointed out in ticket.
    
    Change-Id: I46f482920a42443945698df98a484b92d74b53df
    Reviewed-on: http://review.couchbase.org/17103
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
  9. @aartamonau @alk

    MB-5500 Show all the gauges as N/A if anything goes wrong.

    aartamonau authored alk committed
    Change-Id: If3682f4e0841a9c5fe1b68a6cec7926c0331a7a7
    Reviewed-on: http://review.couchbase.org/17086
    Tested-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Reviewed-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
  10. @aartamonau @alk

    MB-5308 Bound unallocated memory by zero.

    aartamonau authored alk committed
    It could be less than zero if some of the nodes were down.
    
    Change-Id: I8acff37db9be223639bd2ea8a43f367465220970
    Reviewed-on: http://review.couchbase.org/17087
    Tested-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Reviewed-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
  11. @alk @aartamonau

    MB-5307: fixed dialyzer complaints

    alk authored aartamonau committed
    I couldn't understand what it was trying to tell me and if it was
    genuine problem. Anyway new way is a bit more correct as we'll raise
    missing_checkpoint_stat exception even if first checkpoint is 0.
    
    Change-Id: I2c9ab5a1511f21d6eff81f0e4da8604c8d198984
    Reviewed-on: http://review.couchbase.org/17082
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
  12. @alk @aartamonau

    MB-5307: fixed getting of not ready to replicate from vbuckets

    alk authored aartamonau committed
    Implementation of get_open_checkpoint_ids/2 was returning dict with
    stat names as keys instead of vbucket numbers. That was causing us to
    never detect any vbuckets as not ready to replicate from.
    
    I decided to optimize things a bit further by asking specifically for
    vbuckets with open checkpoint 0 and by trying to minimize produced
    garbage that was typical for previous implementation that was building
    dict.
    
    Change-Id: Ice165d2161348bc9f92a99d8340a8ab0bc9f1f5e
    Reviewed-on: http://review.couchbase.org/17081
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
  13. @alk @aartamonau

    MB-5307: only consider stopping rebalance for normal janitor runs

    alk authored aartamonau committed
    We also run janitor as part of failover. And that janitor cleanup run
    is made on orchestrator process itself. So trying to ask orchestrator
    if rebalance is running times out as we're asking ourselves.
    
    The fix is only consider stopping rebalance when it's normal
    scheduled janitor run.
    
    Change-Id: I8967a12343cedc51bbf1e2013f7c4ed97a573e3e
    Reviewed-on: http://review.couchbase.org/17080
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
  14. @alk @aartamonau

    MB-5307: increased size of master activity events history

    alk authored aartamonau committed
    With recent changes we're now routinely cannot hold even single bucket
    rebalance in our 'archive'. Hopefully making it larger helps.
    
    Also it's now somewhat runtime-configurable.
    
    Change-Id: Id66d204ebe217b8769404678a5f90e7041f39a95
    Reviewed-on: http://review.couchbase.org/17079
    Tested-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
    Reviewed-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
Commits on Jun 11, 2012
  1. @aartamonau @alk

    MB-5307: [bp] Fix cluster overview graphs showing blank data

    aartamonau authored alk committed
    Originally reviewed on http://review.couchbase.org/12196
    
    Change-Id: Ic330da56cd50cbcca13df0f2dd73c139969f8aec
    Reviewed-on: http://review.couchbase.org/17049
    Tested-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Reviewed-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
  2. @aartamonau @alk

    MB-5307 Pass global port to memcached via -p.

    aartamonau authored alk committed
    Change-Id: Iafd3c2371863b42581ea1dcd82545cc5dcb862f6
    Reviewed-on: http://review.couchbase.org/17048
    Tested-by: Aliaksey Artamonau <aliaksiej.artamonau@gmail.com>
    Reviewed-by: Aliaksey Kandratsenka <alkondratenko@gmail.com>
Something went wrong with that request. Please try again.