History log of /3.0.3-GA/ep-engine/ (Results 426 - 450 of 4208)
Revision (<<< Hide revision tags) (Show revision tags >>>)Date Author Comments
(<<< Hide modified files)
(Show modified files >>>)
b032b65028-Apr-2014 abhinav dangeti <abhinav@couchbase.com>

Revert "Fix race conidtion when deleting vbuckets"

This reverts commit ca2370576f7387416f3ae6b26f2b2b13539abdd9.

Change-Id: Ie3e376d6a1ddf36c1bfdb97ef6415d2561ea6105
Reviewed-on

Revert "Fix race conidtion when deleting vbuckets"

This reverts commit ca2370576f7387416f3ae6b26f2b2b13539abdd9.

Change-Id: Ie3e376d6a1ddf36c1bfdb97ef6415d2561ea6105
Reviewed-on: http://review.couchbase.org/36425
Tested-by: abhinav dangeti <abhinav@couchbase.com>
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>

show more ...

ea7d269123-Apr-2014 Trond Norbye <trond.norbye@gmail.com>

Allow build from source bundle

Change-Id: Ic0f6690334607411f9ce8cdf560c67fe5a4d234a
Reviewed-on: http://review.couchbase.org/36216
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>

Allow build from source bundle

Change-Id: Ic0f6690334607411f9ce8cdf560c67fe5a4d234a
Reviewed-on: http://review.couchbase.org/36216
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Tested-by: Trond Norbye <trond.norbye@gmail.com>

show more ...

e787298110-Apr-2014 Sriram Ganesan <sriram@couchbase.com>

MB-10411: Stream request response shouldn't contain start seqno

Whenever a stream request is issued with a start sequence number,
the response from the checkpoint and the backfill should

MB-10411: Stream request response shouldn't contain start seqno

Whenever a stream request is issued with a start sequence number,
the response from the checkpoint and the backfill shouldn't
contain the start sequence number

Change-Id: I93c8b7540bbaddb7e8ed0b81a0128c18662d2cef
Reviewed-on: http://review.couchbase.org/35578
Reviewed-by: Michael Wiederhold <mike@couchbase.com>
Tested-by: Michael Wiederhold <mike@couchbase.com>

show more ...

e1907a5b23-Apr-2014 Mike Wiederhold <mike@couchbase.com>

MB-10514: Wait for the backfill queue to drain before scheduling backfill

It is possible that after a vbucket takeover the entire backfill
queue has not drained. This means that if we sc

MB-10514: Wait for the backfill queue to drain before scheduling backfill

It is possible that after a vbucket takeover the entire backfill
queue has not drained. This means that if we schedule a backfill
may miss some items. This change causes the backfill task to wait
until all of the backfill items have been persisted before
scheduling the a new backfill.

Change-Id: I158f49aac40a932fd45431666e134e842df8ba92
Reviewed-on: http://review.couchbase.org/36255
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Tested-by: Michael Wiederhold <mike@couchbase.com>

show more ...

ca23705724-Apr-2014 Mike Wiederhold <mike@couchbase.com>

Fix race conidtion when deleting vbuckets

There is a race condition where we create, delete, and then create
a vbucket. The problem is that the the first create can finish
after the

Fix race conidtion when deleting vbuckets

There is a race condition where we create, delete, and then create
a vbucket. The problem is that the the first create can finish
after the second create is started. When this happens the code
thinks that the last create has finished even though the vbucket
deletion has not. This can cause items to fail to persist to disk.

Change-Id: I66e6450481acda2250bb7e13f921f3e9d20fbee1
Reviewed-on: http://review.couchbase.org/36254
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Reviewed-by: Michael Wiederhold <mike@couchbase.com>
Tested-by: Michael Wiederhold <mike@couchbase.com>

show more ...

a56b094525-Apr-2014 Sundar Sridharan <sundar.sridharan@gmail.com>

Force schedule snapshot task on bucket creation

Because we de-duplicate the snapshot task scheduling, it
is possible that a
Delete,Create gets re-ordered as Create,Delete because the

Force schedule snapshot task on bucket creation

Because we de-duplicate the snapshot task scheduling, it
is possible that a
Delete,Create gets re-ordered as Create,Delete because the
first create task was de-duplicated but the actual task runs
before the Delete task

Change-Id: I2d617ff2a63d4a6ff592ee03f883d5a9e75fd588
Reviewed-on: http://review.couchbase.org/36300
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Tested-by: abhinav dangeti <abhinav@couchbase.com>

show more ...

527009f425-Apr-2014 Sundar Sridharan <sundar.sridharan@gmail.com>

Serialize Shard Level Operations from ExecutorPool

Add a flag in every shard to indicate that a thread is
currently working on it.
In TaskQueue, threads check this flag first to ensu

Serialize Shard Level Operations from ExecutorPool

Add a flag in every shard to indicate that a thread is
currently working on it.
In TaskQueue, threads check this flag first to ensure that
only one thread picks up task from this shard
Mark certain tasks like VBDeleteTask and VBSnapshotTask
as needing serialization

Change-Id: Ib4f0ae14d90a3d95ea0a59531197da4f87f388d3
Reviewed-on: http://review.couchbase.org/36286
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Tested-by: abhinav dangeti <abhinav@couchbase.com>

show more ...

e550173824-Apr-2014 D Liao <liaodw@gmail.com>

MB-10155 Log OS error when CouchKVStore openDB fails

Currently we only log a generic "open failed" error.
Logging OS error gives us more info about the failure.

Change-Id: I5f84

MB-10155 Log OS error when CouchKVStore openDB fails

Currently we only log a generic "open failed" error.
Logging OS error gives us more info about the failure.

Change-Id: I5f84ee98dc8717fc7cd955573a53749e7da4ccd8
Reviewed-on: http://review.couchbase.org/36247
Tested-by: David Liao <david.liao@couchbase.com>
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>

show more ...

fe31ca7e31-Mar-2014 Mike Wiederhold <mike@couchbase.com>

MB-9898: Add byte sizes to upr responses

The sizes in this change represent the amount of bytes that each
upr message requires when being sent over a network connection.

Change-

MB-9898: Add byte sizes to upr responses

The sizes in this change represent the amount of bytes that each
upr message requires when being sent over a network connection.

Change-Id: Ic90bffc6ced9a7f9b4c6624289e0eb42e0caaaf6
Reviewed-on: http://review.couchbase.org/35374
Reviewed-by: Michael Wiederhold <mike@couchbase.com>
Tested-by: Michael Wiederhold <mike@couchbase.com>

show more ...

def96a4916-Apr-2014 Sriram Ganesan <sriram@couchbase.com>

MB-10702: Include the proper NRU bits in an UPR mutation

When sending an UPR mutation, we need to include the proper
NRU bits of the stored value.

Change-Id: If7fd01e34079161212

MB-10702: Include the proper NRU bits in an UPR mutation

When sending an UPR mutation, we need to include the proper
NRU bits of the stored value.

Change-Id: If7fd01e34079161212cf213fb62cafefa9248f7c
Reviewed-on: http://review.couchbase.org/35897
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Tested-by: Chiyoung Seo <chiyoung@couchbase.com>

show more ...

595a0ce923-Apr-2014 Mike Wiederhold <mike@couchbase.com>

MB-10850: Always add backfill remaining stat to stats vb-takeover

This stat is used by ns_server to check for backfill completion. It
was previously not sent as part of the stats if the

MB-10850: Always add backfill remaining stat to stats vb-takeover

This stat is used by ns_server to check for backfill completion. It
was previously not sent as part of the stats if the stream had
already transitioned to dead state. This change causes it to
always be sent even if the stream has transistioned to dead state.

Change-Id: I6b17df09cc8f6188c95591224a86d26089c9e4d9
Reviewed-on: http://review.couchbase.org/36165
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Tested-by: Michael Wiederhold <mike@couchbase.com>

show more ...

555b2ee723-Apr-2014 Chiyoung Seo <chiyoung.seo@gmail.com>

MB-10875 Fix to a bug in updating the persistence cursor offset.

We should NOT set the persistence cursor offset to be equal to the
number of items including meta items after getting all

MB-10875 Fix to a bug in updating the persistence cursor offset.

We should NOT set the persistence cursor offset to be equal to the
number of items including meta items after getting all the dirty
items, as we update the cursor offset while moving the cursor.

Change-Id: Ic80b3153ec7032a8d99d907db0a80e96fd3c5674
Reviewed-on: http://review.couchbase.org/36240
Reviewed-by: abhinav dangeti <abhinav@couchbase.com>
Tested-by: abhinav dangeti <abhinav@couchbase.com>

show more ...

d5a9a26222-Apr-2014 Mike Wiederhold <mike@couchbase.com>

MB-10919: Don't generate new failover entry during vbucket takeover

When a vbucket is transfered from one node to another we should
not generate a new failover entry in the failover log

MB-10919: Don't generate new failover entry during vbucket takeover

When a vbucket is transfered from one node to another we should
not generate a new failover entry in the failover log of the
new active vbucket.

Change-Id: I4e9a222ba67fbdb793035544527122e45bf0adc6
Reviewed-on: http://review.couchbase.org/36157
Reviewed-by: Michael Wiederhold <mike@couchbase.com>
Tested-by: Michael Wiederhold <mike@couchbase.com>

show more ...

09ba91c122-Apr-2014 Sundar Sridharan <sundar.sridharan@gmail.com>

MB-10875: fix flusher stat via checkpoint queue

In Checkpoint's queueDirty, check for checkpoint start message
to ensure that we do not decrement persistence cursor for an
existing i

MB-10875: fix flusher stat via checkpoint queue

In Checkpoint's queueDirty, check for checkpoint start message
to ensure that we do not decrement persistence cursor for an
existing item. We also should not consider checkpoint start/end
items in incrementing / decrementing a cursor offset.

Correct stat to totalEnqueued in queueBackfill.

Change-Id: Ibaa6d2b19de1034b4ca96ea8b5bd3ba78d08d4f4
Reviewed-on: http://review.couchbase.org/36154
Reviewed-by: abhinav dangeti <abhinav@couchbase.com>
Tested-by: abhinav dangeti <abhinav@couchbase.com>

show more ...

e81d80bc16-Apr-2014 Sundar Sridharan <sundar.sridharan@gmail.com>

MB-10849: add stat for priority queue sizes

Unblock QE tests that validate bucket priority UI setting.
'cbstats raw workload' now lists the sizes of futureQueue
and readyQueue as Inc

MB-10849: add stat for priority queue sizes

Unblock QE tests that validate bucket priority UI setting.
'cbstats raw workload' now lists the sizes of futureQueue
and readyQueue as Incoming Q and Outgoing Q respectively...
HiPrioQ_AuxIO:InQsize: 1
HiPrioQ_AuxIO:OutQsize: 0
HiPrioQ_NonIO:InQsize: 8
HiPrioQ_NonIO:OutQsize: 0
HiPrioQ_Reader:InQsize: 4
HiPrioQ_Reader:OutQsize: 0
LowPrioQ_Writer:OutQsize: 0
etc

Change-Id: I023b16e3cbf953eb110b5be6b9c5a630201fa1ed
Reviewed-on: http://review.couchbase.org/35913
Reviewed-by: abhinav dangeti <abhinav@couchbase.com>
Tested-by: Sundararaman Sridharan <sundar@couchbase.com>

show more ...

0e4754aa16-Apr-2014 Sundar Sridharan <sundar.sridharan@gmail.com>

MB-10027: use GlobalTask in flusher to avoid pool mutex

This fix saves reduces the number of lock contentions
in the write path and improves flusher performance as
seen in code profi

MB-10027: use GlobalTask in flusher to avoid pool mutex

This fix saves reduces the number of lock contentions
in the write path and improves flusher performance as
seen in code profiling report analysis

Use GlobalTask * pointer directly instead of ExTask to
avoid reference counting when there is no need

Change-Id: I39246bdc97e19d9e9bc974e99dab61bd4091ba4d
Reviewed-on: http://review.couchbase.org/35900
Reviewed-by: abhinav dangeti <abhinav@couchbase.com>
Tested-by: abhinav dangeti <abhinav@couchbase.com>

show more ...

90af56bb16-Apr-2014 abhinavdangeti <abhinav@couchbase.com>

MB-10828: Handling a possible segmentation fault

- If vb is NULL, there could be a segmentation
fault when using the object to retrieve the vbucket
hash table, etc.
- Return if v

MB-10828: Handling a possible segmentation fault

- If vb is NULL, there could be a segmentation
fault when using the object to retrieve the vbucket
hash table, etc.
- Return if vb is NULL

Change-Id: Ic646a529f9ca03b170629a43edb3fc3dd717c8d6
Reviewed-on: http://review.couchbase.org/35891
Reviewed-by: Sundararaman Sridharan <sundar@couchbase.com>
Tested-by: Sundararaman Sridharan <sundar@couchbase.com>

show more ...

4640c14c16-Apr-2014 abhinavdangeti <abhinav@couchbase.com>

MB-10849: Adding bucket priority to cbstats

Change-Id: I6d44598688a85a5ae46d795791ee6eef87ea6051
Reviewed-on: http://review.couchbase.org/35908
Reviewed-by: Sundararaman Sridharan <s

MB-10849: Adding bucket priority to cbstats

Change-Id: I6d44598688a85a5ae46d795791ee6eef87ea6051
Reviewed-on: http://review.couchbase.org/35908
Reviewed-by: Sundararaman Sridharan <sundar@couchbase.com>
Tested-by: Sundararaman Sridharan <sundar@couchbase.com>

show more ...

0cb591f915-Apr-2014 Sundar Sridharan <sundar.sridharan@gmail.com>

MB-10027 reduce ExecutorPool mutex calls in doneWork

It is sufficient for the doneWork routine to call the
ExecutorPool mutex just once after it is done with a task type

This re

MB-10027 reduce ExecutorPool mutex calls in doneWork

It is sufficient for the doneWork routine to call the
ExecutorPool mutex just once after it is done with a task type

This reduces the lock contention and decreases cpu usage as
per code profiling

Change-Id: I024f3180d67270cb03f2382de5165af329da47e4
Reviewed-on: http://review.couchbase.org/35743
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Tested-by: Sundararaman Sridharan <sundar@couchbase.com>

show more ...

dc81101f15-Apr-2014 Mike Wiederhold <mike@couchbase.com>

Notify memcached when an upr stream is set to dead state

If a stream is currently not sending any items then setting the
stream to dead state (which adds an end stream message) will not

Notify memcached when an upr stream is set to dead state

If a stream is currently not sending any items then setting the
stream to dead state (which adds an end stream message) will not
cause the memcached worker thread to visit the upr stream in order
to send this message. As a result we need to notify memcached when
streams die in order to make sure that consumers don't get stuck
waiting for the end stream message.

Change-Id: Iad46cd18d6d911e8826c0ff235924a8403a90c54
Reviewed-on: http://review.couchbase.org/35748
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Tested-by: Michael Wiederhold <mike@couchbase.com>

show more ...

ed33ae8d09-Apr-2014 Mike Wiederhold <mike@couchbase.com>

MB-10794: Set the open checkpoint seqno properly when backfill ends

When we initially create a replica vbucket we create an open
checkpoint with checkpoint id 0 and this checkpoint will

MB-10794: Set the open checkpoint seqno properly when backfill ends

When we initially create a replica vbucket we create an open
checkpoint with checkpoint id 0 and this checkpoint will start
at seqno 1. Since this replica might receive backfill items when
data is replicated to it we need to properly reassign the seqno
of the start checkpoint item before once backfill is finished.

Change-Id: I9c224f6b0d517fe500b733ce84cc00366b702a2c
Reviewed-on: http://review.couchbase.org/35708
Reviewed-by: Chiyoung Seo <chiyoung@couchbase.com>
Tested-by: Michael Wiederhold <mike@couchbase.com>

show more ...

44fb2b8a14-Apr-2014 Trond Norbye <trond.norbye@gmail.com>

Minor cleanup

Change-Id: I836767aecd72becf213845c5be18d39afa69ccff
Reviewed-on: http://review.couchbase.org/35681
Reviewed-by: Michael Wiederhold <mike@couchbase.com>
Tested-by:

Minor cleanup

Change-Id: I836767aecd72becf213845c5be18d39afa69ccff
Reviewed-on: http://review.couchbase.org/35681
Reviewed-by: Michael Wiederhold <mike@couchbase.com>
Tested-by: Trond Norbye <trond.norbye@gmail.com>

show more ...

1e951c3811-Apr-2014 Sundar Sridharan <sundar.sridharan@gmail.com>

MB-10539: use less than eq comparison in wake()

ExecutorPool::wake() was wrongly using less_tv()
in its wake routine, resulting in un-woken tasks
This was causing large delays in sch

MB-10539: use less than eq comparison in wake()

ExecutorPool::wake() was wrongly using less_tv()
in its wake routine, resulting in un-woken tasks
This was causing large delays in scheduling
Fix by using less_eq_tv

Change-Id: I52e97c6308c70cfc415dd731a50dc8e00a7d834a
Reviewed-on: http://review.couchbase.org/35586
Reviewed-by: abhinav dangeti <abhinav@couchbase.com>
Reviewed-by: Sundararaman Sridharan <sundar@couchbase.com>
Tested-by: Sundararaman Sridharan <sundar@couchbase.com>

show more ...

d8a3dade10-Apr-2014 Sundar Sridharan <sundar.sridharan@gmail.com>

MB-10539-debug: avoid ExecutorPool mutex to snooze in bgfetcher

Use of ExecutorPool->snooze() unnecessarily increases lock
contention within the shared thread pool in bgfetcher

MB-10539-debug: avoid ExecutorPool mutex to snooze in bgfetcher

Use of ExecutorPool->snooze() unnecessarily increases lock
contention within the shared thread pool in bgfetcher

Change-Id: Ia0a0fed93f170d6e504a24f54c2907b32d7851a2
Reviewed-on: http://review.couchbase.org/35577
Reviewed-by: abhinav dangeti <abhinav@couchbase.com>
Tested-by: Sundararaman Sridharan <sundar@couchbase.com>

show more ...

cee0935710-Apr-2014 Sundar Sridharan <sundar.sridharan@gmail.com>

simplify snooze: remove first run and GMT stuff

Change-Id: I21cd48b33cf565bc19e2545b2ced2eb6f8627a5f
Reviewed-on: http://review.couchbase.org/35580
Reviewed-by: abhinav dangeti <abhi

simplify snooze: remove first run and GMT stuff

Change-Id: I21cd48b33cf565bc19e2545b2ced2eb6f8627a5f
Reviewed-on: http://review.couchbase.org/35580
Reviewed-by: abhinav dangeti <abhinav@couchbase.com>
Reviewed-by: Sundararaman Sridharan <sundar@couchbase.com>
Tested-by: Sundararaman Sridharan <sundar@couchbase.com>

show more ...

1...<<11121314151617181920>>...169