History log of /1.7.1/ep-engine/ (Results 1 - 25 of 1707)
Revision (<<< Hide revision tags) (Show revision tags >>>)Date Author Comments
(<<< Hide modified files)
(Show modified files >>>)
Revision tags: v4.6.3, v4.6.4, v4.6.2, v4.6.1, v4.6.0, v4.6.0-DP, v4.5.1, v4.1.2, v3.1.6, v4.5.0, v4.1.1, v3.1.5, v3.1.4, v3.1.3, v4.1.0, v3.1.2, v3.1.1, v4.0.0, v3.1.0, v3.0.2, v3.0.0, v2.5.1, 2.1.1r, 2.1.0r, 2.0.2r, 2.0.1-macosx, 2.0.1-linux, 2.0.0-couchbase, 1.8.1-MB-5845, 2.0.0r, 1.8.1, 1.8-MB-4901, 1.8-MB-4738, 1.8.1r, 1.8.0, 1.8.0r, 1.6.5.6, 1.7.3r, 1.7.2, 1.7.1.1
8d17d20523-Jul-2011 Dustin Sallings <dustin@spy.net>

Merge remote-tracking branch 'origin/master' into branch-171

* origin/master:
Remove debug query print from mbrestore.
Partial restore functionality in mbrestore.
Fix test

Merge remote-tracking branch 'origin/master' into branch-171

* origin/master:
Remove debug query print from mbrestore.
Partial restore functionality in mbrestore.
Fix test suite compilation errors
Fix failing ep-engine atomic_queue_test
Fix failing ep-engine checkpoint unit test
Clear engine specific data before SyncListener object destruction
Implemented __repr__ in TapDescriptor
Implemented a binary multi-set in mc_bin_client.
Add timeout argument to getl operation in mc_bin_client
set_vbucket_state should encode the state into extras

Conflicts:
management/mc_bin_client.py
t/checkpoint_test.cc

Change-Id: I952d9503ce197248142fe57db5b7fc66e1302d09

show more ...


977b8b6d22-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Revert "Check the open checkpoint id in CheckpointManager::hasNext()"

This reverts commit edb76a512b636bd4be5b5135ddb2aa30e4e99877.

Change-Id: If3c501d708475c24fbbc63f6a97123eaeada8

Revert "Check the open checkpoint id in CheckpointManager::hasNext()"

This reverts commit edb76a512b636bd4be5b5135ddb2aa30e4e99877.

Change-Id: If3c501d708475c24fbbc63f6a97123eaeada8036
Reviewed-on: http://review.couchbase.org/8277
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
Reviewed-by: Chiyoung Seo <chiyoung.seo@gmail.com>

show more ...

2b4c2ab922-Jul-2011 Chad Kouse <chad.kouse@gmail.com>

Remove debug query print from mbrestore.

Change-Id: Ia034ce84b014199e2c2d3110a1d0c7fc0d757652
Reviewed-on: http://review.couchbase.org/8276
Reviewed-by: Dustin Sallings <dustin@spy.n

Remove debug query print from mbrestore.

Change-Id: Ia034ce84b014199e2c2d3110a1d0c7fc0d757652
Reviewed-on: http://review.couchbase.org/8276
Reviewed-by: Dustin Sallings <dustin@spy.net>
Tested-by: Dustin Sallings <dustin@spy.net>

show more ...

6ac9f84122-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Check the open checkpoint id in CheckpointManager::hasNext()

If the vbucket's open checkpoint id is 0, this means that the
vbucket has been reset and does not have any items yet.

Check the open checkpoint id in CheckpointManager::hasNext()

If the vbucket's open checkpoint id is 0, this means that the
vbucket has been reset and does not have any items yet.

Change-Id: I457cbaef72fb8d97e7c8b186c867b84ac1bf042f
Reviewed-on: http://review.couchbase.org/8244
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
Reviewed-by: Steve Yen <steve.yen@gmail.com>
Reviewed-by: Srinivas Vadlamani <srinivas@couchbase.com>

show more ...

3c826e8e21-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Send NOOP messages to the eBucketMigrator periodically.

The upstream server should send TAP NOOP messages to
the eBucketMigrator periodically even in non-idle state because
each disk

Send NOOP messages to the eBucketMigrator periodically.

The upstream server should send TAP NOOP messages to
the eBucketMigrator periodically even in non-idle state because
each disk backfill task might take a long time in case of the long
waiting backfill dispatcher queue. Otherwise, the eBucketMigrator
will get timeout if it does not receive any messages from the
upstream server for 300 sec. Note that TAP NOOPs are just dummy
messages and consumed by the local eBucketMigrator only, but not
propagated to the downstream server. Therefore, this change won't
cause any bad overhead.

Change-Id: Idfe06579d29a1832b3779a636b8a30a64759d530
Reviewed-on: http://review.couchbase.org/8210
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
Reviewed-by: Steve Yen <steve.yen@gmail.com>

show more ...

618eb50121-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Stat for the total number of TAP_NOOP messages sent.

Change-Id: Ia3bcbb4f3ae250c495538e6c47368872c436c469
Reviewed-on: http://review.couchbase.org/8207
Tested-by: Chiyoung Seo <chiyo

Stat for the total number of TAP_NOOP messages sent.

Change-Id: Ia3bcbb4f3ae250c495538e6c47368872c436c469
Reviewed-on: http://review.couchbase.org/8207
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
Reviewed-by: Srinivas Vadlamani <srinivas@couchbase.com>

show more ...

c209e84f21-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Skip TAP cursor registration if backfill runs for a given vbucket

If backfill is currently running for a given vbucket, we can simply
skip the TAP cursor registration because the backfil

Skip TAP cursor registration if backfill runs for a given vbucket

If backfill is currently running for a given vbucket, we can simply
skip the TAP cursor registration because the backfill task already
sets the cursor to the beginning of the current open checkpoint when
scheduled.

Change-Id: Ifb36ccc5f21ba38520e6431bbdef4f3b81055984
Reviewed-on: http://review.couchbase.org/8180
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
Reviewed-by: Srinivas Vadlamani <srinivas@couchbase.com>

show more ...

50beb62221-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Support backward compatibility to 1.7 version for TAP backfill.

1.7 TAP producer does not send BACKFILL_END messages for individual
vbuckets. This causes the backward compatibility issue

Support backward compatibility to 1.7 version for TAP backfill.

1.7 TAP producer does not send BACKFILL_END messages for individual
vbuckets. This causes the backward compatibility issues in 1.7.1
TAP consumer. Therefore, the 1.7.1. TAP consumer should detect the
backfill completion for each vbucket by looking at checkpoint_start
messages recevied from 1.7 TAP producer.

Change-Id: I8834f94cb1deac9403191d70266ee580fa52d73d
Reviewed-on: http://review.couchbase.org/8179
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
Reviewed-by: Srinivas Vadlamani <srinivas@couchbase.com>

show more ...

d215f42317-Jun-2011 Perry Krug <perry@couchbase.com>

Partial restore functionality in mbrestore.

Hey guys, I did some overhaul on our mbrestore script (find it attached).

I added options for:
-k KEY, --key=KEY Key name to re

Partial restore functionality in mbrestore.

Hey guys, I did some overhaul on our mbrestore script (find it attached).

I added options for:
-k KEY, --key=KEY Key name to restore
-d DATA, --data=DATA Server side value to match
-i ID, --id=ID vbucketID to restore
-v VERBOSE, --verbose=VERBOSE
verbose logging

Change-Id: I8ba883a88bd8c847feb566ec64894bcaf33cc2d2
Reviewed-on: http://review.couchbase.org/7063
Tested-by: Perry Krug <perry@couchbase.com>
Reviewed-by: Perry Krug <perry@couchbase.com>
Reviewed-by: Dustin Sallings <dustin@spy.net>

show more ...

908007ae15-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Stat that represents the memory usage to store resident keys' values

Change-Id: Id83501c9a6437e48d30da06407da98d7e174b8ef
Reviewed-on: http://review.couchbase.org/8009
Reviewed-by: S

Stat that represents the memory usage to store resident keys' values

Change-Id: Id83501c9a6437e48d30da06407da98d7e174b8ef
Reviewed-on: http://review.couchbase.org/8009
Reviewed-by: Srinivas Vadlamani <srinivas@couchbase.com>
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>

show more ...

790ff91b13-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Pause backfill tasks if memory usage is too high.

Change-Id: Iabf7d6a915e7bdb084c62d3bd49148620b161e63
Reviewed-on: http://review.couchbase.org/7967
Reviewed-by: Matt Ingenthron <mat

Pause backfill tasks if memory usage is too high.

Change-Id: Iabf7d6a915e7bdb084c62d3bd49148620b161e63
Reviewed-on: http://review.couchbase.org/7967
Reviewed-by: Matt Ingenthron <matt@couchbase.com>
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>

show more ...

39c6c2d813-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Deregister the TAP cursor when backfill is required.

If backfill is required for a given vbucket, we can simply remove
the corresponding TAP cursor from that vbucket's checkpoint
dat

Deregister the TAP cursor when backfill is required.

If backfill is required for a given vbucket, we can simply remove
the corresponding TAP cursor from that vbucket's checkpoint
datastructure. Later, we set the cursor to the beginning of the open
checkpoint when backfill is actually scheduled.

Change-Id: I8c19ce97e4c459ba300ad2d7d050439b57664939
Reviewed-on: http://review.couchbase.org/7966
Reviewed-by: Matt Ingenthron <matt@couchbase.com>
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>

show more ...

6da82ced12-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Item's class overhead should be considered in memory accounting

Item instance is used in exchanging item information between
memcached frontend and ep-engine as well as persisting an ite

Item's class overhead should be considered in memory accounting

Item instance is used in exchanging item information between
memcached frontend and ep-engine as well as persisting an item
into disk. Therefore, Item instance's memory overhead should be
included in memory accounting.

Change-Id: I90b823ada9d759fd68895c464f9943133cb1bc71
Reviewed-on: http://review.couchbase.org/7920
Reviewed-by: Matt Ingenthron <matt@couchbase.com>
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>

show more ...

b572011a12-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Use RCPtr in accessing Blob and QueuedItem instances.

From profiling results, we found that using RCPtr consumes less
memory than shared_ptr in accessing Blob and QueuedItem instances.

Use RCPtr in accessing Blob and QueuedItem instances.

From profiling results, we found that using RCPtr consumes less
memory than shared_ptr in accessing Blob and QueuedItem instances.

Change-Id: If4a7ef336374cb5f3a5bf82a28931b55f1c839b7
Reviewed-on: http://review.couchbase.org/7919
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>

show more ...

25429bf812-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Blob class's memory overhead should be included in memory accounting

Change-Id: I30b87b31466a940398bcdf35acf981d77facd56e
Reviewed-on: http://review.couchbase.org/7918
Tested-by: Chi

Blob class's memory overhead should be included in memory accounting

Change-Id: I30b87b31466a940398bcdf35acf981d77facd56e
Reviewed-on: http://review.couchbase.org/7918
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>

show more ...

82050d5d12-Jul-2011 Filipe David Manana <fdmanana@apache.org>

Fix test suite compilation errors

This fixes some missing function declaration errors:

CXX ep_testsuite_la-ep_testsuite.lo
cc1plus: warnings being treated as errors
ep_

Fix test suite compilation errors

This fixes some missing function declaration errors:

CXX ep_testsuite_la-ep_testsuite.lo
cc1plus: warnings being treated as errors
ep_testsuite.cc: In function ‘test_result prepare(engine_test_t*)’:
ep_testsuite.cc:5425: error: no previous declaration for ‘test_result prepare(engine_test_t*)’ [-Wmissing-declarations]
ep_testsuite.cc: In function ‘void cleanup(engine_test_t*, test_result)’:
ep_testsuite.cc:5447: error: no previous declaration for ‘void cleanup(engine_test_t*, test_result)’ [-Wmissing-declarations]
ep_testsuite.cc: In function ‘bool teardown_suite()’:
ep_testsuite.cc:5965: error: no previous declaration for ‘bool teardown_suite()’ [-Wmissing-declarations]
make[2]: *** [ep_testsuite_la-ep_testsuite.lo] Error 1

Change-Id: I2363240c962a09740301a4a85f0559bc04981bcd
Reviewed-on: http://review.couchbase.org/7913
Tested-by: Filipe David Borba Manana <fdmanana@gmail.com>
Reviewed-by: Dustin Sallings <dustin@spy.net>

show more ...

ed2a54a411-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Perform further deduplication for backfilled items.

This will reduce disk IO overhead and resolve mccouch issues caused
by having duplicate items within the same transaction.

Ch

Perform further deduplication for backfilled items.

This will reduce disk IO overhead and resolve mccouch issues caused
by having duplicate items within the same transaction.

Change-Id: I0b2d1552ac094cb221010a4908c8ec772cb4fed4
Reviewed-on: http://review.couchbase.org/7882
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>

show more ...

d79f65ca11-Jul-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

MB-4034 Set open checkpoint Id to 1 if vbucket's state sets to active

If the state of a given vbucket with open checkpoint Id 0 sets to
active, its open checkpoint Id should set to 1 as

MB-4034 Set open checkpoint Id to 1 if vbucket's state sets to active

If the state of a given vbucket with open checkpoint Id 0 sets to
active, its open checkpoint Id should set to 1 as the active vbucket
should always start with checkpoint Id 1.

Change-Id: I8b075b91c525162cea6d0fa5f52521d91e1910cc
Reviewed-on: http://review.couchbase.org/7879
Reviewed-by: Srinivas Vadlamani <srinivas@couchbase.com>
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>

show more ...

0843451208-Jul-2011 Bin Cui <bin.cui@gmail.com>

MB-4039 Launch sqlite command correctly on windows

os.system() won't correctly process db filepath on windows platform.
Instead, use subprocess.popen to execute sqlite command. It works

MB-4039 Launch sqlite command correctly on windows

os.system() won't correctly process db filepath on windows platform.
Instead, use subprocess.popen to execute sqlite command. It works
on both linux and windows.

Change-Id: I3901f523b892a963109ed16649fe54d553b8e93f
Reviewed-on: http://review.couchbase.org/7819
Tested-by: Bin Cui <bin.cui@gmail.com>
Reviewed-by: Steve Yen <steve.yen@gmail.com>

show more ...

7732f65105-Jul-2011 Srinivas Vadlamani <srinivas@couchbase.com>

Fix failing ep-engine atomic_queue_test

Simplified the logic to join the worker threads created by the main
thread. Essentially, since we do not care about the actual return
value fr

Fix failing ep-engine atomic_queue_test

Simplified the logic to join the worker threads created by the main
thread. Essentially, since we do not care about the actual return
value from the threads, it is alright to return a NULL from the
threads and throw it away in the main thread. If at all we must
assert some condition, we should do so in the worker thread. The
previous method of calling pthread_join() was causing a segfault
due to improper casting of the return value.

Change-Id: I244915ff556518bfb470f12d6716cf26c1a1c27e
Reviewed-on: http://review.couchbase.org/7730
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>
Tested-by: Dustin Sallings <dustin@spy.net>

show more ...

5ade399601-Jul-2011 Srinivas Vadlamani <srinivas@couchbase.com>

Fix failing ep-engine checkpoint unit test

The following issues were fixed:
1. Simplified the logic to join the worker threads created by the main
thread. Essentially, since we do

Fix failing ep-engine checkpoint unit test

The following issues were fixed:
1. Simplified the logic to join the worker threads created by the main
thread. Essentially, since we do not care about the actual return
value from the threads, it is alright to return a NULL from the
threads and throw it away in the main thread. If at all we must
assert some condition, we should do so in the worker thread. The
previous method of calling pthread_join() was causing a segfault
due to improper casting of the return value.
2. Fixed the persistence worker thread to read the queue items front
to back instead of back to front. If the persistence thread is slow
for whatever reason, reading the queue back to front may cause the
flush operation to be discovered before one or more of the regular
data items. Consequently, the persistence thread will quit before
it has finished reading the entire queue. This was causing a timing
dependent assertion failure.
3. Fixed the persistence worker thread to also account for checkpoint
start and end operations. The previous code assumed that the
presence of a flush operation indicated end of the queue. However,
this may not be true because new checkpoint start and end ops may
get introduced while processing the regular queue items.
4. Fixed the cleanup worker thread to *not* assert the numItemsRemoved
be zero. The reason this may not be zero is because we may create
new checkpoints during the test and thus end up with one or more
closed and unreferenced checkpoints. Consequently, a call to the
removeClosedUnrefCheckpoints() function may return a non-zero value.

Change-Id: Ie9725ac830ae3ce30a4b3e8c0cb4e9da62b958bd
Reviewed-on: http://review.couchbase.org/7696
Tested-by: Trond Norbye <trond.norbye@gmail.com>
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>

show more ...

b05d98f827-Jun-2011 Srinivas Vadlamani <srinivas@couchbase.com>

Clear engine specific data before SyncListener object destruction

Currently, we attempt to clear engine specific data within the
SyncDestruction callback method. This, however, may intro

Clear engine specific data before SyncListener object destruction

Currently, we attempt to clear engine specific data within the
SyncDestruction callback method. This, however, may introduce a race
condition as a result of which the cookie whose engine specific data
is being cleared may get deleted in a separate thread before it is
accessed in the callback method. The race condition was exposed by a
few failing ep-engine unit tests.

This change set avoids the race condition by clearing the engine
specific data before destruction of the SyncListener object is
dispatched.

Change-Id: I3d6b09adf4e8118918053156a0bd90a8b80cf440
Reviewed-on: http://review.couchbase.org/7602
Tested-by: Dustin Sallings <dustin@spy.net>
Reviewed-by: Dustin Sallings <dustin@spy.net>

show more ...

0898826424-Jun-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

scheduleBackfill() should skip vbuckets that are being backfilled

If a given vbucket is being backfilled from the upstream server,
it should be skipped in scheduleBackfill(). Backfill fo

scheduleBackfill() should skip vbuckets that are being backfilled

If a given vbucket is being backfilled from the upstream server,
it should be skipped in scheduleBackfill(). Backfill for that
vbucket will be scheduled separately once the backfill from the
upstream server is completed.

Change-Id: I297865c51d5a3c2d1e8fcc643ca1d7dff528b553
Reviewed-on: http://review.couchbase.org/7154
Tested-by: Farshid Ghods <farshid.ghods@gmail.com>
Reviewed-by: Farshid Ghods <farshid.ghods@gmail.com>
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>

show more ...

f0b7b24923-Jun-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

MB-4004 Always schedule backfill through scheduleBackfill()

All backfill tasks including the ones from rollback() should be
scheduled through scheduleBackfill() to maintain the consisten

MB-4004 Always schedule backfill through scheduleBackfill()

All backfill tasks including the ones from rollback() should be
scheduled through scheduleBackfill() to maintain the consistency.

Change-Id: I704003f187297f6e67ce2472abb48240be3ecfa5
Reviewed-on: http://review.couchbase.org/7138
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>
Tested-by: Farshid Ghods <farshid.ghods@gmail.com>
Reviewed-by: Farshid Ghods <farshid.ghods@gmail.com>

show more ...

62f8804420-Jun-2011 Chiyoung Seo <chiyoung.seo@gmail.com>

Fix to the checkpoint unit test failure.

The checkpoint unit test should make sure that the persistence
cursor processes all the items from the checkpoint datastructure
at the end of

Fix to the checkpoint unit test failure.

The checkpoint unit test should make sure that the persistence
cursor processes all the items from the checkpoint datastructure
at the end of the test.

Change-Id: I2ecf6f53b7edbd0d2409d8d4d2c5d7d956b0851e
Reviewed-on: http://review.couchbase.org/7076
Tested-by: Chiyoung Seo <chiyoung.seo@gmail.com>
Reviewed-by: Trond Norbye <trond.norbye@gmail.com>

show more ...

12345678910>>...69