Release Notes - Couchbase Server - Version 2.0-developer-preview-4 - HTML format

Bug

  • [MB-3466] - Add bytes as synonym to mem_used in statistics
  • [MB-4068] - memcached crashing after creating 6 buckets each 100MB on machine with 2GB ram
  • [MB-4079] - Querying invalid view breaks view querying completely
  • [MB-4136] - _stats built in reduce gives poor error message when emitted value is not a number
  • [MB-4147] - creating view on million existing items in a single node can cause crashes in couchdb replication and cause the ui and view to hang.
  • [MB-4148] - modifying and regenerating index too often for development views can cause the view to hang and not return the results in the specified timeout
  • [MB-4178] - memcached does not start after multiple bucket creation/deletion
  • [MB-4273] - MacOS should store data in Library/Application Support/Couchbase, not Membase
  • [MB-4310] - Node leaves cluster 3 times when it detects that it was shunned.
  • [MB-4398] - rename /var/lib/membase and other membase binaries to couchbase
  • [MB-4409] - moxi-server init script broken on Ubuntu
  • [MB-4459] - Deadlock in flusher thread and assertion failure
  • [MB-4485] - rebalance failed with error memcached exited with status 255
  • [MB-4498] - getting view results after failing over and rebalancing a node fails {"error":"error","reason":"inconsistent_state"}
  • [MB-4500] - reduce map show different count number at each node
  • [MB-4501] - memcached exited with status 139 after restart couchbase server in failover add back test.
  • [MB-4512] - Incorrect total Cluster Storage , summing up index_path and data_path sizes
  • [MB-4518] - deleting items during rebalance results in desynced view
  • [MB-4526] - Disk Overview shows incorrect value
  • [MB-4536] - creating a flawed design doc hangs cluster
  • [MB-4548] - Going from 3rd to 1st page of configuration wizard leaves server in a bad state
  • [MB-4549] - make CAPI CRUD return $flags and $expiration in replies [was: emitting a doc into the view brings $flags and $expiration]
  • [MB-4555] - Document editing UI must be able to handle (random) documents from a node that is down
  • [MB-4560] - Crash in set-index compaction code
  • [MB-4562] - deleting a bucket and recreating it does not delete it's views
  • [MB-4594] - "Undefined set view `default` for `_design/dev_t` design document." on view after rebalance
  • [MB-4596] - add license files to couchbase 2.0 OSX builds
  • [MB-4600] - server goes unavailable, protocol violations after a period of time
  • [MB-4606] - Memcached should not require linking against libmemcached_utilities.0.dylib
  • [MB-4610] - view merge fails if connection_timeout is passed as string
  • [MB-4611] - creating an invalid view causes memcached to hang
  • [MB-4613] - couchbase-cli still references membase in help description
  • [MB-4620] - memcached make distcheck broken in 2.0
  • [MB-4623] - getting missing partition error when accessing view results if one node is failed over but not rebalanced out yet
  • [MB-4624] - UI error {"error":"badrpc","reason":"nodedown"} when user tries to access view results when a node is failed over
  • [MB-4625] - ui shows "Rebalancing 5 nodes " when i was rebalancing out one of the nodes from a 4 node cluster
  • [MB-4626] - couch_set_view,handle_info,\n [{'DOWN' error when user changes the design_doc during rebalance and runs the view query
  • [MB-4628] - make dist for ep-engine 2.0 missing configuration.json
  • [MB-4629] - ep-engine does not persist the mutated replica items to disk
  • [MB-4635] - rebalancing nodes getting stuck in 1->6 or 2->1 with 100k items
  • [MB-4644] - couch_set_view_group dies with reason {read_loop_died,{badmatch,eof}}
  • [MB-4649] - Couchbase 2.0 doesn't support sync
  • [MB-4650] - Main set view group dies with reason {cleaner_died, {function_clause, [{couch_compress, decompress,[<<...]}]}}
  • [MB-4651] - Some race between insertion and delitions items to couch_setview_name_to_sig table
  • [MB-4655] - Maximum call stack size exceeded" error in the UI logs
  • [MB-4656] - rebalance hangs because capi_set_view_manager-default receives an invalid vbucket state change during the takeover operation
  • [MB-4661] - active/replica item count not 0 when all items are deleted.
  • [MB-4663] - couch_db:update_docs crashes with reason {update_error,compaction_retry}
  • [MB-4664] - mc-kvstore does not retry commands if mc_daemon crashes
  • [MB-4669] - "reason":"{<<\"shutdown\ , [{'DOWN' and eventually reason":"req_timedout when running views during rebalance operation
  • [MB-4674] - UI does not show the red status bar when rebalance fails or stopped
  • [MB-4675] - view returns badarg after rebalancing out a node ( rebalance failed first time with bad_replicas ) but succeeded
  • [MB-4676] - rebalance out did not reset the ejected node's state
  • [MB-4677] - UI shows indexing in progress bar continously when indexer updates the replicas
  • [MB-4678] - couch_set_view_compactor dies with 'badarith' reason
  • [MB-4680] - EPERF - Large scale rebalance failing during perf tests
  • [MB-4681] - View queries hang during rebalance
  • [MB-4682] - UI shows an error when creating a view but the view is created
  • [MB-4683] - rebalance fails if the reduce function of a view is invalid
  • [MB-4684] - rebalance is hung at 100% when rebalancing 6 nodes
  • [MB-4686] - View returns partial results
  • [MB-4692] - Reduce view returns incorrect results
  • [MB-4693] - EPERF - disk write stuck during large scale write load
  • [MB-4694] - After a failed rebalance nodes/self api returns 404 "...unknown to this cluster"
  • [MB-4696] - rpm installation fails: error reading information on service couchbase-server
  • [MB-4697] - Sometimes view queries return {"error":"unknown_error","reason":"killed"}
  • [MB-4700] - /pools/default/tasks returns 500 (Internal Server Error) - UI will display the yellow status bar too often
  • [MB-4704] - couch_set_view_group dies with {badmatch,{error,enoent}} reason
  • [MB-4706] - memcached crashed when create a default bucket in cluster
  • [MB-4707] - Failing to delete request amount of documents during rebalance
  • [MB-4708] - cpu at 90% when user run views with stale=update_after on 4 node cluster where load has stopped 30 mins ago
  • [MB-4710] - view immediately returns ""reason": "conn_failed" if one node is down and not failed over yet. it should return partial results instead
  • [MB-4711] - view request fails Subset of nodes failed with the following error: [ { "from": "http://10.12.78.97:8092/_view_merge/?stale=false", "reason": "connection_closed" } ]
  • [MB-4712] - {what,badarg}, error when running views on 20+ design docs on a bucket with 200k documents
  • [MB-4713] - view errors ( ""Undefined set view `default`" ) and (Database `default/124` doesn't exist.") on 700k docs with 20+ design docs
  • [MB-4714] - When mccouch stops responding ep-engine hangs forever.
  • [MB-4718] - cbstats timings fails
  • [MB-4719] - Incorrect rebalance progress status for all nodes, when refreshing the "Server nodes" tab
  • [MB-4721] - delete and recreate default bucket in windows made memcached hang
  • [MB-4723] - Client side error when deleting a view
  • [MB-4724] - sample bucket loader does not work if python 2.6 is not installed or is not the default python interpreter
  • [MB-4725] - docloader and it's ns_server integration need to detect older python and report correct error
  • [MB-4726] - memcached core dump in a 8 node cluster under load (Flusher::step () )
  • [MB-4731] - Error installing latest .deb
  • [MB-4732] - Compaction seems to be stuck (or not running)
  • [MB-4735] - Data loss during rebalance after service restart
  • [MB-4739] - three nodes stopped responding after view load was running on 20 design docs for few hours ( + compaction also didn't run on one node )
  • [MB-4740] - Undefined set view error when trying to get view results with ?full_set=true
  • [MB-4741] - One node crashed (beam.smp) after view load was going in parallel with stop/start rebalance
  • [MB-4743] - Totally messed up data files after night of constant data updating with compaction running
  • [MB-4744] - Views hunged when rebalancing the nodes out due to process timing out and resulting in rebalancing stuck at 100%
  • [MB-4750] - rebalancing two buckets under light write load is stuck at 50% ( possibly second bucket is not being scheduled for rebalance )
  • [MB-4751] - view merge error.. {couch_view_merger,prepare_set_view,4},
  • [MB-4752] - remove one node and rebalance out the node gets stuck at 100% if clients are still waiting for the view results ( 10 index process still in progress )
  • [MB-4758] - Undefined set view errors and rebalance stuck, when querying 20 design_docs in parallel, rebalancing nodes in with stop/start rebalance and continuously loading data.
  • [MB-4759] - Failed to get view query results due to error in couch_view_merger
  • [MB-4760] - Node is in unusual state after failed rebalance
  • [MB-4763] - native btree_implemenation leaks memory [was: EPERF - couchbase ( memcached and beam.smp) go into 99% swap usage while running large cluster tests]
  • [MB-4770] - View eventually returns empty results when queried under load
  • [MB-4771] - Unable to create bucket "Unexpected server error, request logged."
  • [MB-4773] - Rebalance failing on latest 2.0 build due to wait_for_memcached failures (ns_rebalancer:wait_for_memcached)
  • [MB-4774] - compactor_died while loading data and querying view
  • [MB-4775] - Incorrect node status after failing over and incomplete information on Pending rebalance tab
  • [MB-4778] - add a checkbox to the UI to enable/disable replica index
  • [MB-4781] - start_key_docid returns unexpected (or unintuitive) results
  • [MB-4788] - Stats are showing incorrect timing information
  • [MB-4789] - Rebalance_out failed during viewquery test
  • [MB-4790] - xhr failure and views hang after rebalance_in
  • [MB-4793] - view_merger error during rebalance_out query tests
  • [MB-4800] - View query returns unknown_error {"error":"unknown_error","reason":"undef"} when one node is failed over, but not removed completely
  • [MB-4801] - Unfriendly error message for builtin reduce function _stats
  • [MB-4802] - nodes returning empty vBucketMap
  • [MB-4804] - querying views when new nodes are added and not rebalanced in yet and another node is failed over - Viewquery failing due to {"error":"error","reason":"A view spec can not consist of merges exclusively."}
  • [MB-4805] - ep-engine says warmup is done too early
  • [MB-4806] - 2.0.0 DP3 reports wrong version (1.8.0 DEV)
  • [MB-4807] - Set view group craches with reason {noproc, {gen_server,call, [<0.4370.0>,{drop,<0.4362.0>},infinity]}}
  • [MB-4808] - View queries causing errors in couch_set_view_group function
  • [MB-4809] - couchbase_compaction_daemon failed:
  • [MB-4810] - Rebalance out exited with error
  • [MB-4811] - View query returns {"error":"unknown_error","reason":"killed"}
  • [MB-4812] - Set view group dies with file_already_opened error
  • [MB-4815] - inclusive_end ineffective when results are to be reduced
  • [MB-4823] - too many keys in query result
  • [MB-4825] - Rebalance failed for removing and failover of nodes cluster at same time
  • [MB-4830] - autofailover timeout does not take into account heartbeat time
  • [MB-4832] - failed rebalance followed by failover leads to very large data loss
  • [MB-4834] - TAP mutation flags are in the wrong order
  • [MB-4843] - exception error: no match of right hand side value {error,enoent}
  • [MB-4850] - Rebalance hangs at 100% with error ebucketmigrator_srv:init ... {error,econnrefused}
  • [MB-4857] - memcached crash when rebalance after add nodes
  • [MB-4859] - Full backfill of replicas after any restart
  • [MB-4860] - Mac OS zip package needs README with pointer to CLI tools
  • [MB-4876] - RAM usage in UI shows incorrectly
  • [MB-4878] - socker error while connecting to 10.1.2.69:8091 error timed out:
  • [MB-4895] - items does not write to disk
  • [MB-4914] - Getting {"error":{"xhrStatus":"error","errMsg":""}} when ask for view results
  • [MB-4959] - memcached crashed (at stored-value.hh:168) when adding node back to cluster
  • [MB-5161] - 2.0 dp4 gets stuck when warmping up after memcached crashes
  • [MB-5179] - Dp4 rebalance fails after node restart
  • [MB-5243] - restarting a node with more than 2M items could result in loss of 0.001 percentage of active or replica items due to timing issues
  • [MB-5251] - replica item does not create when load item to cluster
  • [MB-5429] - dp4 - dupe keys in view_merge result
  • [MB-6101] - cbbackup can not backup design docs on DP4 version server

Improvement

  • [MB-4450] - Stop compaction during replication and log a message.
  • [MB-4608] - Allow deleting of a document via web UI
  • [MB-4688] - Reduce the default compaction ratio to 30% and the minimum to 2%
  • [MB-4831] - REST API for 'max_parallel_indexers'
  • [MB-4847] - Show replica index progress for initial run

Edit/Copy Release Notes

The text area below allows the project release notes to be edited and copied to another document.