Release Notes - Couchbase Server - Version 1.7 GA - HTML format

Technical task

  • [MB-2893] - Issue beta for vBucket Java client

Bug

  • [MB-1295] - Run Valgrind under bat
  • [MB-1344] - Emperfect code around vbucket state on commands
  • [MB-1857] - up to 20 ms response time
  • [MB-1872] - Erlang bug makes our UI break with certain timezones
  • [MB-1980] - membase should support ascii protocol in disk>memory scenario
  • [MB-2040] - After removing servers from cluster active browsers for this servers do not see that servers removed without manual reload
  • [MB-2066] - moxi front_cache doesn't work for a2b/b2b
  • [MB-2129] - moxi in mcmux-compatibility mode is doing unnecessary key hashing
  • [MB-2304] - Cluster overview buckets section shows "Waiting for samples" when no buckets are configured
  • [MB-2309] - Inconsistent units in /pools/default
  • [MB-2676] - Data loss after startup
  • [MB-2743] - Rebalance hanging after trying to "swap" all nodes in a cluster
  • [MB-2833] - Unscynronized clocks lead to data not being displayed on all nodes
  • [MB-2979] - UI is ugly when network error happens during rebalance
  • [MB-3146] - key_stats does not work
  • [MB-3160] - Re-adding failed over node doesn't put it on the pending rebalance list
  • [MB-3276] - RPM (and possibly deb?) versioning is incorrect
  • [MB-3318] - ns_config_rep:init calls ns_config:reannounce, which can time out.
  • [MB-3344] - storage size in data bucket page shows incorrect value
  • [MB-3346] - RAM quota shows incorrectly after delete data in bucket
  • [MB-3380] - ep-engine make test fails on OSX
  • [MB-3388] - top keys has repeats
  • [MB-3396] - Rebalance failed when replication is backed up due to heavy write load
  • [MB-3425] - Add specific node information to "rebalance failed" log entry
  • [MB-3426] - Memcached segfaults on Alestic Ubuntu 10.04LTS 32 bit AMI
  • [MB-3428] - Installing membase 1.6.5 on Ubuntu when it is already install reinstalls the software
  • [MB-3433] - stats unavailable in logs while server is warming up
  • [MB-3445] - Need warning message when fail over second node in cluster with one replication
  • [MB-3448] - Rebalance failed: timeout
  • [MB-3453] - 1.7.0r lacks upgrade scripts from 1.6.5
  • [MB-3461] - memcached crashes when vbucketmigrator reuses tap stream name
  • [MB-3470] - moxi 1.6.5.1 stats.c:201 assert
  • [MB-3477] - database files inscrease their size after each time killing vbucketmigrator
  • [MB-3478] - Tap replication paused with millions of items in queue
  • [MB-3479] - moxi binary protocol uses 0x82 OOM / ENOMEM error code too much
  • [MB-3482] - Turn on auth timeouts for Moxi by default
  • [MB-3487] - moxi assert() at cproxy_reserve_downstream: `d->timeout_tv.tv_sec == 0'
  • [MB-3489] - ep.cc:732: bool EventuallyPersistentStore::deleteVBucket(uint16_t): Assertion `stats.totalCacheSize.get() < ((size_t)1<<(sizeof(size_t)*8-1))' failed
  • [MB-3494] - Error running c:\Program Files\Membase\Server\bin\cli\membase.exe on 64-bit Windows installation
  • [MB-3500] - mnesia receiving empty stats
  • [MB-3518] - curr_connections does a counter underflow when no of connection from client is more than the no of conn specified in membase
  • [MB-3520] - some of the stats does a counter underflow after membase is restarted
  • [MB-3536] - replication is very slow when it reaches about 90% of total items
  • [MB-3537] - Moxi exited with status 139 / segmentatation violation
  • [MB-3547] - Data Buckets page does not refresh automatically when new buckets are added/removed via REST API
  • [MB-3570] - web console displays UIs from previous membase server installation ( CentOs )
  • [MB-3575] - Moxi should automatically skip removed or uninitialized nodes
  • [MB-3586] - membase web console does not work when change IP of server
  • [MB-3592] - rebalance fails with bucket_engine error : bucket_engine.c:1876: bucket_engine_release_cookie: Assertion `peh' failed
  • [MB-3602] - memcached set returns error # 130 ( unknown ) during insertion of large keys
  • [MB-3645] - Investigate and fix mysterious 'address family not supported by protocol.' messages from memcached
  • [MB-3649] - add 'key' stats to mbstats script
  • [MB-3688] - membase server failed to start on windows with PGP Desktop pre-installed.
  • [MB-3692] - pool details "push" doesn't work for node state changes (e.g. warmup -> normal)
  • [MB-3697] - memcached is not using its own logger
  • [MB-3716] - source tarball build doesn't startup with the expected welcome page
  • [MB-3718] - memcached suddenly ejects too many items after OOM error ( RAM usage drops to 25% from 99%)
  • [MB-3719] - Kill mnesia usage for stats
  • [MB-3730] - Pending nodes are also shown as down on the cluster overview screen
  • [MB-3734] - Can't change access mode of memcached type bucket
  • [MB-3736] - Potential for data loss during transfer of actively changing vbucket when rebalancing nodes
  • [MB-3738] - ns_config process grows too large and crashes VM under larger clusters
  • [MB-3741] - 'creates per second' graph label is wrong
  • [MB-3742] - Moxi crashes on long bucket name in stats processing
  • [MB-3758] - need a new Getting Started guide.pdf
  • [MB-3759] - UI doesn't allow complete replacement of existing nodes with new nodes
  • [MB-3762] - We should stop shipping and starting couch
  • [MB-3771] - Sequential remote calls to ns_memcached:connected while building pool details response sucks and may timeout
  • [MB-3776] - Active item loss after reblance-in 5th node
  • [MB-3777] - bucket_engine.c:1413 core dump while possibly retrieving TAP stats during rebalance tests
  • [MB-3778] - erl_init.c:1611 core.bean crash during failover and replication tests
  • [MB-3783] - moxi crash in cproxy_on_close_downstream_conn (cproxy.c:695)
  • [MB-3799] - moxi default hashing choice is not compatible with spy, enyim for memcached buckets (which use ketama_weighted)
  • [MB-3807] - A failedover node that was added back shows on the active server list with rebalance button disabled.
  • [MB-3811] - Fix performance changes due to build changes in 1.7
  • [MB-3812] - ep-engine CXXFLAGS is always -O0
  • [MB-3813] - "rebalance completed successfully" message displayed while rebalance is still in progress
  • [MB-3815] - Spymemcached Incr Issue
  • [MB-3824] - Rebalance 25 nodes into a 25 node cluster failed without a warning
  • [MB-3825] - ui doesnt load server and stats pages in amazon ec2 vms
  • [MB-3827] - memcached takes 90% ram after multiple bucket deletion and it doesn't accept any set command
  • [MB-3829] - rebalance fails silently so new nodes go into pending mode (rebalancing 1->20 nodes )
  • [MB-3843] - replica items missing in 4 node cluster( 2 items in 7 million items)
  • [MB-3844] - web request failed", {path,"/pools/default/saslBucketsStreaming" and pools/default
  • [MB-3845] - moxi: cproxy_protocol_a2b.c:1086: cproxy_forward_a2b_simple_downstream: Assertion `uc->cmd_curr != (protocol_binary_command) -1' failed
  • [MB-3851] - rebalance fails if node goes down during the "remove-node" and "rebalance" clicks
  • [MB-3854] - valgrind report - TapProducer::next(bool&) (stdlib.h:390) invalid read of size 1
  • [MB-3856] - Add ip address to moxi SERVER_ERROR timeout responses
  • [MB-3857] - memcached crash in handle_disconnect at bucket_engine.c:305
  • [MB-3866] - rebalance error after adding 10 nodes with which_children,infinity
  • [MB-3867] - many web_request failures on latest basetar-295+ builds on do_cluster_storage_info and check_and_handle_pool_info
  • [MB-3872] - Membase erlang port range changed to 21100-to-21299
  • [MB-3880] - negative percentage for pending resident items
  • [MB-3881] - 200% beam.smp usage during rebalance after failover ( cluster becomes unresponsive) - amazon ec2
  • [MB-3887] - rebalance is stuck when rebalancing in 3x dg bucket with 20 million items (>100 GB)
  • [MB-3889] - addNode fails with timeout and node2:8091.serverNodes doesn't display self node anymore
  • [MB-3893] - # active vbuckets bouncing between 1024 and not-1024
  • [MB-3898] - upgrade instructions need to be updated before 1.7 GA release date
  • [MB-3899] - rest interface allows 4 replicas while the web gui limits it to 3 replicas
  • [MB-3900] - Missing data when running mbrestore
  • [MB-3904] - rebalance can hang when rebalancing 2->3 if the bucket is at 99% capacity which only happens during bulk load operation
  • [MB-3906] - mbrestore doesn't restore flags correctly
  • [MB-3907] - mbrestore doesn't restore highest-order 31st bit of flags correctly
  • [MB-3908] - rebalance hangs due to underflow of disk_write_queue stat after failing over 2 nodes with default bucket replica set to 2
  • [MB-3914] - swap percentage displayed in the ui doesn't match the swap usage in top command
  • [MB-3915] - missing items in replica vbuckets after rebalancing in/out a cluster in dgm (1.7 RC)
  • [MB-3920] - upgrade fails from 1.6.1 to 1.7.0
  • [MB-3921] - rpm -U does not restore the data after upgrade for 1.6.4
  • [MB-3922] - rpm -U does not restore the data after upgrade for 1.6.0,1.6.1,1.6.3
  • [MB-3925] - rpm -U does not restore the data after upgrade for 1.6.5.3
  • [MB-3928] - Need to display the progress of items moved when upgrade in windows
  • [MB-3929] - upgrading 1.6.5.3 to 1.7.0 membase in windows takes 30 minutes for moving 200,000 items(20 MB of data)
  • [MB-3930] - potential data loss while upgrade from 1.6.5.3.1 to 1.7.0 on Ubuntu 10.04 32-bit and 64-bit
  • [MB-3932] - Failed to upgrade from 1.6.0 to 1.7.0 in windows 32bit and 64bit
  • [MB-3933] - Moving data in upgrade to 1.7 in windows 64bit is very slow
  • [MB-3935] - failed to destroy mutex sigabrt with sync replication
  • [MB-3945] - sync_replication crashes memcached if user passes the wrong vbucket and wrong key
  • [MB-3951] - support_forums link should point to 1.7.0 official page instead of developer preview
  • [MB-3957] - tap.py does not seem to work in 1.7.0 builds
  • [MB-3969] - erl_crash.dump Filling Up Hard Drive
  • [MB-3983] - Upgrade from 1.6.5 (pre-WAL) does not enable RO dispatcher
  • [MB-3991] - getl binary protocol seems to ignore the timeout given by client

Epic

  • [MB-2673] - As a Membase developer, I have a tool that will simulate customer workloads as closely as possible so that I can deliver high quality software.
  • [MB-2883] - Sqoop suport for Membase & Hadoop Integration
  • [MB-2885] - Large Scale Clustering
  • [MB-2886] - Improved Monitoring & Alerting

Improvement

  • [MB-1866] - Display alert when OOM errors are occurring
  • [MB-2244] - wizard should short-circuit to failure page instead of step 2 if installation is unhealthy
  • [MB-2355] - Need alpha-channel-based gauge images
  • [MB-2974] - Add getl support to spymemcached
  • [MB-3262] - Add /var/log/messages to collect_info utility
  • [MB-3624] - add timestamp to each message in alert dialog
  • [MB-3745] - Getl timeouts should be specified as parameters to ep-engine
  • [MB-3819] - 0 swap configuration should show N/A instead of 0 usage
  • [MB-3820] - Server status on monitor server screen

Story

  • [MB-2722] - REST Documentation needs to be up-to-date
  • [MB-2771] - Need a way to kill -9 buckets
  • [MB-2887] - Synchronous Replication
  • [MB-2914] - As a user of the load generator I am able to increment the total item count over time so I can simulate an increasing customer load
  • [MB-2918] - As a user of the load generator I am able to change the throughput at a given rate
  • [MB-3095] - As a tester I can quickly set up a new test server environment
  • [MB-3289] - As a user of the load generator I am able to generate load forever

Edit/Copy Release Notes

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