Release Notes - Couchbase Server - Version 1.7.1 - HTML format

Technical task

  • [MB-3149] - GUI portion of dealing with engine load/starup failure along with fail reason error reporting

Bug

  • [MB-1627] - merge libconflate & moxi
  • [MB-1707] - moxi mock_x2b tests only suceeed every other time
  • [MB-2399] - password is logged
  • [MB-2773] - Editing memcached bucket shows incorrect value for "this bucket"
  • [MB-2897] - msec_current_time should use 64 bits
  • [MB-3333] - Document that network filesystems are unsupported
  • [MB-3352] - collect_info unable to find "ifconfig"
  • [MB-3370] - Wrong message while running mbrestore in a directory that does not have the backup files
  • [MB-3401] - Rebalance failed when one node is out of disk space.
  • [MB-3633] - README.markdown requires update
  • [MB-3669] - Tap doesn't handle garbage/new messages that are sent over the same connection
  • [MB-3689] - Getl returns both the key and the value in the body section with the binary protocol
  • [MB-3705] - segfault during vbucketmigrator restart
  • [MB-3708] - Failover warns about dataloss when it shouldn't
  • [MB-3748] - ep-engine crash with core dump ep_engine.cc:97
  • [MB-3752] - getl with timeout
  • [MB-3756] - deleting database directory makes ep-engine 'hang' in TransactionContext::commit
  • [MB-3767] - Rebalance fails completely silently when no ns_orchestrator process is running
  • [MB-3786] - "Loading" screen messed up when network connection dropped
  • [MB-3814] - memcached crash bucket_engine.c:1530 : bucket_tap_notify
  • [MB-3816] - rebalance paused for 5 minutes during rebalance 3->1 nodes in 1.6.5.4 if started after flushing a bucket
  • [MB-3826] - rebalance fails with not my vbucket error
  • [MB-3863] - More docs & usage() info for tap.py
  • [MB-3894] - rebalancing out an already rebalance out node while rebalance is still in progress will make all the nodes unusable
  • [MB-3905] - cpu percenage per server stats is broken in the ui
  • [MB-3927] - swap percentage label is pushed under the "progress bar" if it has more than 3 digits. (0.112%)
  • [MB-3931] - replication does not complete when replica factor =2,3 after rebalance/failover operation which involves backfill in the new nodes
  • [MB-3941] - Says bucket with name already exists, but it doesn't ( only the last character is different )
  • [MB-3942] - Can't see bucket whole bucket create menu on small screen
  • [MB-3943] - crash in CheckpointManager::checkOpenCheckpoint_UNLOCKED during rebalancing
  • [MB-3949] - moxi doesn't pass getl timeout information correctly to Membase
  • [MB-3956] - tap.py should not be included in the product bundle since users tend to use it as a supported tool/cli instead of looking at it as a sample code
  • [MB-3960] - getl uses flags field in binary protocol and it shouldn't
  • [MB-3967] - tap client doesn't disconnect after receiving the last item via tap dump ack stream
  • [MB-3972] - setting a new value for an expired item before the pager cleans up the value from memory causes memory leak
  • [MB-3973] - Moxi returns errors during rebalance
  • [MB-3978] - exp_pager_stime should be in stats and dynamically adjustable
  • [MB-3980] - pager may purge an item before its expiration time if user sets a new expiration time for an already expired key
  • [MB-3982] - Mnesia detected overload during dump_log because of write_threshold
  • [MB-3984] - statstore returns error when architecture is unkown
  • [MB-3988] - Upgrade membase server to 1.7.0/1.7.1 from 1.6.5.4 failed if cluster already upgrade from 1.6.5 to 1.6.5.4
  • [MB-3989] - latest 1.7.1r build installation does not work (1.7.1r-27)
  • [MB-3990] - ep-engine crash in ep_engine.cc:2536 during replication in 1.7.0 GA build
  • [MB-3992] - UI error in restoreranalytics.js:229 "Uncaught TypeError: Cannot call method 'concat' of undefined"
  • [MB-3997] - Unable to click Save button on bucket creation dialog
  • [MB-3999] - bucket files are not deleted when node is rebalanced out when bucket is in 5x dgm mode
  • [MB-4004] - replication chains seems to be broken for 3 vbuckets after rebalancing 3->4 nodes in replica 3
  • [MB-4008] - Rebalance exited with reason {badarg,[{ns_config_rep,push,0}, {lists,foreach,2}]} when rebalancing out one node in 5x dgm
  • [MB-4013] - membase-server does not start on windows 2003 if vc 2005 redistributable (8.0.59193) is not installed on the machine
  • [MB-4018] - rebalance fails (Rebalance exited with reason {{badfun,#Fun<ns_vbucket_mover) when upgrading a 1.7.0 to 1.7.1 cluster (rolling upgrade)
  • [MB-4021] - adding a 1.6.5.4 node to 1.7 node or vice versa can make cluster unusable if user clicks on rebalance after addNode
  • [MB-4024] - autofailover "test email" button reloads the page and throws an error
  • [MB-4034] - rebalance fails (mover_failed,{badmatch,true}}) after failing over one node on default bucket with three replica
  • [MB-4036] - mismatch of active and replica items when user stops the rebalance during rebalance/in or out
  • [MB-4038] - "create bucket" floating dialog buttons are unreachable in EI at 1280x800
  • [MB-4043] - membase cli should have a flag for printing out all the server nodes without health info
  • [MB-4056] - replication doesn't replicate all items when failing over nodes from 6->5->4
  • [MB-4057] - update cluster overview screen + display an alert or log message when quota reaches 1 after autofailover kicks in
  • [MB-4063] - one of the nodes comes up with 0 items after cluster-restore
  • [MB-4067] - cluster_restore : memcached fails to start with "unable to open database file" error after restoring data files
  • [MB-4074] - Autofailover count reset button handling code is missing error handling completely
  • [MB-4090] - missing items after windows rolling upgrade from 1.7.0
  • [MB-4091] - windows upgrade fails from 1.7.0 to 1.7.1
  • [MB-4100] - windows upgrade from 1.7.0 to 1.7.1 loses data and config
  • [MB-4101] - ubuntu upgrade from 1.7.0 to 1.7.1 fails with missing 1.7.0 config
  • [MB-4107] - multinode ubuntu 1.7.0 to 1.7.1 upgrade doens't load old config
  • [MB-4120] - moxi launched from ns_server in membase 1.7.0 doesn't start on OSX Lion
  • [MB-4467] - Enable core generation by default

Improvement

  • [MB-2485] - Allow user to hit "enter" to active the next screen of the setup wizard
  • [MB-2823] - Handling of startup time and dynamic engine parameters needs to be unified so any parameter can be changed without restarting the engine (where it makes sense).
  • [MB-2828] - Define detailed procedure for recovering a cluster
  • [MB-3201] - Need to make mbRestore command more user friendly
  • [MB-3512] - Make ns_memcached timeout configurable to make GDB-ing to memcached easier
  • [MB-3834] - Rebalance 1.7.1 nodes into older version of membase
  • [MB-3835] - Autofailover

Story

  • [MB-2650] - As a membase user (or Perry), I want to dynamically change embedded moxi's configuration in a simple way (without having to ask engineers to create and users to deploy a new beam file), so that the customer's cluster can be tuned for more happiness.
  • [MB-2855] - Config fixup script to rename nodes while shut down - restore backup of data to cluster of same size
  • [MB-3515] - Editing bucket may switch access method

Edit/Copy Release Notes

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