Release Notes - Couchbase Server - Version 1.8.2 - HTML format

Bug

  • [MB-2160] - Disk space from deleted items needs to be reclaimed
  • [MB-2660] - fix downstream_conn_max to not block during add/remove nodes
  • [MB-3334] - under certain conditions, old items may be read instead of updated items
  • [MB-3466] - Add bytes as synonym to mem_used in statistics
  • [MB-3878] - rebalance failure with Bad replicators after rebalance
  • [MB-3892] - mbrestore not "killable"
  • [MB-3901] - negative # of items and backfill_remaining under TAP_Queue clients section
  • [MB-3938] - mbrestore partial restore if one of the threads encounters an error during restore
  • [MB-4060] - Moxi stack track: invalid pointer
  • [MB-4070] - memory watermark underflow if max_size isn't set
  • [MB-4168] - moxi memory leak?
  • [MB-4218] - manual is not clear about how rebalance affects memcached buckets
  • [MB-4220] - bucket recreate test failure - keys are not entirely deleted from couch files after bucket deletion
  • [MB-4263] - replication keeps going even after rebalance fails or stopped by the user
  • [MB-4269] - inaccurate vb_active_ht_memory value reported in a cluster after few rebalance operations ( replica=2)
  • [MB-4292] - ep-engine does not flush expired replica items if expiry time is 2-3 seconds (memory leak)
  • [MB-4349] - mbrestore doesn't print out any information if user runs it on a data files from a failed over node
  • [MB-4358] - sending backoff to tap producer when disk write queue > 1M is not optimal for large cluster with> 100M items and lot of RAM
  • [MB-4367] - rebalance gets stuck even if ack_seqno is correct and has_queued_item is true and total_backlog_size > 1000
  • [MB-4376] - setup get HDD and memory totale and free space wrong (0Mb)
  • [MB-4428] - GETL leaks memory?
  • [MB-4439] - rebalancing after failing over a node which was down during a previous rebalance causes rebalance failure with bad replicators
  • [MB-4456] - running mbbackup on a huge sqlite database with fragmentation rate of 75% might hang
  • [MB-4457] - warming up from a non-vacuumed large sqlite database might take a long time
  • [MB-4461] - replication cursor stuck from slave 1 to slave 2 , hence high number of checkpoint items in slave1
  • [MB-4476] - adding a node to a cluster which has the same otpCookie causes issues ( happens when vm is cloned or created from an AMI/VM_Template where couchbase was already installed)
  • [MB-4578] - memory spike due to growing number of replica checkpoints if some of the nodes have slow disk write
  • [MB-4595] - rebalancing in a new 1.8 node if there are only 1 checkpoint open results in an imbalanced cluster because ep-engine in the source node does not backfill items from the open checkpoint
  • [MB-4609] - AdminUI: Incorrect info/text for Active disk queue drain rate stat
  • [MB-4613] - couchbase-cli still references membase in help description
  • [MB-4658] - Deprecation message on windows incorrect
  • [MB-4666] - modify copyright test to 2012 in the web console
  • [MB-4729] - Need more documentation on provisioning node via CLI/REST API (with examples)
  • [MB-4738] - compile ep-engine on windows with tcmalloc ( memory accoutning issues in windows when node is actively ejecting items when it reaches low watermark ( delta between mem_used and OS reported allocated memory up to 4 GB ))
  • [MB-4754] - crashed with assertion in bucket_engine - bucket_engine.c:831: get_engine_handle: Assertion `es' failed
  • [MB-4780] - moxi segmentation fault on 1.8 regularly under load
  • [MB-4783] - overstressed memcached keeps conns in close_wait even if the client has already closed its conns
  • [MB-4796] - Need usage example for cbcollect_info during --help
  • [MB-4813] - backfill remaining stat is wrong
  • [MB-4816] - ported documentation really looks bad in the manual
  • [MB-4828] - rebalancing multiple nodes can hang if a bucket has less than 100k items due to a race condition in tap take-over
  • [MB-4834] - TAP mutation flags are in the wrong order
  • [MB-4842] - ep-engine sets ep_num_non_resident to bogus value after warmup
  • [MB-4844] - cbrestore should work with python 2.4 and python 2.6
  • [MB-4859] - Full backfill of replicas after any restart
  • [MB-4864] - rebalancing can get stuck due to a bug detecting the backfill completion during vbucket takeover
  • [MB-4877] - wrong_number_takeovers happens between 1.8.0 nodes casues rebalance to fail
  • [MB-4882] - cbflushctl set max_size doesn't work on a cluster
  • [MB-4883] - memory usage percent on web ui is incorrect
  • [MB-4884] - couchbase backup utility should run integrity check after backup and before vacuuming
  • [MB-4891] - Ctrl+C during a cbbackup/mbbackup process does not cancel the process
  • [MB-4896] - moxi mem leak when using haproxy roundrobin load-balancing
  • [MB-4904] - memcached memory accounting gets worse over time if cache miss is high and memcached is ejecting items in parallel to cache miss constantly
  • [MB-4906] - autofailover may failover two nodes automatically within 1 minute if the master node is failed over and the old master nodes is elected as the master again
  • [MB-4909] - couchbase server console failed to start after reinstall
  • [MB-4922] - rebalancing multiple nodes can hang if a bucket has less than 100k items due to a race condition in tap take-over
  • [MB-4923] - warmup crash from 1.8.1
  • [MB-4935] - cbdbmain --vacuum doesn't work
  • [MB-4961] - increasing the high watermark and decrease it after rebalance operation could help make rebalancing faster
  • [MB-4999] - Incorrect dates showing up in "day" view of UI
  • [MB-5029] - autofailover is enabled but ns-server does not fail over the node after it detects its down ( it fails it over 2 mins later)
  • [MB-5096] - Declaration of 'allocator_stats' shadows a global declaration from memcached
  • [MB-5110] - race condition clearing config during node ejection (was: ns_config:clear may silently fail if config saver is running at call time leading to ejected node having part old config)
  • [MB-5128] - Deadlock during ns_doctor restart during node renaming
  • [MB-5203] - collect_info should include "diag" information
  • [MB-5248] - ebucketmigrator exception error: no match of right hand side value {error,closed}
  • [MB-5252] - show SQLite logic errors in the UI
  • [MB-5255] - rebalance exited with reason mover_failed,wrong_takeover
  • [MB-5296] - branch-1.8.1.xml contains modules that are only used by voltron
  • [MB-5324] - --enable-64bit no longer works due to C++ code
  • [MB-5325] - mcbasher.cc does not compile with g++ on smartos
  • [MB-5327] - Fix SPRO CC options
  • [MB-5328] - Compile failure due to unused variable
  • [MB-5334] - sqlite files not forced into wal mode on startup
  • [MB-6068] - cbrestore does not ignore invalid kv tables which are created when moving data when rebalancing but is purged by database over time
  • [MB-6846] - Impossible to use 1.8.1.xml manifest outside the mountain view office
  • [MB-7148] - cbbackup/cbrestore from 181 to latest 2.0 build, or 181 to 181, with wrong CAS and item_flag value restored.

Improvement

  • [MB-1467] - need to be able to query from a replica vbucket
  • [MB-4480] - We need a stat that tracks meta data usage
  • [MB-4814] - Checkpoint only needs to maintain keys and subset of meta data fields. ( memory optimization)
  • [MB-5159] - Replica read unit tests

Edit/Copy Release Notes

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