[2.0.2] memcachetest+moxi/mcsoda: downstream timeout( with automation tests)

Memcached ops sometimes take an extraordinarily long amount of time

[RN 2.0.1] XDCR docs to replicate for bucket queue is draining unevenly.

Rebalance is slow when indexing/compaction and query load are going on in parallel

Couchbase Server started too early in boot sequence...IP address wasn't yet ready

erlang's global naming facility apparently drops globally registered service with actual service still alive (was: impossible to change settings/autoFailover after rebalance)

[windows] Rebalance exited with reason {unexpected_exit,{badmatch,{error,etimedout}

[windows] Rebalance exited with reason {bulk_set_vbucket_state_failed, {timeout, when data mutation is in progress

ep-engine needs 1.5 minutes to create 1k vbuckets. Seems too slow (but gets fast with barrier=0)

xdcr: data compaction doesn't catch up

Complete Metadata ejection when item becomes non-resident

1 read-only user for UI and REST API

MB-7960 remote_cluster_info module should return remote memcached access info

MB-7960 Implement batch getMeta and setMeta/delMeta operations in ep_engine

MB-7960 Implement ns_server side memcached API for new get_meta_batch and update_meta_batch

MB-7960 XDCR core infrastructure change to enable replicate to remote memcached

Mac OS X App should be signed by a valid developer key

Replace XDCR CAPI protocol on the destination with memcached protocol

erlang crashed with error: "no match of right hand side value {error, closed}...

[Doc'd] failover of node that's completely down is still not quick (was: Rebalance exited with reason {not_all_nodes_are_ready_yet after failover node)

Disk not flush on xdcr dest node

performance testing for cbrecovery tool

Memcached crash on a node on the destination cluster

cbbackup loops infinitely

bgfetcher performance regression

[windows] Rebalance exited with reason {detected_nodes_change, {ns_node_disco_events,

Crash caused by task running while a bucket is shutting down

Update the incorrect information on the Automated Index Updates

[Doc'd] Appends can cause large amounts of memory fragmentation in tcmalloc

Sever performance dip during rebalancing observed in demo environment

Need a purging mechanism for Couchstore to remove deleted / expired items

[RN 2.0.2] config.dat and other potentially security sensitive files are world readable in world readable directories

Binary Protocol on Metadata

Memcached drops tap connections without warning

[Done- RN 2.0.2] Modify a bucket resets unspecified values to the default value

disk channels (threads) are configurable, weighable for different scenarios / workloads

couchstore fsyncs and file per vbucket create very fragmented files (was: couchstore compactor is very poor on reading data from src couch file)

couchstore compactor interleaves byseq btree nodes and doc bodies

System Test : Higher ep_commit_time, beam size on source nodes during xdcr.

View query on Rebalance.Out fails with Reason: A view spec can not consist of merges exclusively.

[RN 2.0.2] Couchbase continually crashes and restart when Couch bucket is created in OpenVZ environment

XDCR Stats enhancements

UI logs showing multiple errors, (XDCR bi and uni, at the start of a rebalance-in operation): Server error during processing: ["web request failed" ...

XDCR: replication lag on ec2 periodically reaches 1 minute and more

[system test] warmup access log corrupt during warmup in windows

race condition while loading sample data via setup wizard

separate disk path for replica index (and individual design doc) disk path

Complete and integrate file I/O NIF to improve couchdb-side latency and throughput

Crash in memcached's release_cookie

Tools to Index and compact database/indexes when the server is offline

Finalize doc editing API and implementation

delete bucket waiting is raceful

XDCR: High CPU utilization on destination nodes

Crash of master node may lead to autofailover in 2 minutes instead of configured shorter autofailover period

Get unit tests running on windows

public api to delete index files

MB-8067 behavior when running out of disk space should be TMPOOM

hotfix release should reflect the change(version#) on the web console

Meaningful alert when low-level packet corruption on node

ns_server should use unique memcached admin/password for each cluster instance

We should hash our REST API passwords

moxi returns "key not found" on ascii-gets during rebalance_in and _out

Certain stats cannot be aggregated for all nodes

windows upgrade fails if machine's ip have changed from the original IP address

Raise a warning message about time on local machines does not match when adding a node.

configurable lifespans in moxi should be 64-bit and parsed as unsigned (safe_strtoul)

open service after loading x items during warmup( starting a server with many items can potentially take hours to fully load hashtable and values)

Our log contains information about the target system we shouldn't reveal

Swap usage shoots up with view querying or indexing

Monitoring and Best Practices Out of Date/Incorrect

Allow for "grouping" servers for replication affinity (rack awareness)

bucket-flush takes over 8 seconds to complete on an empty bucket

making lots of streaming connections in general, or lots of streaming connections in a short period of time causes unexpected resource usage and/or crashing owing to OOM

moxi does not allow a noop before authentication on binary protocol

After two nodes crashed, curr_items remained 0 after warmup for extended period of time

Find a way to avoid creating too many erlang crash dumps

2.0 needs to support use couchbase for all the REST endpoints. (no membase)

Avoid the full rematerizaliztion when the cluster is restarted.

Allow XDCR configuration at bucket replication level, not just the cluster level

Better protect passwords in /opt/membase/var/lib/membase/data/isasl.pw

Allow for resynchronizing of failed-over node after rebalance

Password encryption for SASL buckets over the wire from Client to Server

Update Sizing Calculator for 2.0.2

failover is not quick when any node (including being failed over) is not responding

cbcollectinfo failed to get logs in windows

Couchstore's mergesort module, currently used for db compaction, can buffer too much data in memory

"version" on Windows listed as linux

remove var\log\couchdb and var\lib\couchdb from windows and linux packages

[system test][xdcr+views] rebalance failed due to timeout

error restarting couch_server: {{read_loop_died, {problem_reopening_file,

[system test] warmup access log corrupt during warmup in windows

Memory leak? on one of the buckets on destination cluster (single node) with source cluster on v_iber workload

Add memcached timings stats to memcached

[system test]rebalance failure due to bad replicators after rebalance

Fsync optimizations (remove double fsyncs)

lack of fsyncs in view engine may lead to silent index corruption

[Doc when ready] distribute couchbase-server through yum and debian package repositories

cbcollect_info only collects couchbase.log and misses other logs

Number of document mutations pending XDC replication is rapidly growing

Observed rebalance regression from 2.0.1 to 2.0.2

Rebalance time regression in large scale DGM vperf tests

Checking for purged deletions and update the view accordingly

High latency variation and regression in query throughput

Make XDCR more adaptive at run-time especially when source and destination clusters are highly imbalanced

Improvements to view query performance / latency

unit test to ensure that error strings do not change (used in client API)

Document how to properly shutdown a complete cluster

didn't get error in logs after stopping server in second time: Could not auto-failover node ('IP'). Cluster was too small, you need at least 2 other nodes.

log when the 2.0.2 server stops little information compared to version 2.0.1: the status of a stop of the server isn't displayed

CLI tools need virtualenv sandbox so that they don't conflict with user packages

Couchbase should handle running out of disk space gracefully

Document & cleanup our REST API

[system test] backfill remaining shows overflow value

Time incorrectly displayed in the web UI

Metadata Size Value Discrepancy as a part of documentation

Need documentation on View/Query errors and limitations

Remove the mutation log

Aggregate local docs into a single file

timeout error: having about 100 sasl streaming connection unable to open rest connection during rebalance

Allow for batch set vbucket state commands

Allow batch vbucket deletes

UI should indicate nodes with "problems" or misconfigurations

Unsuccessful upgrade from 2.0.0 to 2.0.1

The installer should check the state for the windows firewall

Inconsistency in the use of symbols in the web UI

The installer should have the user to explicitly have to OK to install on an underprovisioned system

Issues when scaling XDCR on single node

XDCR throughput is unstable and drops to zero when xdcrMaxConcurrentReps is set to high value (256)

Spatial: Use byte sizes to determine fill rate

Linux: Bad messages when restarting service

Web UI truncates fields instead of using the available space on the screen

Couchbase UI must have an option of displaying current mem fragmentation ratio

automated rebalancing feature particularly useful in cloud environments this needs option in Rightscale templates

Incorrect stats displayed on ep_tap_queue_backfillremaining on node, after reboot entire cluster

[system test] curl does not support https protocol in windows

Warmup to continue if the initial access log is quite small or all expired

aggregate all kinds of ops in ops/sec stat (was: Replica Reads don't show up in the UI)

[RN]Items need to be loaded from the mutation log incrementally

Add streaming client connection monitoring

MB-8067 Running out of disk space causes logging module to crash

[system test] [Windows] run more cbtransfer process in windows offline upgrade to speed it up

post 2.0.1 release note items

Need "engineered" way of determining whether certain hotfixes have been applied

XDCR code refactoring

[XDCR + views] we should not try and failover a node when rebalance is already in progress

failover alert for cases where users might lose data should be more quantifiable

Windows installation sometimes exhausts TCP port allocation

windows] memcached connection is lost and rebalance failed with reason {{bulk_set_vbucket_state_failed

Understand pros and cons of having couchbase-server package in public Ubuntu/RedHat repositories

reAddNode returning HTTP 200 on garbled input

Docs: Improve best practices docs

memcached ignores ns_server request(s) to deregister replica building tap names after adding back failed-over node

MB-7761 UI: Getandlock doesn't show up in any operations

UI: CAS operations not aggregated into total ops/sec

Add info about Avast Anti-Virus Software

Inventory on SDK "Advanced Topics"

Move background job queue to CouchKVStore

Supporting data structures for a C view compactor and indexer, and Erlang integration

Allow for default value to append/prepend

add a JSON append feature

Draining rate drops and becomes choppy when more than 5% cache miss happens

Rewrite view compactor in C (in couchstore repository)

XDCR observe

As a user i would like to have a User Defined Function. Client app should be able to call the user defined function in sync/async fashion, something like execute(functionName, key, object,..). This will help me avoid moving data over the network.

We don't replicate the lock time related to "get and lock"

Rabalance button should change status depending on node statuses

Memcached Ascii client doesn't receive TMPFAIL error when deleting data while bucket flush is in progress

Couchbase "Runbooks"

Synchronous writes seem to collapse the database

"backfill remaining" and "remaining on disk" stats are zero while backfill tasks are being executed for restablishing the replicas after the cluster restarts

Compressed documents through similar field names

Couchbase can't use RAM allocated for system disk cache.

Compress XDCR traffic

Compress logs to allow for much longer time period to be captured

MB-7761 Disk reads for append/prepend/incr/decr operations are not correctly accounted in estimating the cache miss ratio

Incorrect return code from deletion of a nonexistent design document

Community AMI credentials

Improve performance for cbdocloader

cbexport - export couchbase data onto external files

couchstore python exposure of header_position off_t is 64-bits

Add ability to modify document bodies during compaction

need an API to control replica-index config per bucket ( disable or enable )

Store revision trees in CouchStore

Total Number of documents information missing in Doc Editing UI

memcached.log file max size is 10*100MB but could be 200MB instead to be consistent with other log files under logs folder

[system test] web console down after failover out from cluster

cbbackup - should monitor disk space and error before running out of space

Using couchbase-cli to setup couchbase(cluster-init) on non default port returns error

can create more than 10 buckets( max by default) with sample buckets

[system test] disk and memory show negative number in windows

MB-6595 bad map function may cause timeouts in entire erlang VM (was: rebalance is stopped by janitor after failover and add back node having views with bad map function)

MB-6595 [system test] rebalance failed with erorr "wait_checkpoint_persisted_failed" due to sync_shutdown_many_i_am_trapping_exits

[RN 2.0.2] resetting the data path to a location which had bad data will cause couchbase server to shut down and not respond ( couch server shuts down first )

In a mix cluster with 181 and 200 nodes, from the UI of 200, not able to see Disk Usage stats.

Inconsistent HTTP response codes for invalid view requests

if flush times out final stage (janitor creating vbuckets back) it returns success causing clients to see TMPFAIL after flush succeeds (WAS: there are reports that even after invoking FLUSH nodes return TMPFAIL...)

Provide way users can populate data from RDBM sourfce to couchbase via csv conversion tool ( e.g cbtransfer)

couchbase view UI should have what's-this tooltips

view Filter Results startkey/endkey fields needs double-quotes, but error message is unhelpful ( hint in the UI would be helpful )

view filter results Reset button doesn't reset reduce or stale controls for buckets other than sample-loader

cbdocloader dumps stack trace on auth / 401 errors

UI - Spatial views don't have "Edit"/"Show" button like other views

Adaptive parallelism in the vbucket move

Create cluster Reference for XDCR should notify port information as part of IP/hostname field

Fix UI to support Non-JSON values ( was: Non-JSON values which are not integers show up as base-64 encoded when viewed through the UI document viewer)

Provide chef/puppet modules and/or examples

Allow to prioritize some view updates over others

Killing myself errors after upgrade to 1.8.1

Beta 2.0 - high fragmentation levels on GAT operations

Please rename lib/couchbase module so that it doesn't conflict with public sdk

Process of removing expired items from memory when the high water mark is hit before ejecting the active non-expired to disk

SASL password with space not being handled properly

Return document ID entry to views screen

XDCR: enable conflict management

memcached gets stuck in shutdown code

Rest API gives back wrong errors

"make test" EXC_CRASH on OsX 10.8, Homebrew, Xcode 4.5.2

moxi doesn't seem to failover on downstream hung

Couchbase server can't handle hundreds of view queries with unlimited number of results at the same time

bucket creation silently ignore unknown options over RESTinterface

sample loader fails when run on a cluster where not all nodes are 2.0

automatically stop rebalance if ns_server detects that it's running out of disk space

Couchbase server package installer must verify the current setting for number of open descriptors

cbworkloadgen should divide and load different key spaces when specifying multi theads

Simplify bucket access control in UI

Remove pending vbuckets from UI - cleanup

dont use md5 hash and dont use openssl

UI log console does not contains information related to upgrade of nodes

[system test] unhandled error: TypeError: prevSamples[keyName] is undefined

enable mac command line tool for mac osx installation

[system test] UI should not display bucket information on newly added nodes for buckets that are not yet rebalanced

Abstract out network related code from memcached (improve error reporting)

healthchecker - Multiple tab support for different scales for healthchecker report

healthchecker - New navigation approach for healtherchecker report : cluster/bucket/node vs cluster/node/bucket

healthchecker - Collect stats in all scales within one run of healthchecker

healthchecker - Hide/expand node detail for healthchecker

win/touchme.bat file works with cygwin but not with mingW

use after free in shutdown path found by valgrind

requesting compressed (zip) diags before sending over http

beam.smp eats up all memory on node beyond the membase quota and causes oom killer

(MacOS) startup time (restart) is quite long considering the number of items on recent builds

UI sometimes show duplicate entries for index compaction

rebalance is failed when change bucket password during rebalance

Don't use pandora for libconflate

Adding node to 1-node cluster when doing heavy IO (e.g. compaction) may cause join to fail (was: rebalance is failed due to timeout when User-triggered compaction of view)

TAP queues show drain rates, yet show no items in queue, no pending backfills and no

Check for available disk space and warn if not sufficient when adding node to cluster

Logout cycle needed to refresh Views list

Timeout in UI doesn't seem to be updated in Views and is still absurdly short

Gap between active and replicated items during load phase

Remove retry logic from CouchKVStore::setVBucketState()

Use the wrapper function getStrError() if possible

We're leaving some disk performance on the table

We need to expose information on half-deleted/still-being-deleted buckets on UI

windows install (and possibly uninstall/reinstall) fails if antivirus is running

cbtransfer gen: performance drops after data load phase

CPU usage shoots up on a cluster with 10 buckets, especially on the node that has UI open

Consider disk write queue size and disk usage to determine if compaction is required or not.

Provide the current disk usage

Moxi uses up a large amount of RAM during a rebalance

MB-7761 Better error responses for append/prepend "not stored"

MB-7761 More operation stats please

"item_alloc_sizes" not correct when using append

View read per sec. should not add per-node counters but max

Add alert when a node does not have any swap configured

Replication should differentiate between deleted and expired items

Support https on REST API

backup/restore - backup into timestamped directory

couchbase logs contain user/password credentials

error in console: Got unhandled error: Uncaught RangeError: Maximum call stack size exceeded At: http://10.1.3.114:8091/js/jquery.js:4029 Backtrace: Function: collectBacktraceViaCaller

cluster is broken when create buckets with some specific names( isasl.pw ns_log _replicator.couch.1 _users.couch.1)

Documents pane may not refresh the UI when a document was deleted

Stale does not mean what it should

cbcollect_info on OSX grabs diags and cbstats but fails to read REST credentials from nsconfig.dat

memcached stucks in shutdown code

Create unit test for management scripts

Blank graphs in UI when one node is timing out stats requests

allow changing the # of memcached worker threads through ns-server apis

Remove (Web UI) and Rebalance (Command-line) Does Nothing

As a user of Couchbase I want to update a single field in a document

XDCR: users should be able to pause and resume an ongoing XDCR process.

moxi assertion fail

web console should display major page faults alongside swap usage specifically for windows

standalone moxi segfault with high connection count multiget and rebalance

Unable to view documents after inserts on recreated bucket

malformed request crashes Moxi

bucket having port number is not mutually exclusive with having a password

client side moxi sigabort if sigstopped during rebalance and multiget

UI ignores rebalance failure if it happens too quickly

Consider gathering REST API usage stats and showing them on stats page

Include vbucket filter on "consumer" TAP stats

"consumer" TAP stats shows "anon" for name instead of more useful name

moxi GET misses stats does not match server side stats

developer can use pre-installed Windows VM's to fix windows issues

MB-7761 gat operation not reporting statistics properly