Release Notes for Couchbase Server 7.1
Release 7.1.4 (March 2023)
Couchbase Server 7.1.4 was released in March 2023. This release contains fixes to issues.
This release contains the fixes listed below.
Cluster Manager
| Issue | Description |
|---|---|
Alerts reports "IP address seems to have changed" for nxdomain errors. |
XDCR
| Issue | Description |
|---|---|
XDCR panic when filtering |
|
Backfill Request Handler deadlock |
|
CheckpointMgr hang on P2P RespCh |
|
bucket topology service concurrent map iteration and map write |
Query Service
| Issue | Description |
|---|---|
query using IntersectScan vs UnionScan |
|
FTS SEARCH() with memory_quota fails |
|
INSERT/UPSERT options should not be shared |
Index Service
| Issue | Description |
|---|---|
Log flooded with "FlushTs Not Snapshot Aligned." |
|
Address plasma rpVersion (uint16) overflow |
|
Panic in NodeTable::Get - Logging improvements |
|
Change log level for watcher connection terminations |
|
Rebalance is hung on a dataplane since >1 hour. |
|
Optimise projector CPU during XATTR processing |
|
perf tests stuck due to failed cbindex |
|
Use streamId instead of index.Stream to determine stream catchup pending |
|
Index build stuck on "Check pending stream" during shard rebalance testing |
|
Index build can hang in mixed mode due to projector skipping transaction records |
Release 7.1.3 (November 2022)
Couchbase Server 7.1.3 was released in November 2022. This release contains fixes to issues.
XDCR
| Issue | Description |
|---|---|
XDCR - Unable to create replications |
|
AdvFilter upgrade happens pre-emptively leading to missed documents |
|
Inter Cluster XDCR failing |
|
XDCR Metakv callbacks racing when remote cluster ref is added/changed |
Query Service
| Issue | Description |
|---|---|
LIMIT clause is not working properly with ORDER BY clause |
|
Potential for request stall if stream operator fails to notify request that it has terminated |
Release 7.1.2 (October 2022)
Couchbase Server 7.1.2 was released in October 2022. This release contains new features and fixes to issues.
Features
The following new features are provided.
-
The Search, Eventing, and Analytics Services now support the Magma storage engine. See Storage Engines.
-
The Search Service now supports a Hebrew analyzer. See Understanding Analyzers.
-
The Analytics Service now supports the Parquet file format, for external datasets. See Creating a Collection on an External Link and Analytics Collection Specification.
-
A REST API is now provided to ensure that only nodes with conformant FQDN and CIDR patterns can be added to the cluster. See Restrict Node-Addition.
-
A user’s password can now be changed by means of the REST API without roles needing to be specified. See Create a Local User and Assign Roles.
-
The Search Service now supports higher dimensional spatial structures via GeoJSON, for both query requests and documents. See Geospatial Queries.
-
The Index Service can now optionally create indexes on missing leading keys. See Index Key Attributes.
-
Couchbase Server now provides configurable alerts, to be triggered when memory thresholds are exceeded. See Setting Alerts.
-
The Eventing Services now allows multiple collections to be listened to. See Eventing Keyspaces.
-
Direct backup to Azure blob store using cbbackupmgr CLI or the Backup Service is GA in 7.1.2. See Cloud Backup.
New Supported Platforms
This release adds support for the following platforms:
-
ARM v8 now supported on Ubuntu 20.04 (AWS Graviton)
See Supported Platforms for the complete list of supported platforms, and notes on deprecated platforms.
Deprecated Platforms and Procedures
The following platforms and procedures are deprecated:
-
SUSE Linux Enterprise Server 12 versions earlier than SP2 are deprecated: in the future, they will no longer be supported.
See Supported Platforms for the complete list of supported platforms, and notes on deprecated platforms.
Data Service
| Issue | Description |
|---|---|
wait for seqno persistence won’t timeout on idle vbucket |
|
Memcached crashes in 20 bucket throughput test due to exception |
|
Vbucket stats call to KV can timeout during delta node recovery preparation |
Query Service
| Issue | Description |
|---|---|
Improve pagination queries with fetch |
|
Race condition between stop signal and timeout |
|
Push filters to index scan to potentially reduce fetch size |
|
UNNEST query 'usedMemory' issue when using Query Memory Quota |
|
Include missing key Index syntax on Index workbench doesn’t show include keyword in definition |
|
subqueries should be advised, explained and monitored |
Index Service
| Issue | Description |
|---|---|
With Collection Indexer should index leading MISSING entries |
|
indexer blocked during storage warmup on MOI storage, causing rebalance failure. |
|
Include missing key Index syntax on Index workbench doesn’t show include keyword in definition |
Analytics Service
| Issue | Description |
|---|---|
Select * query throws Failure contacting server for parquet files |
Eventing Service
| Issue | Description |
|---|---|
Eventing function deployment taking a long time |
|
appcode rest api returns bytes instead of string |
|
Log function scope for lifecycle operation audit logs |
|
Unable to modify function settings when user has only eventing_manage_functions role |
|
Eventing Multi Collection: Inter function recursion not detected in case of sbm handler |
|
Eventing Service should honor the CPU & Memory limits set in cgroups |
|
Memory limits are not checked while setting eventingMemoryQuota via REST API |
|
multi-collection-eventing: eventing leaks source bucket mutation to eventing consumer |
|
Function app-log Write hangs when called after Close |
|
Incorrect query param encoding for curl binding and path param is not encoded |
|
Change in error returned when non-existing bucket used in function creation |
|
Eventing function should be able to listen to multiple collections in a bucket at the same time |
|
Timers handler stuck in deploying state after offline upgrade from 6.6.5 to 7.1.1 |
|
Log a system event when an eventing function is auto undeployed due to RBAC changes |
|
Number of cpu cores mentioned in UI warning does not take into account container limits |
|
LCB_ERR_TIMEOUT thrown when keyspace for a bucket binding does not exist |
|
Eventing Multi Collection: Function deployment successful for a function listening at scope level even though scope does not exist |
|
Unable to modify function settings when user has only eventing_manage_functions role |
|
Unable to modify function settings when user has only eventing_manage_functions role |
Release 7.1.1 (July 2022)
Couchbase Server 7.1.1 was released in July 2022. This maintenance release contains fixes to issues.
Data Service
| Issue | Description |
|---|---|
Memcached hangs when no passphrase is passed for encrypted private key |
XDCR
| Issue | Description |
|---|---|
XDCR does not update memcached flag/body after txn xattribute removal if user xattr is not found |
|
Add authType back to bucket properties in pools/default/buckets/bucket-name |
Query Service
| Issue | Description |
|---|---|
Negative integer in the 64bit range causes rounding |
|
Stop session hangs |
|
IN/NOT IN filters not using Hash for evaluation - continued |
|
LEFT JOIN breaks with between operator on non-existing attribute |
|
WITH clause distribution over union queries deviates from sql standard |
|
Adhoc query index selection issue with LIKE as index condition and query parameters |
|
Refresh_cluster_map fails with ERROR 199 : N1QL: Invalid query service endpoint |
Eventing Service
| Issue | Description |
|---|---|
Function causing recursion is missing from ERR_INTER_BUCKET_RECURSION error description |
Release 7.1 (May 2022)
Couchbase Server 7.1 was released in May 2022. This release contains new features, enhancements, and fixes.
New Features
This section highlights the notable new features and improvements in this release.
-
Analytics shadow data may now be replicated up to 3 times to ensure high availability. Refer to General Settings.
-
Analytics now supports Analytics views and tabular Analytics views. Refer to Analytics Views.
-
The new Tableau Connector provides integration between tabular Analytics views and the Tableau interactive data visualization platform. Refer to Couchbase Tableau Connector.
-
The Analytics Service now supports external datasets on Azure Blob storage. Refer to Managing Links and Analytics Links REST API.
-
Analytics now supports array indexes. Refer to Using Indexes and Data Definition Language (DDL).
-
The cost-based optimizer may now consider different join orders, and can choose the optimal join order based on cost information. Refer to Join Enumeration.
-
The Query service now supports optimizer hints within queries using a specially-formatted hint comment. Refer to Optimizer Hints.
-
Couchbase Server now permits multiple root certificates to maintained in a trust store for the cluster. See Using Multiple Root Certificates.
-
Couchbase Server now supports PKCS #1 and PKCS #8 — in each case, only for use with private keys. See Private Key Formats.
-
Use of encrypted private keys is now supported for certificate management. Registration procedures are provided for encrypted private keys associated with node-certificates. See JSON Passphrase Registration.
-
System Events are now provided, to record significant events on the cluster. See System Events.
-
New roles are provided for the administration of Sync Gateway, especially in the context of Couchbase Capella. These roles are listed at Roles.
-
TLS 1.3 cipher-suites can now by used by all services; and by the Cluster Manager, XDCR, and Views. See On the Wire Security.
-
Heightened security is now provided for adding nodes to clusters. Once a cluster is using uploaded certificates, a node that is to be added must itself be provisioned with conformant certificates before addition can be successfully performed. The new node is now always added over an encrypted connection. See Adding and Joining New Nodes.
-
The scalability of indexing is enhanced by the flattening of arrays. See Format of Query Predicate.
-
Automatic failover can now fail over more than three nodes concurrently. See Automatic Failover. This improvement has permitted the removal of pre-7.1 interfaces that were specific to triggering auto-failover for server groups. Consequently, in order now to ensure successful auto-failover of a server group, the maximum count for auto-failover must be established by the administrator as a value equal to or greater than the number of nodes in the server group.
Note that the pre-7.1 interfaces for triggering auto-failover for server groups have been removed from 7.1: therefore, programs that attempt to use the pre-7.1 interfaces with 7.1+ will fail.
Note also that in 7.1, automatic failover of the Index Service is supported.
Updated interfaces for 7.1+ are documented in Node Availability, Enabling and Disabling Auto-Failover, and Retrieving Auto-Failover Settings.
-
Improvements have been made to rebalancing algorithms so that active buckets, services, and replicas will be spread across different server groups, even when server groups are unequal. See Server Group Awareness.
-
The Magma Storage Engine has been added to 7.1 as an Enterprise Edition feature, allowing for higher performance with very large datasets. Magma is a disk-based engine, so is highly suited to datasets that will not fit in available memory. You can find more details on Magma in Storage Engines.
-
The Eventing Service now has full RBAC support allowing non-administrative users to create and manage Eventing Functions subject to the user’s assigned resource privileges. You can find more details on Magma RBAC in Eventing Role-Based Access Control.
-
The Index Service now uses smart batching to reduce the time and resources required to move index metadata, and to rebuild indexes at their new locations during rebalance. See Smart Batching.
Enhancements
The following enhancements are provided in this release:
-
The Analytics function
object_concathas been updated to support dynamic uses, similar to the more general OBJECT constructor functionality that is available in the Query Service. Refer to object_concat. -
XDCR checkpointing is now entirely persistent through topology-changes on the source cluster. This provides improved performance when failover and rebalance occur on the source cluster.
-
The Plasma Storage Engine has been enhanced with per page Bloom filters and in-memory compression. For information, see Plasma Memory Enhancements.
-
Root and intermediate certificates can now be managed while node-to-node encryption is enabled. See Certificate Management and Node-to-Node Encryption.
New Supported Platforms
This release adds support for the following platforms:
-
Apple macOS v11.6 (Big Sur) for development only
-
Apple macOS v12.x (Monterey) for development only
-
Amazon Linux (ARM)
-
Debian 11.x
-
Microsoft Windows Server 2022
See Supported Platforms for the complete list of supported platforms, and notes on deprecated platforms.
Deprecated Features and Platforms
Deprecated and Removed Platforms
The following platforms are deprecated and will be removed in a future release:
-
Apple macOS v10.14 (Mojave) – removed
-
Apple macOS v10.15 (Catalina) – deprecated
-
CentOS 7.x – deprecated
-
CentOS 8.x – removed
-
Debian 9.x – removed
-
Microsoft Windows Server 2016 – removed
-
Microsoft Windows Server 2016 (64-bit, DataCenter Edition) – removed
-
Oracle Linux 7.x – deprecated
-
Red Hat Enterprise Linux (RHEL) 7.x – deprecated
-
Ubuntu 18.x – deprecated
Deprecation of Certificate Upload API
The POST method and /controller/uploadClusterCA URI, which historically have been used to upload an appropriately configured certificate to the cluster, so that it becomes the root certificate for the cluster, are deprecated in 7.1.
For security reasons, in versions 7.1 and after, by default, this method and URI can continue to be used on localhost only. However, this default setting can be changed, if required. For details, see Deprecated Certificate Management APIs.
Note that new methods and URIs for certificate management are summarized on the page Certificate Management API.
Removal of pre-7.1 Server-Group Auto-Failover Interfaces
Automatic failover can now fail over more than three nodes concurrently: this improvement has permitted the removal of pre-7.1 interfaces that were specific to triggering auto-failover for server groups. Consequently, in 7.1+, in order to ensure successful auto-failover of a server group, the maximum count for auto-failover must be established by the administrator as a value equal to or greater than the number of nodes in the server group.
Note that the pre-7.1 interfaces for triggering auto-failover for server groups have been removed from 7.1: therefore, programs that attempt to use the pre-7.1 interfaces with 7.1+ will fail.
An overview of auto-failover is provided in Automatic Failover. Updated interfaces for 7.1+ are documented in Node Availability, Enabling and Disabling Auto-Failover, and Retrieving Auto-Failover Settings.
Fixed Issues
This release contains the fixes listed below.
Installation
| Issue | Description |
|---|---|
Fix cbupgrade for single node IPv6 clusters |
|
Windows installer always rollbacks during install |
Cluster Manager
| Issue | Description |
|---|---|
The old bucket 'sasl_password' should be effectively removed |
|
The versions REST API should be authenticated |
Storage
| Issue | Description |
|---|---|
Cleaning up of the cluster fails with "Rebalance exited with reason {buckets_shutdown_wait_failed" |
Data Service
| Issue | Description |
|---|---|
Limit the Checkpoint memory usage |
|
Cannot make persistent change to num nonio/auxio threads |
|
Align roles to updated permissions in memcached |
Views
| Issue | Description |
|---|---|
ViewEngine doesn’t handle the case of empty default-collection |
|
Views 8092 REST API leaking version info |
Analytics Service
| Issue | Description |
|---|---|
On corrupt remote link details in metakv, analytics cluster becomes permanently unusable on restart |
Query Service
| Issue | Description |
|---|---|
Query log format |
|
Support FTS’s docid_regexp mode for N1QL |
|
Mutation fail may not report the error |
|
Public interface documentation on parsing 12009 DML error |
|
Like functions escape character should be optional |
Index Service
| Issue | Description |
|---|---|
Smart Batching Index Builds During Rebalance |
|
Rebalance button not enabled post Quorum Loss failover even when indexing has partitioned indexes |
|
Internal Server error is raised while performing backup on a index node using cbbackupmgr |
|
Index build stuck during rebalance due to large number of pending items |
Search Service
| Issue | Description |
|---|---|
Rebalance optimisations via index file transfer across nodes |
|
Bind only to IPv4 addresses when invoked with IPv4-Only cluster-wide setting |
|
FTS - Apply RBAC only for target collections in a multi-collection index |
|
n1fty to upgrade to blevesearch/sear for verification phase |
|
Support encrypted certificate / key / password - Search |
|
System Event Log - Search |
|
Multiple Root CA Certs - FTS |
|
Search UI should be able to accept queries as objects |
|
Add Croatian language (hr) to the list of supported languages |
|
SEARCH_META().score behaves different from SEARCH_SCORE() in some N1QL queries |
Known Issue
This release contains the following known issue.
Query Service
| Issue | Description |
|---|---|
Summary: Implement defs.CheckMixedModeCallback for mixed mode checks Any attempt to execute a function with N1QL udfs replicated from a 7.1 node will fail with "no library found in worker" on a 7.0.x node. Workaround: If possible, all nodes in cluster should be running under version 7.1 or higher. |