3.0
TigerGraph Get TigerGraph

Release Notes - TigerGraph Server 3.0

Release Date: June 30, 2020.

Release Notes for Previous Versions:

For the running log of bug fixes, see the Change Log.

Compatibility with TigerGraph 2

Major revisions (e.g., from TigerGraph 2 to TigerGraph 3) are the opportunity to deliver significant improvements. While we make every effort to maintain backward compatibility, in selected cases APIs have changed or deprecated features have been dropped, in order to advance the overall product.

Data migration: A tool is available to migrate the data in TigerGraph 2.6 to TigerGraph 3.0. Please contact TigerGraph Support for assistance.

Query and API compatibility:

Major New Features and Enhancements

TigerGraph 3.0 is a major revision, with several major new features, many smaller feature additions or improvements, and important performance improvements. Users who are currently running

GSQL

  • ✅ (NEW) User-defined Indexing for Vertex types Users can decide where to add indexes to speed up query performance. Also, available in GraphStudio.

  • ✅ (NEW) Dynamic Querying Queries can now be installed in a generic format, so that the schema specifics (graph, vertex, edge, and attribute names and types) are given as run-time parameters.

  • ✅ More powerful Pattern Matching

    • Perform ACCUM computation and graph updates on any vertex or edge match in the pattern.

    • (BETA) Multiple (Conjunctive) Path Patterns: Easily search for complex graph patterns by combining multiple linear paths.

  • ✅ More powerful Distributed Mode queries

    Queries run in Distributed mode support nearly the full scope of the GSQL language.

GraphStudio

Admin Portal

System Management

  • ✅(NEW) Export/Import Utility

    • Export/Import options for just schema, just one graph, with or without user profiles

  • ✅(NEW) Parallel Installer

    • Significantly faster installation time, especially for clusters

    • Allows users to maintain more than one set of binaries to be installed

Additional New Features and Changes

  • Admin - Better standardization of gadmin command format

  • Admin - license key is not displayed in gadmin output

  • GSQL - ⚠️ STRING COMPRESS should be used with caution. See Data Types

  • GSQL - Complex Edge types: one edge type can be used for multiple pairs of vertex types

  • GSQL - accumulate += vertex and edge attributes in ACCUM & POST-ACCUM

  • GSQL - can use Global SetAccum accumulators in a FROM clause

  • GSQL - TYPEDEF accumulator type, for returning results in nested queries

  • GSQL - FROM clause: Source vertex set has same flexibility as Target vertex set

  • GSQL - backquote escape to encapsulate username strings

  • GSQL - string comparison in lexicographical order

  • GSQL - User's OAuth secret is masked in the output of SHOW USER / SHOW SECRET

  • GSQL - Reserved Word lists have been greatly reduced

  • GSQL - Change behavior of CREATE OR REPLACE QUERY - do not install the new query

  • GSQL/GraphStudio - session timeout with user-configurable time limit

  • GSQL/GraphStudio - installing a query does not block other users from running queries

  • REST - query endpoints can use GET (parameters in URL) or POST (parameters in payload)

  • REST /requesttoken endpoint can use GET or POST

  • REST - Correction: Use port 14240 to access GSQL server, not 8123

  • GraphStudio - Users can search for a vertex type on the Design Schema page, to aid the design of large schemas

  • GraphStudio - Support for Microsoft Internet Explorer and Edge browsers

  • GraphStudio - Remove 500MD limit on file upload size

  • GraphStudio - Set query timeout limit

  • GraphStudio - Refresh the Graph Exploration client with the latest data on the database server

  • GraphStudio - Improved graph display panel, with persistent menu for zoom in/out, refresh display (re-fetch from the database server), and change layout style.

  • GraphStudio - Write Queries button to download the query to a file

  • GraphStudio - accumulator values persist after adjusting the display

  • GraphStudio - other users can run other queries while one user is installing a query

  • Deprecated Features for TigerGraph 3.0

Documentation Updates

  • Clarified that outdegree() might not report the latest values under heavy workload.

Features in 3.0.5 Version

Release Date: 2020-09-05

System Management

  • ✅Support for SSH removal from Cluster operations

    • Remove SSH requirement for Platform services except for one time operations like Install, Upgrade and Start of executor daemon process.

  • ✅Support gadmin commands with failed nodes in cluster

    • Robustness improvement to allow HA service reliability for cluster configuration management

  • ✅Backup/Restore Enhancements

    • Log the directory tree of backup directories

    • Remove ssh dependency

    • Support JSON output for gbar list

    • Record GSQL password and add password check for restore

  • ✅System reliability Improvements

    • Zookeeper load reduction by caching static/runtime nodes in dictionary

GraphStudio

  • ✅Data Loading Improvement

    • JSON File Upload support

GSQL

  • ✅LOAD statement have NEW_VERTEX_ONLY option

    • Load only if the vertex does not yet exist (true Insert semantics instead of Upsert semantics

  • ✅DROP USER is disallowed by Admin users. Only allowed by Superusers.

Known Issues

Applications

System Management

  • Automated Cluster Scale Out

    • In TigerGraph 3.0, automated cluster scale-out is not supported. There is workaround available using the Backup/Restore utility. Customers are requested to reach out to TigerGraph support for Cluster scale-out operation.

  • HA/Node Failure Scenarios: Core data service operations like Querying and Updates will work as usual.

    • Impact of Failure of a node in the cluster:

      • gadmin operations will not work if one of the nodes is offline [NOTE: fixed in v3.0.5]. There is workaround available. Metadata operations will not work if one of the nodes is offline. Metadata operations include Install Query / GBAR / GSQL commands such as User and Token management.

      • m1 node is required to run user-facing applications such as GSQL and GraphStudio. This is because GSQL/Graph Studio metadata is stored on m1 node only.

GSQL

  • Multiple (Conjunctive) Path Patterns:

    • There are no known functional problems, but the performance has not be optimized. Your feedback would be appreciated.

  • DML type check error in V2 Syntax:

    • GSQL will report a wrong type check error for Query block with multiple POST-ACCUM clauses and Delete/Update attribute operation.

Utilities

  • gstatusgraph (Graph Statistics)

    • Graph statistics (gstatusgraph) may give inaccurate number when the cluster has node names too similar, where a node name is a part of another node name. It's usually seen in cluster has 10+ nodes.

    • A quick fix is available at: https://tigergraph.freshdesk.com/support/solutions/articles/5000859550-how-to-patch-gstatusgraph-in-3-0-0

PreviousRelease Notes, FAQs and TroubleshootingNextChange Log

Last updated