Release Notes

Guavus SQLstream V5.2.2 Release Notes

1. Introduction
—————

Guavus SQLstream is a massively scalable, real-time Big Data
management platform. Data feeds can be anywhere and in any
format, including message buses, log files, sensors, networks,
social media, databases, and web feeds. Real-time alerts,
aggregated information, and in-memory operational intelligence
can be visualized immediately while streaming the information
directly to external systems, applications and databases.
Ultimately, you gain immediate insight into your data, enabling
you to respond while it matters.

Guavus s-Server 5 is the core data processing engine in
Guavus SQLstream and can process 1,000,000 records per second per core in a commodity server. This level of performance makes for lower
hardware and Cloud costs with more predictable scaling across
the board.

Guavus s-Server is the only standards-based streaming SQL
platform for managing live data feeds in real time. s-Server
is not SQL-like; it’s standard SQL. SQL is powerful, popular,
fast, and scalable. Your existing RDBMS queries can be ported
directly and executed as real-time, continuously executing
queries in s-Server. s-Server is aligned with the SQL:92,
SQL:99, SQL:2003, SQL:2008, and SQL:2011 standards, with an
implementation of the WINDOW operator to support continuous SQL
queries over fixed and rolling time windows.

2. What’s New
————-

Guavus SQLstream 5 introduces the new Extensible Common Data
Adapter (ECDA) and native support for reading from and writing
to relational databases, integrating a collection of purpose
specific adapters and agents into a single software
infrastructure capable of reading and writing to and from the
file system, network sockets, message buses, databases, web
services, and more.

ECDA exposes external streams and tables via SQL/MED (SQL
Management of External Data), making it simple for SQL
developers to incorporate streaming and tabular data from a
wide range of sources into their data processing pipelines.

ECDA supports Kafka, AMQP 0.9, AMQP 1.0, and MQ Series message
buses. It also supports Teradata, Oracle, Microsoft SQL Server,
IBM DB2, MySQL, Postgres, Paraccel, Ingres, Vectorwise, and
HSQLDB relational databases.

ECDA has been tuned for excellent performance and provides a
wide range of parsers including CSV (delimited), JSON, Key/Value
Pair, XML, Protobuf, W3C, Regular Expression, Fixed Column, and
more. ECDA can examine a payload of data and automatically
discern its structure whether it is in CSV (delimited), JSON,
or XML format, making it easy for developers to easily handle
semi-structured data.

ECDA also allows developers to write custom, pluggable filters
enabling programmatic selection of which data is passed along
to s-Server.

3A. SQLstream s-Server System Requirements
——————————————

* A 64-bit version 2.6 or later Linux kernel distribution with
Libc version 2.5 or later. Other Linux kernel versions are not
supported. Tested Linux kernel distributions include:

* CentOS 5.x
* CentOS 6.x
* CentOS 7.x
* Redhat Enterprise Linux 5.x
* Redhat Enterprise Linux 6.x
* Redhat Enterprise Linux 7.x
* Ubuntu 12.04 LTS (Precise Pangolin)
* Ubuntu 14.04 LTS (Trusty Tahr)
* Ubuntu 16.04 LTS (Xenial Xerus)

* JDK (Java Development Kit) version 8. Earlier versions of
Java are not supported. If the JAVA_HOME environment variable
is not set or does not point to the JDK, the installer will
prompt for its location. 64-bit Java is required for the
64-bit s-Server. If the JDK is not installed on your system,
the Installer will attempt to download and install OpenJDK 8
for you.

* 3G RAM

* 2G available disk space

* When running under VMWare, 3G physical RAM is required on the
guest system.

3B. SQLstream s-Studio System Requirements
——————————————

* Windows 10, Windows 8/8.1, Windows 7, or Windows Vista. Both
32-bit and 64-bit versions are supported.

* A 64-bit version 2.6 or later Linux kernel distribution with
Libc version 2.5 or later. Other Linux kernel versions are
not supported. Tested Linux kernel distributions include:

* CentOS 5.x
* CentOS 6.x
* CentOS 7.x
* Redhat Enterprise Linux 5.x
* Redhat Enterprise Linux 6.x
* Redhat Enterprise Linux 7.x
* Ubuntu 10.04 LTS (Lucid Lynx)
* Ubuntu 12.04 LTS (Precise Pangolin)
* Ubuntu 14.04 LTS (Trusty Tahr)
* Ubuntu 16.04 LTS (Xenial Xerus)

* A 32-bit or 64-bit JRE (Java Runtime Environment) version 8.
Earlier versions of Java are not supported. Use the 32-bit
JRE for 32-bit Windows s-Studio installations. Use the 64-bit
JRE for 64-bit Windows or Linux s-Studio installations.

* 2G RAM

* 1G available disk space

* When running under VMWare, 2G physical RAM is required on the
guest system.

* All SQLstream s-Studio system memory and disk requirements
are above and beyond SQLstream s-Server system requirements
when running on the same physical or virtual machine.

3C. SQLstream s-Dashboard System Requirements
———————————————

* A 64-bit version 2.6 or later Linux kernel distribution with
Libc version 2.7 or later. Other Linux kernel versions are
not supported. Tested Linux kernel distributions include:

* Centos 6.x
* Centos 7.x
* Redhat Enterprise Linux 6.x
* Redhat Enterprise Linux 7.x
* Ubuntu 12.04 LTS (Precise Pangolin)
* Ubuntu 14.04 LTS (Trusty Tahr)
* Ubuntu 16.04 LTS (Xenial Xerus)

* The WebAgent component is required for s-Dashboard to connect to
s-Server requires JRE (Java Runtime Environment) version 8.
Earlier versions of Java are not supported.

* Microsoft Edge, Internet Explorer 11 (or later), Google
Chrome 31 (or later), Mozilla Firefox 17 (or later), or Apple
Safari 7 (or later).

* 1G RAM

* 1G available disk space

* When running under VMWare, 1G physical RAM is required on the
guest system.

4. Installation
—————

Quick Guide to Installing SQLstream s-Server

The following guide offers basic instructions for preparing to
install and installing SQLstream s-Server. For more detailed
information, please see www.sqlstream.com/docs.

Preparing to Install SQLStream s-Server

To prepare for installation, follow these steps:

1. Check that your system has a version of libc greater than
2.4.x.

On most versions of Linux, you can see which version of
libc you have by executing the following command:

ldd –version

s-Server will work with any version of libc greater than
2.4.

2. Check that libuuid is available on your system. Most Linux
distributions install libuuid by default. Others may need
to install it manually. On an Ubuntu system, you may need
to run:

sudo apt-get uuid-dev.

If the command

/sbin/ldconfig -p | grep libuuid

returns anything, you should be fine.

3. Ensure that your system has Internet access so that it can
download and install Linux packages that might be required
such as JDK 8.

SQLstream s-Server requires JDK 8. The JDK (Java Development
Kit) is distinct from the Java Runtime Environment (JRE). If it
is not installed, the Installer will attempt to download and
install OpenJDK 8 for you. Alternatively, you can download and
install Oracle JDK 8 by visiting the following URL:

http://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html

You can point s-Server to a specific JDK 8 installation
location on the command line during installation by specifying
JAVA_HOME on the command line as follows:

sudo JAVA_HOME=/usr/local/java/jdk1.8.0_65 ./SQLstream-sServer-5.2.3-x64.run

This method avoids the necessity of setting JAVA_HOME as an
environment variable.

Installing SQLstream s-Server

Proceed with the installation by performing the following steps:

1. If you want to install s-Server as a service, be sure to
log in as a super user (root) or be sure you have access to
sudo. If you run the installer as standard user, s-Server
will be installed, but not as a service.

2. Navigate to the directory where you have downloaded the
installation file (such as SQLstream-sServer-5.2.3-x64.run).

3. Give the installation file permission to run by entering
chmod +x SQLstream-sServer-5.2.3-x64.run

4. From a terminal prompt enter ./SQLstream-sServer-5.2.3-x64.run

For Ubuntu systems, enter sudo ./SQLstream-sServer-5.2.3-x64.run

To complete installation, follow the prompts in the installation wizard.

5. Documentation
—————-

Documentation for this release can be found at ${SQLSTREAM_HOME}/doc.

You can also view documentation in your browser by opening
https://sqlstream.com/docs

6. Known Issues
—————

0038: SIMILAR does not include support for SQL20003 8.6 General
Rule 7L, SQL2003 8.6 Note 190, handling of blanks at the
end of a pattern, or SQL2003 8.6 General Rule 7T.

0589: The SQLLine !metadata directive called with no parameters
throws the IndexOutOfBoundsException. Put any string
after it to receive an error message that lists valid
usages.

0804: The ORDER BY clause for windowed aggregation is not
supported.

0904: ALTER PUMP START can fail silently when it fails to
connect to a foreign database and therefore cannot find
column metadata.

1058: s-Server can crash when its file system fills completely.

1170: The Foreign Stream Definition Editor allows the user to
add columns to a Foreign Stream coming from the log file
adapter. This is not meaningful, and should not be done.

1171: Automatic Java code generation on the Java tab of various
editors in s-Studio does attempt to deal with reserved
words in Java.

1230: The AVG aggregation function returns incorrect results
for very large numbers (where the sum of the numbers
exceeds 2^63 scaled).

1232: If an aggregation function for a real or double precision
over a given window gives a result of infinity, then all
subsequent windows will show a result of infinity, even
if the out-of-range condition has left the window.

1235: s-Studio does not enforce required FOREIGN SERVER
options.

1256: DROP with CASCADE does not delete Pumps.

1316: Renaming a SCHEMA does not propagate the name change to
all of the dependent objects. It is necessary to manually
change them.

1330: It is possible to inadvertently hide portions of the
s-Studio workspace. Workaround:

* close s-Studio

* open a shell window and execute the following commands:

* cd ${SQLSTREAM_HOME}
* cd ui/workspace/.metadata/.plugins/org.elipse.ui.workbench
* rm workbench.xml

* restart s-Studio

1382: Renaming local tables is not currently supported by
s-Studio.

1411: Schema changes made via the use of redirected scripts
with SQLLine are not checkpointed to the s-Server
catalog. Use !run in SQLLine engine instead.

1446: Window rank functions (CUME_DIST, PERCENT_RANK,
DENSE_RANK, RANK and ROW_NUMBER) are not supported.

1458: The s-Studio Advisor supports only queries, not DDL or
DML. Therefore the validation and formatting buttons are
only available in the query editor and the view editor.

1617: User defined types (UDTs) are not supported.

1619: The SUM aggregation function fails for CHAR, VARCHAR,
DATE, and TIMESTAMP data types.

1635: SQLLine’s !manual directive does not display a manual.

1656: The FLOAT data type is not supported by the variable
column log file parser.

1667: ALTER TABLE ALTER COLUMN can crash s-Server.

2062: SIMILAR TO does not process the ‘.’ character correctly
and erroneously assumes it is a special expression
matching directive. Work around this by using ‘[.]’ in
your SIMILAR TO expression.

2078: When writing GROUP BY expressions on ROWTIME, it is
necessary to fully qualify the rowtime with
stream.ROWTIME.

2143: RANGE INTERVAL durations in WINDOW clauses cannot exceed
30 days.

2144: The SOME() Streaming Aggregate function on booleans is
not supported.

2159: SQLlineClient doesn’t accept tab characters as whitespace

2402: The Filewriter Adapter IDLE and IDLE_ONLY options are not
implemented.

2688: Parentheses around windowed JOINS can give a syntax error
or an assertion error. The workaround is to eliminate
the unnecessary parentheses.

2852: The error message presented when inspecting a Write-Only
FOREIGN STREAM is confusing. It mentions that a SELECT
is not possible rather than explaining the actual
problem.

2987: s-Server starts but doesn’t listen if IPv6 is enabled.

3074: IN() expressions with more than 19 values can fail to
plan.

To work around this problem, use multiple IN()
expressions connected together with AND or OR operators.

3088: A View that calls a UDF can assert when its CASE
statement includes this an IN() expression.

To work around this problem, use AND or OR operators.

3232: s-Studio can time out when asked to inspect a query that
takes more than 30 seconds to plan.

3259: s-Studio allows users to reorder columns in an stream
INSPECT, but fails to remember the order change upon
subsequent inspections. This can cause confusion.

3568: Queries that use a streaming ORDER BY, i.e. an ORDER BY
without T-sorting, cannot promote ROWTIME. To workaround
this, used a nested query to compute the ORDER BY and
then promote ROWTIME in the main query.

3906: s-Server can crash when it shuts down.

7. Corrected Issues
——————-

* Corrected from Version 5.1.1

1438: JOIN OVER (ROWS 1 PRECEDING) returns incorrect results

2964: Pumps should automatically be started in topological
order

3220: FOREIGN STREAMs against MySQL with TINYINT(1) column
types cannot be validatated

4209: When using ECDA to write a file with the
filename_date_format set to ‘yyyy-MM-dd-HH:mm:ss’ the
first file is created as test-1970-01-01-00:00:00.xml,
but subsequent filenames correctly use the current date
and time.

4245: Discovery on JSON arrays returning poor results

4282: Discovery on parsing on JSON timestamps is inconsistent

4355: Installer should validate that user supplied email
addresses are valid

4571: With SSL enabled, queries can intermittently fail.

4584: Lack of Internet connectivity causes 3.5 minute delay
on s-Server startup

4651: Declaring an ECDA source column as BINARY or VARBINARY
raises an exception at ingest time

5235: ALTER USER SET PASSWORD for SA (System Administrator)
user does not work

* Corrected from Version 5.0.1

1116: Double clicking on a “Local SQLstream Instance” in
s-Studio should connect to it

1180: Joining a stream against a table is not currently
supported without the use of a UDX.

1681: PUMPs should be schema objects

3531: MERGE to external table fails with internal error: node
could not be implemented

4122: Installer error messages associated with LSB (Linux
Standard Base) not installed are unfriendly

4145: The Installer ignores the –prefix option for the
catalog

4189: Must Write Version of the JDBC Driver Used To Connect to
a Remote Database to Trace.Log

4222: ECDA Formatters need to handle file rotation properly

4293: The CREATE FOREIGN STREAM syntax diagram in the SQL
Reference Guide is incorrect

4330: SQLstream services give error messages whenb starting up
on Ubuntu 15.10

4342: Declaring a FOREIGN STREAM with a VARBINARY column gives
a strange erorr message “Error: Missing required property
‘FILENAME_PREFIX OR FILENAME_DATE_FORMAT OR
FILENAME_SUFFIX’. & null (state=,code=0)”

4343: Internal error while starting a PUMP that MERGEs or
INSERTs into a a FOREIGN TABLE

4357: The Guavus SQLstream appliance should also be released in VHD format
for Hyper-V

4360: INSERT INTO foreign table fails if ROWTIME is in the
select list

4375: Bidirectional ECDA FOREIGN STREAMs are confusing and
should be replaced with specific DDL for input and output
FOREIGN STREAMs

4376: Guavus SQLstream documentation should list all TCP/IP port which a
user needs to open if behind a firewall

4553: Metadata queries do not return the names of primary keys

4647: Limitation of stream/stream JOINs to equijoins needs to
be documented

4651: Declaring an ECDA source column as BINARY or VARBINARY
raises NPEs at ingestion-time

4656: There are problems corecing JSON numbers to non-numeric
SQL types

4657: There are problems corecing JSON boolean lterals to
various SQL types

4658: s-Dashboard documentation links to Vega need to be
updated

4629: Wrong pump names show up in the global error stream

4649: JSON parsing does not work with TIME columns

4662: Ingesting JSON arrays and objects should raise an error
unless the target type is CHAR/VARCHAR

4663: Ingesting JSON string types should handle quotes,
numbers, and boolean types correctly

4664: Ingesting into CHAR columns does not properly add
space padding

4669: Kafka ECDA agent documentation is incomplete

4700: Global error stream reports a late insert incorrectly

4724: Late rows are logged with the wrong error_class to the
global error stream

4729: Multi-row INSERTs corrupt VARCHAR, VARBINARY, and BINARY
data

4742: Provenance columns are not set when ingesting CSV
formatted data through an ECDA agent

4762: s-Server crahes when reading JSON array with no embedded
objects

* Corrected from Version 5.0.0

4377: Redundant versions of GeoIP functions have been removed.
They should now be accessed by loading the GeoIP library
as follows:

CREATE OR REPLACE JAR “GeoIp”
LIBRARY ‘file:plugin/geoip/geoip.jar’ OPTIONS(0);

* Corrected from Version 4.0.4

1379: Clearing an s-Studio Inspect window can take a long time.

1986: To view a streaming query, SQLLine Client needs the
directive !set incremental TRUE. This is not the default
setting in the SQLLine Client distributed with s-Server.

3103: Enhancement: s-studio should align data in numeric
columns in inspect windows

3248: s-Dashboard requires a Linux installation with libc 2.7
or later. Since s-Server runs on Linux installations
with libc 2.6 or later, it is possible to successfully
install s-Server on a machine along side an installation
of s-Dashboard that won’t run because the underlying libc
library is too old.

3440: The use of the STDDEV_SAMP() SQL function can cause a
divide by zero error. Use STDDEV_POP() as a work around.

3596: Studio should not allow “Delete Cascade” for pumps.

3637: REGEX_LOG_PARSE should skip null columns.

3699: The error message when a primarykey uniqueness constraint
is violated needs improvement.

3804: Nested NTH_VALUE() expressions won’t plan.

3825: When a user is logged in as a different user than the
user set to run s-Server, they get a poor error message
which needs improvement.

3850: When StreamLab needs to modify the views that are being
shown in the guide tables (because the user modified the
script), it currently issues STOP commands on the /read
sockets and then issues the SQL on the /sqlstream socket,
assuming that the queries will be terminated before the
new SQL is executed. This works under light loads, but
fails when the streams are running at high speed.

3932: Kafka adapter does not work with DECIMAL data type.

3942: s-Dashboard does not install behind a firewall.

3959: Installation on CentOS 7 fails due to syslog changes.

3993: JSON parser causes Null Pointer Exception when given an
object as a JSON value.

4058: webAgent and s-Dashboard continue to run after a
deinstallation.

4085: Under some circumstances, multiple clients connecting to
the webAgent using multiplexed reads could causes a
deadlock.

* Corrected from Version 4.0.3

1047: s-Server can leave large txnNNNNNNNN.dat files in its
home directory when running.

3534: Under some circumstances, the Mochi demo does not run.

3719: The s-Dashboard Bar Graph for the Mochi demo is blank.

3860: DECIMAL values are not properly handled by the ECDA
parser.

3913: Under heavy load, the webAgent can crash.

3914: The stopDemo.sh script fails to halt the Mochi demo.

3927: Under large numbers of concurrent conncepts the WebAgent
can crash.

3931: The WebAgent can crash when stopping all streams of
flowing data.

3941: When using the WebAgent with a multiplexed /read command
with the same query on several channels, data was only
delivered to one channel.

3953: CONTROL-C does not terminate the WebAgent.

3978: All installed desktop icon need to provide vision
feedback when used.

4028: The s-Studio 64-bit release was actually the 32-bit
release.

4042: Using a SQL MERGE into multiple tables can cause a
deadlock.

* Corrected from Version 4.0.2

3383: The AMQP adapter must support the AMQP 1.0 protocol.

3773: The SQL Planner can be very slow when issuing a CREATE
OR REPLACE VIEW statement underneath active views.

3812: The documentation for TIMESTAMP_TO_CHAR misspells the
required schema as SYS_BOOT_MGMT; it should be
SYS_BOOT.MGMT.

* Corrected from Version 4.0.1

2999: Inserting into a native table which has no primary key
can fail

3686: TableReader fails to stream out the last row when a
TimeStamp is used as a primary key

3690: When ECDA attempts to read a file in a missing directory
it gives the misleading error message “Error: Invalid
Format ‘VCLP’ does not match any known format: ‘CSV,
XML, JSON, Custom’

3694: Exporting code from s-Studio does not export pumps

3778: The ECDA comma separated value FileReader needs an option
to handle quoted strings

3788: The webAgent /getcontinuous/session method can return too
few rows

3806: webAgent can give a ConcurrentModificationException under
high load

* Corrected from Version 4.0.0

3646: The webAgent fails to time sessions out when a client
fails

3687: When s-Studio tries to inspect a foreign stream that is not
inspectable, the error message provided is not friendly

3707: FIRST_VALUE or NTH_VALUE won’t work when inside
a conditional expression

3764: The webAgent /getColumns method can return incorrect
results

3772: The sys_boot.jdbc_metadata.columns_view can return
the wrong precision 0 for simple types like INT, FLOAT

3774: The webAgent should return native table contents in a
predictable, deterministic order

3794: The TableReader fails to reconnect after a client
connection is killed

* Corrected from Version 3.0.0

0640: s-Server Should Report Startup Failures Gracefully

0859: Under certain statement error conditions, UDX calls can
cause the calling client to hang instead of report an
error.

0887: Inserting into a TABLE from a STREAM is not currently
supported without the use of a UDX.

1173: SQL statements are sometimes expanded into unexpected
and complex expressions.

1183: You cannot use the JDBC adapter to write data into a
Foreign Table. A UDX must be used instead.

1410: EXTRACT(<time unit> FROM <date time>) not documented

1714: DROP JAR followed by CREATE JAR doesn’t reload jar file

1758: Local table updates can crash the s-server.

1800: Under some circumstances, INSERT or UPDATE statements on
local tables can give an assertion error instead of a SQL
exception.

1869: CTRL-C does not always cancel a streaming query running
in SQLlineClient.

2018: Cannot use column aliases for the column list sent to
UDXes

2053: VARIABLE_COLUMN_LOG_PARSE() shouldn’t throw out data with
insufficient columns or wrong types. [Fixed through the
use of the new Extensible Common Data Adapter.]

2095: Implement support for SQL “RESPECT NULLS” and “IGNORE
NULLS” for FIRST_VALUE(), LAST_VALUE(), and NTH_VALUE().

2270: INTERVAL types are not supported via the s-Server JDBC
driver.

2342: Documentation needs more information on meaning of
standard error conditions

2387: UDX with two cursor inputs hangs

2399: The Filewriter Adapter can throw a Null Pointer Exception
if its target directory doesn’t exist

2401: s-Server Freezes After Thousands of Connects and
Disconnects

2619: Support for native tables is incomplete

2849: s-Studio can sometimes be unresponsive

2904: nsAgent should have option for timely/correct for each
stream being written

2936: Adding or editing a UDF is not possible using the
s-Studio object editor

2998: TableUpdate Into Microsoft SQL Server With NULLS Can Fail

3007: TableReader Sometimes Fails To Get Data Across 3 or More
Tables

3026: Email Adapter requires configuration workarounds to be
usable

3029: When a UDX is instantiated twice, only one of the
instances execute

3095: Refreshing the current view on s-Studio does not always
show new objects

3104: s-Server Can Crash When an Agent Disconnect

3107: Streaming GROUP BY with ORDER BY ignores ORDER BY

3121: Internal Histograms for MIN(), MAX(), and NTH_VALUE() Can
Leak Memory

3192: Pump page in documentation has a broken link

3199: The TableUpdate UDX can failed with a null pointer
expcetion when writing to a MySQL table when writing to
a table with mediumtext data type column

3213: s-Server does not exit gracefully when license expires

3218: Rowtime bounds can be sent too early from UDXs

3224: Need to be able to install multiple versions of s-Server
on the same machine when s-Server does not run as a
service

3226: Closing a pump can lose rows

3234: Server fails to start if IPv6 is in /etc/hosts and
server not on DNS. [The Installer now checks for this.]

3242: The FileWriter agent can misnames a rotated files when
the filename base contains a period

3269: Log File Agent and Adapter need option to skip first row
in a tailed log file

3275: Java Calcs do not pass rowtime bounds

3286: Autocp Directory Plugins Do Not Appear In the s-Studio
Drop Down List

3294: Trace log goes to incorrect directory

3297: The licensing system reports an evaluation license
expiration despite a production license being installed

3301: Connecting to SQLstream using perl and the JDBC cpan
module requires perl 5.8 or earlier

3309: Precision and scale calculations for NUMERIC types
is undocumented

3314: All output agents need hook to to flush their buffers
upon termination. [Added to the new Extensible Common
Data Adapters and Agents.]

3315: Cannot upsert into native tables

3328: SMTP adapter could throw an
ArrrayIndexOutOfBoundsException

3335: console.log missing after non-service install

3336: s-Studio does not always reflect the correct state of
a PUMP in its tree view.

3356: Certain Log File Agent options cause it to stop following
file rotations

3367: AMQP Adapter should allow connection parameters in
jdni.properties

3372: JDBC incompatibility should be reported gracefully

3373: AMQP adapter only support AMQP 0.9 not AMQP 1.0

3378: TableReader/TableUpdate documentation does not list
valid SQL Dialects for some databases

3382: Filewriter overwrites file when s-Server is restarted

3385: The TableUpdate UDX fails when writing to Oracle tables
with lowercase names

3397: The s-Studio SQL Advisor can be slow over WAN connections

3398: s-Server chokes on bad whitespace in DDL generated by
s-Studio

3404: Table Reader UDX options are case sensitive

3409: s-Studio converts all foreign stream options to delimited
names

3423: s-Dashboard needs to detect incompatible browsers

3436: $SQLSTREAM_HOME/support/sql/showStatements.sql fails

3446: Need means to identify agent version numbers. [Added to
the Extensible Common Data Agents.]

3449: When s-Server is down, LogfileAgent tries to reconnect
for only a finite amount of time

3451: TableLookup generates wrong Lookup Query for Compund
Names

3452: TableLookup should not query a database on a cache miss
when target is prefetched

3463: Rapidly stopping and starting s-Server makes installation
unusable

3465: TRIM function examples in documentation are not clear

3479: Start s-Server as root the first time it is run causes
subsequent catalog permissions problem and misleading
error message

3480: s-Studio Must Clean Up Disconnected SQL Sessions Upon
Restart

3481: Dynamic parameters are not allowed for UDX/UDF parameters

3482: The Developer Guide documentation index is missing the
example chapter

3484: Two Inspect windows on the same stream can hang s-Studio

3488: The QuickStart Guide is missing information on
TableReader UDX state table

3506: Need to include and document the new AesEncryptDecrupt
UDX

3512: Mochi demo SQL documentation needs clean up

3515: When s-Server is running under load, interactive clients
aren’t responsive

3516: Mail sender credentials should be validation at start of
execution time not at query planning time

3517: Some pumps create and leak phantom sessions/statements

3518: When starting pumps cause an s-Server startup to be slow,
inbound clients can hang trying to connect

3549: Network Socket Agent Transmits Wrong Number of Columns
When -id Paramter is Set

3552: Documentation erroneously says that can use two dashes
before “su” in the logfileagent.sh command

3607: s-Studio’s Stop Pump progress dialog says “Starting
pump…”

3617: s-Studio disables Delete Cascade when select multiple
schemas

3642: The JDBC driver URL in ERnterprise Integration Guide is
incorrect

* Corrected from Version 2.5.1

0878: JDBC extensions don’t work with connection pools

0909: Need Indicator That Inspect Window Query is Still
Starting

0965: Need Oracle Database Adapter

1047: s-Server sometimes leaves spurious files named
txn####.dat in its home directory. These can be deleted
with no ill effect.

1051: Normalization UDR Needs Documentation

1152: Invalid View Stamp (Red X) No Longer Shows Up

1242: CREATE OR REPLACE SCHEMA will do nothing of a schema has
active objects.

1270: Need to support JDK 1.6 and JDBC 4.0 extensions

1278: Validator allows all window specs but offset windows are
not implemented

1392: JDBC driver should allow no-wait polling

1410: EXTRACT (<time unit> FROM <date time>) is not supported

1419: s-Server does not respond to system shutdown events
correctly. On restart, s-Server will recover, but changes
since the last successful shutdown will be lost.

1482: The s-Studio object editor definition tab is not updated
after manually edited SQL is saved.

1516: Outer join physical plan swaps its left & right inputs

1521: Need Multi-select export in Studio

1567: The s-Server planner cannot currently implement multi-way
JOINs. To work around this, implement your multi-way JOIN
as nested two-way JOINs.

1724: Using CAST on DATE to CHAR or VARCHAR without a length
causes an overflow error. Be sure to specify CHAR and
VARCHAR sizes when casting from DATE fields.

1765: Multiple INSERT statements on the same client connection
can cause a deadlock

1815: The aggregation functions FIRST_VALUE() does not work on
boolean datatypes.

1837: Need documentation for CURSOR

1857: Windowed join causes “node could not be implemented”
planner assertion

1886: If a PUMP fails to start, no error is reported to the
client. To discover the error, one must look at the
trace log.

1904: CREATE PUMP statement fails when ROWTIME column is
specified in “insert-column-list”

1936: UNBOUNDED PRECEDING windows are now implemented without
saving all data that has passed through the window.
Instead, incremental aggregates are computed instead.

2070: Planner blows up if query has ‘n = n’ in join condition

2087: Embedded SELECT Clauses With Streaming GROUP BY Causes Java
Array Out of Bounds Error

2100: AssertionError on FLOOR(<value> TO SECOND)

2118: MONOTONIC function needs documentation

2142: Aggregate functions cannot be written on BOOLEAN
expressions. Work around this by mapping BOOLEANs to
INTEGERs and aggregating those.

2174: FileWriter Documentation Does Not Mention CSV Output
Ability

2306: Export pumps in STOPPED mode

2309: AMQP adapter does not work with latest
rabbitmq-java-client 2.1.1

2353: s-Studio missing a way to return to a previous help page

* Corrected from Version 2.5.0

1765: Multiple INSERT statements on the same client connection
can cause a deadlock when one of the statements closes
while the other is blocking with full buffers. There are
two ways to work around this:

– Before closing an INSERT statement on a shared client
connection, punctuate the statement by calling
setRowtimeBound() for a very long time into the future,
such as:

setRowtimeBound(new Timestamp(Long.MAX_VALUE – 1))

then wait two seconds before closing the statement; or

– Use separate connections for each INSERT statement
through the client driver.

2136: Extend the Log File Adapter to provide a means of
stepping over old log files during a restart.

2192: IPv6 /etc/hosts can cause the s-Server startup procedure
to fail.

2212: The s-Server startup script should report when pumps are
being started.

2272: The s-Server Installer should create a clean checkpoint.

2300: Unnecessary trace messages appear on the console when
starting the s-Server.

2320: After editing a column in the Studio Definition Tab from
VARCHAR or CHAR to INTEGER and attempting to save, an
error is given.

2327: Support was added for exponentially decaying averages in
windowed aggregation.

2359: A program using JDBC cannot reconnect to s-Server after a
failure.

2386: Database Adapters and UDXes should use the same
connection definitions

2400: Backpressure causes CPU to spike to 100%

2445: The Socket Adapter now handles multiple character
delimiters

* Corrected from Version 2.0.2

1050: Punctuating a stream that joins multiple streams doesn’t
always emit all data rows that should be available at that
point in time. Workaround: adding additional data forces
the straggling data rows to be pushed through.

1261: The MINUS operator is not supported on DATETIME data.

1284: Dropping a running pump does not stop the underlying
stream. Only a server restart properly resets it.

1480: The s-Studio can show ad hoc queries as unsaved, right
after they have been saved.

1605: The s-Studio object tree is not automatically refreshed
after a stream or view is renamed. Use the refresh menu
option to force a manual refresh.

1780: Unaliased aggregation functions used with streaming GROUP
BY statements report an error. To work around this,
always use aliases:

SELECT STREAM COUNT(*) AS failrate
FROM failures
GROUP BY FLOOR(failures.ROWTIME TO SECOND);

not

SELECT STREAM COUNT(*)
FROM failures
GROUP BY FLOOR(failures.ROWTIME TO SECOND);

1798: The s-Studio does not properly enable the Format and
Validate buttons on newly created objects. Close the
window and reopen to enable them.

1856: Streaming Aggregation must include ROWTIME calculations
as the first item in the GROUP BY clause or s-Server may
exit with a stack overflow error.

* Corrected from Version 2.0.1

0993: DROP FOREIGN DATA WRAPPER CASCADE intermittently crashes.

1640: <timestamp> – <timestamp> is not supported in query
expressions.

1641: SQL Line Client can’t handle case where a line ends in a
semicolon followed by white space.

1691: Row parser support for row parsing functions and DDL is
inconsistent.

1812: Certain files left behind after s-Server is uninstalled.

1832: Under certain circumstances, stopping a query in SQL Line
Client via CTRL-C can cause the s-Server to crash.

1839: The confirmation message for killing a statement in
s-Studio is wrong.

1840: Streaming GROUP BY with explicit ROWTIMES is not
supported.

1863: Support for foreign streams coming from separate
instances of s-Server is not supported.

1870: Installing into directory ‘.’ renders CLASSPATH unusable.

1883: Large ASpenApplicationTrace.log files makes s-Studio
unusable.

1887: The Table Update adapter requires a primary key in the
destination table.

1890: Selected from pumped streams can cause an assert in the
s-Server scheduler.

1905: The log file reader returns inconsistent results with
explicit row times.

1917: The checkpointing script fails with “cp: preserving times
for ‘$file’: Operation not permitted” on certain Linux
distributions.

* Corrected from Version 2.0.0

1884: When an active log file input stream returns a buffer
with null bytes embedded, the Log File Agent abruptly
stops processing data.

1877: When a regular expression used in parsing a FOREIGN
STREAM fails to match an input string, the Log File
Adapter stops working.

1875: The Log File Agent does not properly recover from a
server restart.

1820: Some of the logfile parsing functions (notably the W3C,
VARIABLE_COLUMN, and FIXED_COLUMN parsers) can create
excessively large output streams. For example, parsing a
4K input stream into 20 columns with the W3C_LOG_PARSE
function, each of the 20 columns is also 4K, which
exceeds the 64K SDP buffer limit. If you get SDP buffer
errors, try reducing input and output stream sizes or
parsing fewer columns per view.

* Corrected from Version 1.1.2

1079: In certain circumstances a JDBC rollback causes a server
crash.

1226: MIN and MAX aggregation functions on DECIMAL types round
their result values to the nearest integer.

1231: Aggregation functions on exact numeric types currently
ignore overflow conditions, so results overflowing 2^63
will produce incorrect results without warning.

1389: Mixing windowed aggregation and traditional aggregation
in the same query can cause s-Server to crash.

1578: WINDOWs of more than 24 days (approximately 2^32
milliseconds) are not yet supported.

* Corrected from Version 1.1.1

0845: Under some circumstances, the optimizer can take a very
long period of time to find an optimal query plan.

0886: Streaming GROUP BY is not yet supported. Attempts to use
it can cause a crash.

0990: The Database Adapter cannot handle the SQL Server types
MONEY, NTEXT or NVARCHAR

0997: After long periods of time with the Database Adapter
running idle, s-Server can crash due to a known memory
leak in Java.

1000: The Database Adapter can sometimes detect and therefore
apply transactions out of order when transactions commit
in a different order from their final operation and when
they commit within ~10ms.

1149: The Export Schema function can write objects out of
topological order, causing them to fail on import.
Workaround: edit the exported file and reorganize the
objects in topological order.

1173: Under some circumstances, the s-Studio can expand a SQL
statement into its underlying representation, overwriting
the “native” implementation provided by the user.

1344: If there is no text in the ad-hoc query editor, the
Execute, Check Syntax, and Pretty Print buttons
produce error messages.

1365: Copying and pasting a stream within the Repository
Navigator causes a server assertion.

1384: It is possible for s-Studio’s inspect window to display
rows out of order.

1387: It is not possible to CAST columns to the ROWTIME
reserved column.

1438: The s-Server may return incorrect results when processing
offset windows in a WINDOW clause.

* Corrected from Version 1.1.0

1463: Import does not properly report errors.

1462: Export does generate a SET SCHEMA command in its output
file.

1447: The definition editor erroneously show ANY for the data
type on new columns.

1436: The distribution media needs installation instructions in
its root folder.

1421: Installer crashes on Kubuntu Linux distribution.

1420: Server crashes when computing certain Windowed
aggregation results.

1407: XML Xpath User Defined Function fails.

1405: s-Studio can report “connection lost” when opening
FOREIGN STREAMs.

1403: Need to include a s-Studio reset script within release
bits.

1400: Need to dramatically simplify Bincho demonstration setup.

1399: The s-Server release DVD should include a distributable
virtual machine that has all patches and settings in
place.

1397: The s-Server release DVD should include a separate tree
with SQL Server Bincho demo files.

1386: The Ad Hoc Query editor never closes its statement.

1383: s-Studio intermittently loses connecting while saving
changes to STREAM definitions.

1356: s-Studio does not refresh itself properly after importing
from a file.

1352: The Add Option dialog for a STREAM or TABLE definition
comes up with an error icon.

1341: Using the Java tab for various object editors causes
s-Studio to crash.

1340: Attempting to connect s-Studio while s-Server is not
running gives the confusing error message “Cannot create
PoolableConnectionFactory”.

1282: The query editor ignores the default schema setting.

1239: Unsupported aggregation functions are allowed by the
s-Server parser and are not properly caught by the
s-Server validator, generating the confusing error
message, “Calculator program implementation error:
Unknown operator”.

1217: After creating a new object with s-Studio and saving it,
the object’s tab still shows a *, indicating that is has
not been saved. This only happens when creating and
immediately saving an object.

1215: The s-Studio SQL Editor’s real time validator cannot
properly handle DDL expressions. It shows syntax and
semantic errors where there are none.

1188: Using s-Studio’s Kill Statement functionality can cause
it to hang.

1165: The s-Studio disconnects from the s-Server once in a
while.

1129: New User Defined Function menu item does nothing.

1074: There are missing scroll bars from various s-Studio
definition tabs.

1007: Cannot declare a TIMESTAMP column in a STREAM.

0967: BIGINT values can get corrupted in certain conditions.

0918: The database adapter grabs locks in a conflicting order
causing deadlocks.

0611: Welcome screen obscures other views in a way that makes
it unclear how to proceed.

0554: Client applications are now notified of abnormal end of
result sets stemming from a user cancellation or an
operator kill.

8. License
———-

Please see ${SQLSTREAM_HOME}/LICENSE.txt for the license governing SQLstream products.

Guavus SQLstream V5.1.0 Release Notes

1. Introduction
—————

2. What’s New
————-

ECDA also allows developers to write custom, pluggable filters
enabling programmatic selection of which data is passed along
to s-Server.

3A. SQLstream s-Server System Requirements
——————————————

* A 64-bit version 2.6 or later Linux kernel distribution with
Libc version 2.5 or later. Other Linux kernel versions are not
supported. Tested Linux kernel distributions include:

* CentOS 5.x
* CentOS 6.x
* CentOS 7.x
* Redhat Enterprise Linux 5.x
* Redhat Enterprise Linux 6.x
* Redhat Enterprise Linux 7.x
* Ubuntu 10.04 LTS (Lucid Lynx)
* Ubuntu 12.04 LTS (Precise Pangolin)
* Ubuntu 12.10 (Quantal Quetzal)
* Ubuntu 13.04 (Raring Ringtail)
* Ubuntu 14.04 LTS (Trusty Tahr)
* Ubuntu 16.04 LTS (Xenial Xerus)

* 3G RAM

* 2G available disk space

* When running under VMWare, 3G physical RAM is required on the
guest system.

3B. SQLstream s-Studio System Requirements
——————————————

* Windows 10, Windows 8/8.1, Windows 7, or Windows Vista. Both
32-bit and 64-bit versions are supported.

* A 64-bit version 2.6 or later Linux kernel distribution with
Libc version 2.5 or later. Other Linux kernel versions are
not supported. Tested Linux kernel distributions include:

* CentOS 5.x
* CentOS 6.x
* CentOS 7.x
* Redhat Enterprise Linux 5.x
* Redhat Enterprise Linux 6.x
* Redhat Enterprise Linux 7.x
* Ubuntu 10.04 LTS (Lucid Lynx)
* Ubuntu 12.04 LTS (Precise Pangolin)
* Ubuntu 12.10 (Quantal Quetzal)
* Ubuntu 13.04 (Raring Ringtail)
* Ubuntu 14.04 LTS (Trusty Tahr)
* Ubuntu 16.04 LTS (Xenial Xerus)

* 2G RAM

* 1G available disk space

* When running under VMWare, 2G physical RAM is required on the
guest system.

* All SQLstream s-Studio system memory and disk requirements
are above and beyond SQLstream s-Server system requirements
when running on the same physical or virtual machine.

3C. SQLstream s-Dashboard System Requirements
———————————————

* A 64-bit version 2.6 or later Linux kernel distribution with
Libc version 2.7 or later. Other Linux kernel versions are
not supported. Tested Linux kernel distributions include:

* Centos 6.x
* Centos 7.x
* Redhat Enterprise Linux 6.x
* Redhat Enterprise Linux 7.x
* Ubuntu 12.04 LTS (Precise Pangolin)
* Ubuntu 12.10 (Quantal Quetzal)
* Ubuntu 13.04 (Raring Ringtail)
* Ubuntu 14.04 LTS (Trusty Tahr)
* Ubuntu 16.04 LTS (Xenial Xerus)

* The WebAgent component is required for s-Dashboard to connect to
s-Server requires JRE (Java Runtime Environment) version 8.
Earlier versions of Java are not supported.

* Microsoft Edge, Internet Explorer 11 (or later), Google
Chrome 31 (or later), Mozilla Firefox 17 (or later), or Apple
Safari 7 (or later).

* 1G RAM

* 1G available disk space

* When running under VMWare, 1G physical RAM is required on the
guest system.

4. Installation
—————

Quick Guide to Installing SQLstream s-Server

The following guide offers basic instructions for preparing to
install and installing SQLstream s-Server. For more detailed
information, please see www.sqlstream.com/docs.

Preparing to Install SQLStream s-Server

To prepare for installation, follow these steps:

1. Check that your system has a version of libc greater than
2.4.x.

On most versions of Linux, you can see which version of
libc you have by executing the following command:

ldd –version

s-Server will work with any version of libc greater than
2.4.

2. Check that libuuid is available on your system. Most Linux
distributions install libuuid by default. Others may need
to install it manually. On an Ubuntu system, you may need
to run:

sudo apt-get uuid-dev.

If the command

/sbin/ldconfig -p | grep libuuid

returns anything, you should be fine.

3. Ensure that your system has Internet access so that it can
download and install Linux packages that might be required
such as JDK 8.

http://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html

From that page, download the correct file for your system, and
unpack the compressed binary file, by entering

tar xvf jdk-8uXX-linux-x64.tar.gz

Files should uncompress to a directory such as jdk1.8.0_XX. The
location of the uncompressed files, such as
/usr/local/java/jdk1.8.0_XX, is the home directory for the JDK.
You will need to set JAVA_HOME to this location.

Setting JAVA_HOME and PATH

The steps for setting JAVA_HOME depend on your system. For all
systems, we recommend setting such variables for the whole
environment. We recommend against, for example, using .bashrc
to set such variables, since .bashrc sets variables for a
single user.

* In Ubuntu 14.04 or 12.02, the only way to set JAVA_HOME for
the whole environment, for desktop icons as well as command-
line shells, is to edit the file /etc/environment.

* In CentOS or Redhat, the only way to set JAVA_HOME for the
whole environment, for desktop icons as well as command-line
shells, is to create and edit a script in the directory
/etc/profile.d.

For both systems, you will need to add lines to the file or
script to set and export $JAVA_HOME, as well as to add $JAVA_HOME
to the system path. For example, if you installed the JDK in
/usr/local/java/jdk1.8.0_65, you could use the following code:

JAVA_HOME=/usr/local/java/jdk1.8.0_65
PATH=$PATH:$HOME/bin:$JAVA_HOME/bin
export JAVA_HOME
export PATH

Setting JAVA_HOME on the Command Line During Installation

Alternately, you can point s-Server to the JDK 8 installation
location on the command line during installation, using code
along the lines of

sudo JAVA_HOME=/usr/local/java/jdk1.8.0_65 ./SQLstream-sServer-5.1.0-x64.run

This method avoids the necessity of setting JAVA_HOME as an
environment variable.

Installing SQLstream s-Server

Proceed with the installation by performing the following steps:

2. Navigate to the directory where you have downloaded the
installation file (such as sServer-5.1.0-x64.run.

3. Give the installation file permission to run by entering
chmod +x SQLstream-sServer-5.1.0-x64.run

4. From a terminal prompt enter ./SQLstream-sServer-5.1.0-x64.run

For Ubuntu systems, enter sudo ./SQLstream-sServer-5.1.0-x64.run