I cannot comment on the stability of revision 62 as I am not personally familiar with it. I have included a list of the issues corrected in revision 62. I hope it will address the issues you have experienced.
At the moment, revision 60 is the most recent developer edition that is publically available. Higher revisions will be made available as soon as possible, as has always been the case. I hope it will be available by the end of this week, and I can check with my colleagues to confirm this.
I locked the other thread because a generic plea does not help us to fix specific issues. The comments and replies here regarding specific problems that users have experienced do help us to fix these issues. The fact that revision 62 itself exists demonstrates our commitment and intention to improve the software.
There was certainly no intention or attempt to censor you; users freely post their issues here and they are addressed by SAP HANA Product Management, and new revisions and features are highlighted here by Product Management and community members as soon as they are available.
I do understand your frustration, and we can certainly try to address the current situation regarding the availability of the new revisions, but clearly the fact that we continue to make new revisions available, as soon as is practical, shows that we take on board the concerns and criticisms of users who need greater stability and functionality.
I hope you feel that this has clarified the situation and addressed your concerns. We are most certainly listening.
Myles
Revision 62 release notes:
•Fixed high memory consumption of Pool/RowEngine/MonitorView
•If there is an insert/upsert trying to insert an already existing record into a columnstore table and a concurrent update on the same record which is rolled back, a duplicate record issue could arise.
•Removed internal length check for lob datatypes which could lead to SQL "inserted value too large for column" errors.
•Fixed a bug wich could crash the statisticsserver. Call stall contains "TrexThreads::ThreadManager::getStatistics" and "InternalHandler::getSimpleStat"
•Fixed wrong decimal value.
•Slow performance encountered during SPO access because filter transport for IN-lists were not supported.
•Due to a logical programming error, performance issue occurs when there are too many identical GROUP BYs being generated when perforing a select on an analytic view.
•Queries with CASE expressions return wrong result if the CASE expression produces a constant number.
•Fixed inconsitent data for column INDEX in the monitoring view HOST_SERVICE_COMPONENT_MEMORY.
•Fixed aborting upgrade during hdbmount.py --sudoers call when using a Storage Connector in a distributed enviroment.
•An indexserver crash might occur when the table is corrupted. The indexserver crash dump contains the following stack trace: "AttributeEngine::Parallel::MultiValueExpansionFilter::filter".
•Cache infrastructure will now distinguish between active and standby node and cache lookup requests are no longer sent to standby nodes.
•Some queries or DML statements with a correlated sub query having a TOP operator may return wrong results. For more details see SAP Note 1885029.
•Some user sessions could be hidden from the HANA DB Studio Session Monitor. This revision enables the Session Monitor to display all sessions created by users.
•Fixed a bug where accessing a temporary table within a stored procedure could result in incorrect data for preceeding selects to this temporary table when the stored procedure is called a second time.
•NOT BETWEEN queries triggered via DBSL (e.g. from BI) can show wrong results for data stored in delta. SQL based queries are not showing this behavior. For details see SAP Note 1893911.
•The Statistics Server is now dynamically obtaining memory needed for "SYS.M_STATISTICS_LASTVALUES" monitoring view. The overall memory consumption of the Statistics Server is reduced.
•Fixed a bug where 'alter system clear column result cache' command does not clear all remote result caches in a distributed system.
•Fixed indexserver crash when querying on an OLAP cube that contains a time dependent join and a pre-aggregation calculated keyfigure.
In the indexserver crash dump the following entry can be found:
"int AttributeEngine::Parallel::_merge<false, TrexTypes::FixedAttributeValue8,AttributeEngine::Parallel::sum<TrexTypes::FixedAttributeValue8> >(AttributeEngine::Parallel::ResultData<TrexTypes::FixedAttributeValue8>&, unsigned int, unsigned int, TRexUtils::CachedMemBlock<TRexUtils::Mapping>::iterator&, AttributeEngine::Parallel::ResultData<TrexTypes::FixedAttributeValue8> const&) + 0x100".
•Fixed indexserver crash at startup caused by an old deprecated procedure paramter (@PARAM@) definition.
In the indexserver crash dump the following entry can be found:
"RTSchema.cpp(01030) : _SYS_RT schema is marked as inconsistent in database. Refusing to change or upgrade content.repositorySetup.cpp(00413) : Permanently setting schema rt as in consistent in database.
RTSchema.cpp(00804) : Failed to update _SYS_RT schema on master host, RT will not be available."
•Fixed a bug where an 'ALTER TABLE ADD PRIMARY KEY' statement could fail with the following error:
"cannot CREATE UNIQUE INDEX; duplicate key found: [5]".
Several documents with the same ID exist in the index if the related table contains more than 2^31 rows (and therefore is partitioned)".
•In the scenario for SAP Business Objects Analysis, the extension to the 'BIMC_VARIABLE_VIEW' for 'VALUE_ENTITY' is added.
•When the ODBC driver of the SAP HANA client (Revision 52 to 61) is used to open a DB connnection, viewing reports on the SAP Business Intelligence platform may cause the Crystal Reports processing server to unexpectedly stop. Also, creating or viewing reports in SAP Crystal Reports based on a HANA ODBC connection established by the ODBC driver of the SAP HANA client (revision 52 to 61) may cause SAP Crystal Reports to freeze.
•Improved performance of queries that need to access "M_ACCESSIBLE_VIEWS" and "M_EFFECTIVE_STRUCTURED_PRIVILEGES".
•Fixed a bug where SQL queries could crash the indexserver. The stack trace contains the following function call sequence:
"TRexConfig::Table::getKeyAttributes", "TRexConfig::CachedMetaData::getKeyAttributeIndices", "TRexConfig::CachedMetaData::initAllKeyAttributeAccessors".
•Fixed a bug where sessions executing long running DML-SQL statements on large column tables like 'CREATE COLUMN TABLE A LIKE B WITH DATA' could not be terminated using either 'ALTER SYSTEM CANCEL [WORK IN] SESSION' or 'ALTER SYSTEM DISCONNECT SESSION' commands.
•Fixed the "Index syncpoint mismatch" error thrown from TRexAPI::DeltaIndexManager::jeReadValueIds4Values that can occur after a delta merge.
•Fixed a bug where executing an OLAP query could crash the indexserver (signal 11). The related call stack contains multiple entries related to: "AttributeEngine::Parallel::GroupByColumnInfoContainer".
•Fixed the issue when querying on OLAP cube and the following error occurs:
"failed with rc 6900; Attribute engine failed;calcEngine search on olapIndex failed."
•Fixed the issue in HDBSQL: /i does not allow file names with spaces properly.
When we are using /i to pass in a filename that contains spaces in HDBSQL, it is not possible to enclose the file name in double nor single quotes.