Issues fixed in MarkLogic Server 10.0-11

Incorrect document filtering with multiple LIKE conditions

When using SQL queries with multiple LIKE conditions, it is possible that the query will return the wrong results.

"Automatic" cache sizing only works for hosts with up to 256 GB of RAM

Automatic cache sizing does not allocate caches using the recommended ratios for hosts with more than 256 GB of RAM.

Journal Archiving on incremental backup does not recover when the disk corruption occurs

When there is a disk failure while archiving journals, journal archiving will not resume until a full backup is performed.

Data missing after restore from backup and forests failover to replicas

There is a potential for data to be lost if a database with local disk failover configured is restored from a backup and the forests fail over to the replicas when the restore completes.

BUG-60265

BUG-57915

Restoring the security database can break https app servers

When restoring a database backup that includes the Security database, if the backup of the Security database does not include certificate templates or external security configurations that are in use by Admin app server (they may have been added after the backup was taken), no warning that the Admin UI could become inaccessible is provided to the user.

The V8 engine causes segfaults under heavy load or overloaded infrastructure

Under heavy load, the V8 engine can timeout when instantiating JavaScript modules and cause MarkLogic to segfault.

CTS queries in JSON payloads are not correctly created by RMA in partition-query and alerting

When doing a PUT of a partition query with an array of CTS queries in JSON format via RMA, only the last query in the array of the queries will be stored in the database.

When doing a POST of an alert rule or a PUT of alert rule properties with an array of CTS queries in JSON format, only the query in the array of the queries will be stored in the database.

BUG-60211

BUG-56402

xdmp:document-filter() does not detect the file type in some cases

When running xdmp:document-filter() against a binary node that is only in memory (has not be retrieved from the database), the file type may not be detected correctly because it does not know the file extension which otherwise comes from the document URI.

sec:amps-change-modules-database throws an XDMP-COMPARE error

The sec:amps-change-modules-database() function does not correctly cast the database ID and throws an XDMP-COMPARE error.

Performance regression for cts:triples() with triple range queries

An optimization for certain cts:triples() patterns that was added in 10.0-9 has caused other triple range queries with cts:triples() queries to be slower.

SSL Secure Client-Initiated Renegotiation is not disabled

SSL Secure Client-Initiated Renegotiation option cannot be disabled.

Replica databases show size zero in the Admin UI

In the Admin UI, when database replication is used, replica databases where database encryption is enabled report that the encrypted size is zero even though the data is loaded.

sjs referencing XQuery modules with the same name may use the wrong module

When multiple XQuery modules with the same filename, but different paths exist in the modules database, SJS modules that reference the XQuery modules using the require function may not retrieve the correct module relative to the calling module.

mlcp has CWE-400 due to 3rd party dependency

mlcp has CWE-400 due to dependency on: com.fasterxml.jackson.core: jackson-core: 2.14.0-rc3

mlcp has a CWE in 3rd party dependencies

mlcp has CWE-611 due to dependency on: org.apache.hadoop: hadoop-common : 3.3.4, org.apache.hadoop : hadoop-hdfs-client : 3.3.4.

cts:triple-range-query() can throw "Critical: Bad malloc"

cts:triple-range-query() throws "Critical: Bad malloc" when too many triple values are provided.

Backup to S3 with object lock and compliance retention does not work

When the object lock and compliance retention is activated on S3 buckets, MarkLogic is unable to backup data to them because of a missing HTTPS header.

BUG-60131

BUG-60480

Select all via ctrl+a is not working in Query Console

In Query Console, select all via ctrl+a is not working correctly when selecting in the query buffer. Once a query is run, ctrl-a stops working in the query buffer. ctrl+a in the results panel only selects one line if a line is clicked on when it should select all of the results.

xdmp:transaction-locks() times out in a cluster when transaction has no locks

When running xdmp:transaction-locks() against a cluster, if there are no transactions, the call will hang and eventually timeout.

The character "&" is being escaped as "&" in the query console "table" output

When running a SQL or SPARL query that outputs a table in query console, & characters are escaped as &.

Private keys are not stripped from support dump when requesting all versions of configuration files

The cleansing process that strips out private keys from support dumps does not strip out private keys when Version is set to All. The private key is only stripped from the Latest configuration files.

sem:query-results-serialize produces invalid JSON when serializing constructed triples

When using sem:query-results-serialize to serialize constructed triples as JSON, the JSON produced may be invalid.

The XDMP-FORESTTAKEN error reports incorrect host ID when the forest label is corrupt

If the forest label file is corrupt, the XDMP-FORESTTAKEN error may be thrown, incorrectly indicating that a host has taken over a forest. This hides actual error which is that the forest label file has been corrupted.

POST /manage/v2/databases/{id|name}/partition-queries adds spurious namespace to element name

When posting certain queries via REST, there is an extra namespace added to the element name when using an element-range-query.