SQL injection in Hive Metastore direct SQL partition-name resolution in Apache Hive before 4.2.1 on all platforms allows authenticated users with access to Hive Metastore APIs to read, modify, or affect unintended partition metadata (including statistics updates, truncation targets, and file-metadata cache operations) via crafted partition names in metastore RPC requests when direct SQL is enabled (the default). Users are recommended to upgrade to version 4.2.1, which fixes this issue.

Details about the issue: Several Hive Metastore RPCs resolve partitions by full partition name (PARTNAME) through direct-SQL helpers. In those paths, client-supplied partition names are embedded into SQL using string concatenation (DirectSqlUpdatePart.quoteString() → '...') instead of bind parameters. A partition name containing a single quote (and crafted SQL) can alter the generated WHERE clause so that lookups intended for one partition match additional rows. That can affect reads, stats updates, truncate targets, metadata-cache targets, and related operations when metastore.try.direct.sql is enabled (default: true). An authenticated or network-trusted caller with the ability to invoke Hive Metastore partition-name APIs against a target table (directly or via Hive/other clients), when direct SQL is enabled can perform this attack. Also, the impact is mainly within table & partition targeting (read/update/truncate/drop/cache the wrong partitions in a table they can reference), not arbitrary cross-database access via this bug alone.

Server-Side Request Forgery (SSRF) in Avro SerDe schema resolution in Apache Hive before 4.2.1 allows an authenticated remote attacker with CREATE TABLE privilege to cause the Hive server to fetch an attacker-controlled URL when resolving the avro.schema.url table property on an Avro table that is subsequently queried. This can expose cloud instance metadata, internal network services, or local server files to the Hive process identity. Users are recommended to upgrade to version 4.2.1, which fixes this issue.

Attacker access requirements:

Network access to HiveServer2 / Metastore: required (remote attacker model). Valid Hive authentication: required. CREATE TABLE (or equivalent) privilege: required, so the attacker can set avro.schema.url in table properties. SELECT privilege on the malicious table: not required for the creator, who can typically query their own table; any other user granted SELECT can also trigger the fetch. Write access to the table LOCATION: not required; the attack uses the schema URL, not the data path. Admin / superuser privileges: not required; an ordinary authenticated user with DDL rights is sufficient. External tables enabled: typically required in practice, and enabled by default in most deployments.

Detection guidance:

Inspect metastore / Hive table metadata for Avro tables whose avro.schema.url uses unexpected schemes such as http, https, file, or ftp, or points at link-local / cloud metadata addresses (for example 169.254.169.254) or other internal hosts. Review HiveServer2 and Metastore logs around CREATE/ALTER TABLE and queries against Avro tables for schema-resolution failures or outbound fetches of avro.schema.url. Correlate CREATE TABLE / ALTER TABLE activity that sets avro.schema.url with subsequent SELECT activity on the same table, especially when the URL target is unusual for schema distribution. On cloud deployments, check instance / VPC flow logs and metadata service access logs for unexpected requests from Hive host identities shortly after Avro DDL or query activity.

An improper authentication vulnerability in HiveServer2 SAML bearer-token validation in Apache Hive 4.0.0 through 4.2.0 (and later unreleased branches) on deployments using HTTP transport with hive.server2.authentication=SAML allows an unauthenticated network attacker to authenticate as an arbitrary Hive user and obtain an authenticated HiveServer2 session via a forged Authorization: Bearer token sent to the /cliservice HTTP endpoint. Users are recommended to upgrade to 4.2.1 version that includes the fix for this issue.

Access / authorization required: No Hive credentials, SAML IdP login, or knowledge of the server signing secret is required. The attacker only needs network reachability to the HiveServer2 HTTP port (typically /cliservice), directly or through a reverse proxy such as Apache Knox that forwards unauthenticated requests to HS2. The instance must have SAML authentication enabled in HTTP mode. Deployments where Knox handles SSO and HiveServer2 uses LDAP/Kerberos (not native SAML mode) are not affected by this specific issue.

Hive creates a credentials file to a temporary directory in the file system with permissions 644 by default when the file permissions are not set explicitly. Any unauthorized user having access to the directory can read the sensitive information written into this file. Users are recommended to upgrade to version 4.0.1, which fixes this issue.

Severity: important

Affected versions:

- Apache Hive 1.1.0 before 4.0.1

Description:

Hive creates a credentials file to a temporary directory in the file system with permissions 644 by default when the file permissions are not set explicitly. Any unauthorized user having access to the directory can read the sensitive information written into this file. Users are recommended to upgrade to version 4.0.1, which fixes this issue.

Credit:

Andrea Cosentino (reporter)

References:

https://github.com/apache/hive https://github.com/apache/hive/commit/20106e254527f7d71b2e34455c4322e14950c620 https://issues.apache.org/jira/browse/HIVE-28134 https://hive.apache.org/ https://www.cve.org/CVERecord?id=CVE-2024-29869

Use of Arrays.equals() in LlapSignerImpl in Apache Hive to compare message signatures allows attacker to forge a valid signature for an arbitrary message byte by byte. The attacker should be an authorized user of the product to perform this attack. Users are recommended to upgrade to version 4.0.0, which fixes this issue.

The problem occurs when an application doesn’t use a constant-time algorithm for validating a signature. The method Arrays.equals() returns false right away when it sees that one of the input’s bytes are different. It means that the comparison time depends on the contents of the arrays. This little thing may allow an attacker to forge a valid signature for an arbitrary message byte by byte. So it might allow malicious users to submit splits/work with selected signatures to LLAP without running as a privileged user, potentially leading to DDoS attack.

More details in the reference section.

Apache Hive before 3.1.3 "CREATE" and "DROP" function operations does not check for necessary authorization of involved entities in the query. It was found that an unauthorized user can manipulate an existing UDF without having the privileges to do so. This allowed unauthorized or underprivileged users to drop and recreate UDFs pointing them to new jars that could be potentially malicious.

Apache Hive cookie signature verification used a non constant time comparison which is known to be vulnerable to timing attacks. This could allow recovery of another users cookie signature. The issue was addressed in Apache Hive 2.3.8

Apache Thrift is vulnerable to a denial of service, caused by improper input validation. By sending specially-crafted messages, a remote attacker could exploit this vulnerability to cause a large memory allocation.

Jodd before 5.0.4 performs Deserialization of Untrusted JSON Data when setClassMetadataName is set.

In Apache Hive 2.3.3, 3.1.0 and earlier, Hive "EXPLAIN" operation does not check for necessary authorization of involved entities in a query. An unauthorized user can do "EXPLAIN" on arbitrary table or view and expose table metadata and statistics.

In Apache Hive 2.3.3, 3.1.0 and earlier, local resources on HiveServer2 machines are not properly protected against malicious user if ranger, sentry or sql standard authorizer is not in use.

This vulnerability in Apache Hive JDBC driver 0.7.1 to 2.3.2 allows carefully crafted arguments to be used to bypass the argument escaping/cleanup that JDBC driver does in PreparedStatement implementation.

In Apache Hive 2.1.0 to 2.3.2, when 'COPY FROM FTP' statement is run using HPL/SQL extension to Hive, a compromised/malicious FTP server can cause the file to be written to an arbitrary location on the cluster where the command is run from. This is because FTP client code in HPL/SQL does not verify the destination location of the downloaded file. This does not affect hive cli user and hiveserver2 user as hplsql is a separate command line script and needs to be invoked differently.

In Apache Hive 0.6.0 to 2.3.2, malicious user might use any xpath UDFs (xpath/xpathstring/xpathboolean/xpathnumber/xpathdouble/xpathfloat/xpathlong/xpathint/xpathshort) to expose the content of a file on the machine running HiveServer2 owned by HiveServer2 user (usually hive) if hive.server2.enable.doAs=false.

Apache Hive 2.1.x before 2.1.2, 2.2.x before 2.2.1, and 2.3.x before 2.3.1 expose an interface through which masking policies can be defined on tables or views, e.g., using Apache Ranger. When a view is created over a given table, the policy enforcement does not happen correctly on the table for masked columns.

Apache Hive (JDBC + HiveServer2) implements SSL for plain TCP and HTTP connections (it supports both transport modes). While validating the server's certificate during the connection setup, the client in Apache Hive before 1.2.2 and 2.0.x before 2.0.1 doesn't seem to be verifying the common name attribute of the certificate. In this way, if a JDBC client sends an SSL request to server abc.com, and the server responds with a valid certificate (certified by CA) but issued to xyz.com, the client will accept that as a valid certificate and the SSL handshake will go through.

The authorization framework in Apache Hive 1.0.0, 1.0.1, 1.1.0, 1.1.1, 1.2.0 and 1.2.1, on clusters protected by Ranger and SqlStdHiveAuthorization, allows attackers to bypass intended parent table access restrictions via unspecified partition-level operations.

The LDAP implementation in HiveServer2 in Apache Hive before 1.0.1 and 1.1.x before 1.1.1, as used in IBM InfoSphere BigInsights 3.0, 3.0.0.1, and 3.0.0.2 and other products, mishandles simple unauthenticated and anonymous bind configurations, which allows remote attackers to bypass authentication via a crafted LDAP request.

Apache Hive before 0.13.1, when in SQL standards based authorization mode, does not properly check the file permissions for (1) import and (2) export statements, which allows remote authenticated users to obtain sensitive information via a crafted URI.