Known Issues and Workarounds in Cloudera Navigator Data Management

Access, Login, or Cloudera Navigator Console Issues

Unlimited description through API only

As of Cloudera Manager version 5.15.0 and 5.14.2, name values can be as long as 500 characters and description values are unlimited. However, the Navigator console limits manual entry of the name to 40 characters and descriptions to 500 characters. When you edit the existing metadata for an entity, if the text is longer than the console limits, you won't be able to save changes to the text.

Fixed in version: Cloudera Manager 5.15.0, 5.14.2, 5.13.3 (Navigator 2.14.0, 2.13.2, 2.12.3)

Cloudera Issue: NAV-6228

Internet Explorer 11 document mode issue

When using Microsoft Internet Explorer 11 to access the Cloudera Navigator console, the login page does not display properly due to an issue with that browser's Compatibility View and certain document modes (see the Microsoft Technet article for details). As of Cloudera Navigator 2.10.1, a warning prompt now displays if the Internet Explorer's document mode is set to IE7 or lower, to alert you to the issue, but prior releases display a blank page rather than the login page.

Workaround: Open the Compatibility View Settings (under the Internet Explorer options menu) and remove the setting for the domain or website. Or use another browser to access the Cloudera Navigator console. Supported browsers include Mozilla Firefox, Google Chrome, or Microsoft Edge.

Affects version: Cloudera Manager 5.11.0, 5.10.x (Navigator 2.10.0, 2.9.x)

Fixed in version: Cloudera Manager 5.11.1 (Navigator 2.10.1)

Cloudera Issues: NAV-3743, NAV-4381

Error message ("not authorized") when logging in

The Cloudera Navigator console saves the state of the last URL accessed when you log out, and opens the same page at your next login. If two different users log in to Navigator using the same browser tab, the state of the first user applies to the second. If the second user does not have permissions to that section of the page, that user receives an error message.

Workaround: Close the browser tab and log in on a new tab. The state is cleared, and the access error message does not appear.

Affects version: Cloudera Manager 5.8.x, 5.7.x (Navigator 2.9.x, 2.8.x)

Fixed in version: Cloudera Manager 5.11 (Navigator 2.10)

Cloudera Issue: NAV-3425

Cloudera Manager Configuration Issues

Adding a blank audit filter removes filter configuration property

In Cloudera Manager, when adding an empty rule to a service's Audit Event Filter and then saving the change, all existing audit event filters are lost. The filter configuration property is removed from Cloudera Manager's list of configuration properties. Reverting the change in the History and Rollback does not restore the previous filters nor reproduce the filter property.

Affected versions: All Cloudera Navigator versions

Fixed in version: N/A

Cloudera Issue: NAV-6096

Overriding safety valve settings disables audit and lineage features

Customers or third party applications such as Unravel may require that hive.exec.post.hooks is configured in a HiveServer2 safety valve. Cloudera Manager will comment out the hive.exec.post.hooks value that is configured if audit or lineage is enabled for Hive. The safety valve content shows the commented code:

<!--'hive.exec.post.hooks', originally set to
'com.cloudera.navigator.audit.hive.HiveExecHookContext,org.apache.hadoop.hive.ql.hooks.LineageLogger'
(non-final), is overridden below by a safety valve-->

This automated change disables Navigator's auditing and lineage features without notification.

Workaround: To fix this problem, manually merge the original HiveServer2 safety valve content for hive.exec.post.hooks with the new value. For example, in the case of Unravel, the new safety valve would look like the following:

<property>
  <name>hive.exec.post.hooks</name>
  <value>com.unraveldata.dataflow.hive.hook.HivePostHook,com.cloudera.navigator.audit.hive.HiveExecHookContext,org.apache.hadoop.hive.ql.hooks.LineageLogger</value>
  <description>for Unravel, from unraveldata.com</description>
</property>

Cloudera Issue: NAV-5331

Cloudera Manager audit events case-sensitive when using PostgreSQL

When Cloudera Manager and Navigator Audit Server are installed using PostgreSQL databases, the behavior of queries run from the Navigator console is different between the two databases. The result is that Cloudera Manager events are returned only if the query values match the case of the event values as they are stored in the Cloudera Manager database.

For example, use "HiveReplicationCommand" instead of HIVEREPLICATIONCOMMAND as it appears in the Audit log.

Cloudera Issue: NAV-6141

Multi-Cluster Environments

Extractors not working in multi-cluster environment

When Navigator Metadata Server is started with a fresh storage directory on versions 5.12.0 or later, extractors are not working when a single Cloudera Manager is managing multiple clusters.

Affects version: Cloudera Manager 5.15.0, 5.14.3, 5.14.2, 5.14.1, 5.14.0, 5.13.1, 5.13.0, 5.12.1, 5.12.0 (Navigator 2.14.0, 2.13.3, 2.13.2, 2.13.1, 2.13.0, 2.12.1, 2.12.0, 2.11.1, 2.11.0)

Fixed in version: Cloudera Manager 5.16.1, 5.15.1, 5.14.4 (Navigator 2.15.1, 2.14.1, 2.13.4)

Cloudera Issue: NAV-6145

Cloudera Navigator is not supported in installations in multi-cluster environments for Cloudera Manager versions 5.12.0, 5.12.1, and 5.13.0

Navigator is not able to uniquely identify more than one cluster in installations of Cloudera Manager that support multiple clusters. This problem only occurs for Cloudera Manager deployments of version 5.12.0, 5.12.1, or 5.13.0. Avoid multi-cluster installations with these releases. This problem does not affect single cluster deployments or Altus clusters.

Affects version: Cloudera Manager 5.13.0, 5.12.1, 5.12.0 (Navigator 2.12.0, 2.11.1, 2.11.0)

Fixed in version: Cloudera Manager 5.14.0, 5.13.1, 5.12.2 (Navigator 2.13.0, 2.12.1, 2.11.2)

Cloudera Issue: NAV-5615

Table-to-HDFS links not established when Navigator supports multiple, high-availability clusters

When Navigator extracts metadata for multiple clusters and when the clusters are configured for high availability operation, Navigator does not correctly link tables to their HDFS backing data. The result is that lineage between Hive tables and their physical data are not created. In addition, some Hive table metadata that is derived from the backing files is not available.

You may see errors in the log such as the following:

2018-02-19 09:01:32,999 ERROR com.cloudera.nav.persist.impl.CompositeLinker [CDHExecutor-0-CDHUrlClassLoader@01010d7e]: Internal error while linking.java.lang.IllegalArgumentException: expected one element but was: <com.cloudera.nav.core.model.Source@b22e68ba, com.cloudera.nav.core.model.Source@5b1bac2c>

Affects version: Cloudera Manager 5.13.x, 5.12.x (Navigator 2.12.x, 2.11.x)

Fixed in version: Cloudera Manager 5.14.0 (Navigator 2.13.0)

Cloudera Issue: NAV-5749

AWS and Amazon S3

Update AWS Credentials from the same AWS account

After configuring Cloudera Navigator with a specific set of AWS Credentials for Amazon S3, any future changes to the credentials (for example, if you rotate credentials on a regular basis) must be for the same AWS account (IAM user). Changing the AWS Credentials to those of a different IAM user results in errors from the Amazon Simple Queue Service (used transparently by Cloudera Navigator).

Workaround: If a new key is provided to Cloudera Navigator, the key must belong to the same AWS account as the prior key.

Affects version: Cloudera Manager 5.10 and later (Navigator 2.9)

Cloudera Issues: NAV-3990

Unnamed folders on Amazon S3 not extracted

Unnamed folders on Amazon S3 are not extracted by Navigator, but the content of the folders is extracted. For example, a top-level folder the top level folder in the bucket has no name (for example, /bucket//folder/file), it is extracted as /bucket/folder/file.

Affects version: Cloudera Manager 5.10 and later (Navigator 2.9)

Cloudera Issue: NAV-3981

Implicit folders are not marked as deleted

If an implicit folder is deleted in Amazon S3, it does not appear as deleted in Cloudera Navigator console.

Workaround: To prevent folders deleted from Amazon S3 from appearing in Navigator Search results, include implicit:false in the search query.

Affects version: Cloudera Manager 5.10 and later (Navigator 2.9)

Cloudera Issue: NAV-3802

Extraction delayed while Amazon S3 inconsistent

Inconsistencies that occur in AWS (for example, due to eventual consistency) can delay Navigator extraction of metadata and lineage from Amazon S3. When Cloudera Navigator detects an inconsistency, extraction may stop until the inconsistency is resolved in AWS. Cloudera Navigator will retry at the next scheduled extraction.

Affects version: Cloudera Manager 5.10 and later (Navigator 2.9)

Cloudera Issue: NAV-4028

Microsoft Azure

Less Secure Credentials Protection Policy can expose Azure credentials in audit logs

When you use Cloudera Manager to configure the ADLS Connector service using the Less Secure option for the Credentials Protection Policy, it is possible for Hive audit logs to include Microsoft Azure credentials. If you are using Navigator Audit Server, these credentials may appear in audit reports. To mitigate this problem, make sure that access to Hive logs is appropriately controlled and that Navigator users with Auditing Viewer roles are cleared to have access to the Hive credentials.

Fixed in version: Cloudera Manager 5.16.1, 5.15.1, 5.14.2 (Navigator 2.15.1, 2.14.1, 2.13.2)

Cloudera Issues: NAV-5861, CDH-56241

Data Stewardship Dashboard

Counts displayed in Dashboard and Search may differ

Counts for databases, tables, views, and other entities displayed in Cloudera Navigator Dashboard can differ from Search values for the same objects. The Dashboard displays data from the Navigator Audit Server, which contains actual counts captured continuously. On the other hand, Search (Solr) returns data that has been periodically extracted from HMS (Hive Metadata Server). Tables, databases, views, or other objects created or destroyed between extracts are not reflected in values displayed by Search (Solr), but they are contained in the audit data displayed in the Dashboard.

Fixed in version: Cloudera Manager 5.11.0 (Navigator 2.10.0)

Cloudera Issue: NAV-4192

Tables Populated count reflects INSERT and UPDATE statements

The Tables Populated display widget in the Data Stewardship Dashboard reflects the number of times that a table has been loaded with data, such as through INSERT and UPDATE statements—not the number of unique tables loaded. For example, a single table to which data is added (through 6 INSERT statements) and that has also had 4 UPDATE statements submitted in the same period would report Tables Populated as 10.

Cloudera Issue: NAV-3886

Search and Managed Metadata

Hive, Hue, Impala, Pig

Hive lineage missing

Because of race conditions within a single JDBC session, Navigator may not show lineage when Hive queries run in parallel within the same session. Hive maintains lineage per session and it clears all lineage information from the session when a query completes. The result is that any query that does not generate lineage will clear the lineage for any other queries that are running in parallel. It can occur for any JDBC clients, such as Hue. Hue uses same session to run user queries and background refresh queries. If you are using JDBC to connect to Hive and you issue concurrent queries using a single JDBC connection, you may encounter this issue.

Affects version: Cloudera Manager 5.12.0 and later (Navigator 2.14.0)

Fixed in version: Cloudera Manager 6.0.0 (Navigator 6.0.0)

Cloudera Issue: CDH-55693

Impala and Hive audit events fail to be captured when one audit event includes 4-byte characters, such as an emoji

This problem applies when the Navigator Audit Server database is a MySQL database configured to use the "UTF8" character set.

When a query includes an emoji or other Unicode supplementary-plane character that is encoded as four bytes in UTF-8, Navigator Audit Server fails to process the event and any following events from the same service.

Workaround: There are work-arounds for this problem; contact Cloudera Technical Support for details.

Cloudera Issue: NAV-4845

Hive service configuration in auditing component

For Hive services, the auditing component does not support the "Shutdown" option for the "Queue Policy" property.

Severity: Low

Workaround: None.

Cloudera Issue: OPSAPS-11537

Hue service audit log and Unknown IP address

The IP address in a Hue service audit log displays as "unknown".

Severity: Low

Workaround: None.

Cloudera Issue: OPSAPS-11986

Changing auditing configuration for Hive or Hue requires restart

Whenever a change is made to a Hive service's audit configuration, Beeswax must be restarted so that the Hue service audit log can reflect the change.

Severity: Low

Workaround: None.

Cloudera Issue: OPSAPS-12274

Viewing Navigator tags in Hue overloads Metadata Server heap

When viewing Cloudera Navigator tags through Hue, Navigator uses more memory than usual and does not release the memory after logging out of Hue. Eventually, the calls between Hue and Navigator will occupy the majority of the heap space allocated to Navigator Metadata Server.

To work around this problem you may need to restart the Navigator Metadata Server periodically to clear the heap usage.

Affects version: Cloudera Manager 5.11.x and later (Navigator 2.10.x)

Cloudera Issue: NAV-4326

HiveServer1 and Hive CLI support deprecated

Cloudera Navigator requires HiveServer2 for complete governance Hive queries. Cloudera Navigator does not capture audit events for queries that are run on HiveServer1/Hive CLI, and lineage is not captured for certain types of operations that are run on HiveServer1.

HiveServer1 and Hive CLI are deprecated in CDH 5. If you use Cloudera Navigator to capture auditing, lineage, and metadata for Hive operations, upgrade to HiveServer2 if you have not done so already.

Cloudera Issue: TSB-185

Navigator Metadata Server

Navigator Metadata Server fails to start with InvalidClassException, local class incompatible

Navigator Metadata Server fails to start (red status in Cloudera Manager) and the Metadata Server log includes the following exception:

2018-09-06 10:38:20,786 WARN jndi [main]:
java.io.InvalidClassException: com.mchange.v2.c3p0.WrapperConnectionPoolDataSource; local class incompatible: stream classdesc serialVersionUID = 7806429541739165290, local class serialVersionUID = -7086951306718003710

Even if Navigator Metadata Server starts once, it may encounter this problem at a later time, depending on which JAR file is loaded.

Workaround: See Knowledge Base Article Navigator Metadata Server fails to start with "local class incompatible" error

Affects version: Cloudera Manager version 5.15.1 (Navigator 2.14.1)

Fixed version: Cloudera Manager version 5.15.2 (Navigator 2.14.2)

Cloudera Issue: NAV-6724, TSB-342

Navigator does not mark HDFS entities as deleted when in bulk extraction takes too long to complete

In large HDFS deployments, the fsimage takes a long time to index. When an HDFS checkpoint occurs, it creates a new fsimage. However if the previous fsimage is still in the process of being indexed, Navigator cannot use the incremental changes found in the inotify stream because it refers to the newly created fsimage.

When this happens, Navigator attempts to start indexing the newer fsimage, creating a loop where Navigator can never take advantage of the more efficient change processing through inotify. The immediate fallout of this delay is that no HDFS entities deleted in the cluster are marked as deleted in Navigator.

Fixed in version: Cloudera Manager 5.16.1 (2.15.1)

Cloudera Issue: NAV-6456

ClassCastException in Navigator Metadata Server log

When a Hive view is created, then dropped, and then subsequently recreated as a table with the same name as that of the original view, the Hive extraction process shows this exception in Navigator Metadata Server logs:

java.lang.ClassCastException: com.cloudera.nav.hive.model.HView cannot be cast to com.cloudera.nav.hive.model.HTable

Affects version: Cloudera Manager 5.14.0 and later (Navigator 2.13.0)

Fixed in version: Cloudera Manager 5.16.1, 5.15.2 (2.15.1, 2.14.2)

Cloudera Issue: NAV-5939

Navigator cannot use PostgreSQL 10.x

Navigator Metadata Server will not work with PostgreSQL versions 10.x. This version incompatibility will result in a message similar to this:

2018-08-06 00:00:22,899 WARN com.mchange.v2.async.ThreadPoolAsynchronousRunner [Timer-1]: com.mchange.v2.async.ThreadPoolAsynchronousRunner$DeadlockDetector@6544a122 – APPARENT DEADLOCK!!! Creating emergency threads for unassigned pending tasks!

Products affected: Cloudera Navigator using PostgreSQL as the database for Navigator Metadata Server

Releases affected: Cloudera Navigator as packaged with Cloudera Manager versions 5.15.0 and 5.15.1.

Users affected: Installations using Cloudera Navigator using PostgreSQL as the database for Navigator Audit Server and Navigator Metadata Server

Severity (Low/Medium/High): High

Impact: Navigator Metadata Server does not function with PostgreSQL database version 10 and above.

Immediate action required: Do not upgrade the Navigator Metadata Server database to PostgreSQL 10.x (allow the Navigator Metadata Server database to remain on a previous working version of PostgreSQL ) or upgrade to future Cloudera Manager release where the issue is fixed.

Affects version: Cloudera Manager 5.15.1, 5.15.0 (2.14.1, 2.14.0)

Fixed in version: Cloudera Manager 5.16.1, 5.15.2 (2.15.1, 2.14.2)

Cloudera Issue: DOCS-3686, TSB-334

Navigator Embedded Solr can reach its limit on number of documents it can store

Navigator Metadata Server extracts HDFS entities by performing a one-time bulk extraction and then switching to incremental extraction. In Cloudera Manager releases 5.10.0, 5.10.1 and 5.11.0 (Navigator releases 2.9.0, 2.9.1, and 2.10.0), a problem causes HDFS bulk extraction to be run more than one time, resulting in duplicate relations created for HDFS. Over time, embedded Solr runs out of document IDs that it can assign to new relations and fails with following error:

"Caused by: java.lang.IllegalArgumentException: Too many documents, composite IndexReaders cannot exceed 2147483519"       

When this happens, Navigator stops any more extraction of data as no new documents can be added to Solr.

The workaround is to upgrade to Cloudera Manager releases 5.10.2, 5.11.1, or 5.12.x (Navigator 2.9.2, 2.10.1, 2.11.x) or later where this issue is fixed. In addition, see Repairing metadata in the storage directory after upgrading.

Affects version: Cloudera Manager 5.11.0, 5.10.1, 5.10.0 (2.10.0, 2.9.1, 2.9.0)

Fixed in version: Cloudera Manager 5.12.0, 5.11.1, 5.10.2 (Navigator 2.11.0, 2.10.1, 2.9.2)

Cloudera Issue: NAV-5600

"Lineage Diagram is limited to 3000 entities" Error

Navigator caps the number of lineage parents that could be used at one time, however this cap included relations with second level parents (columns). The Navigator Metadata Server log includes an error such as "Lineage Diagram is limited to 3000 entities".

Work-around: Set nav.capacity.max_nodes_limit=1000 in the Cloudera Manager configuration Navigator Metadata Server Advanced Configuration Snippet (Safety Valve) for cloudera-navigator.properties. Restart Navigator Metadata Server role and check the lineage. The 1000 value assumes that the lineage only touches max of 1000 tables.

Affects version: Cloudera Manager 5.10.0 and later (Navigator 2.9.0)

Fixed in version: Cloudera Manager 5.16.1, 5.15.1, 5.14.4 (Navigator 2.15.1, 2.14.1, 2.13.4)

Cloudera Issue: NAV-6371

Lineage relations are missing for some views with "Endpoint2 must not be null or empty" error

Lineage relations are missing for some views and the Navigator Metadata Server log includes errors with a stack trace that looks something like this:

2017-10-25 19:00:40,780 ERROR com.cloudera.nav.persist.impl.CompositeLinker [CDHExecutor-0-CDHUrlClassLoader@381e3]: Internal error while linking.
java.lang.IllegalStateException: EndPoint2 must not be null or empty

Note that the exception is “EndPoint2 must not be null or empty”. If you see errors that don’t exactly match this exception, see one of the following similar errors:

While Navigator is determining lineage relations, it found a relation that was created with the target entity missing, hence Endpoint2 is null. This situation can occur when a Hive view is altered without changing the columns extracted from the source table, such as when the name or other table metadata on the view is changed. For example, the following operation would produce the error:

alter view customers_sales_sw set tblproperties("region"="southwest");

This problem can occur in Navigator Metadata Server versions 2.9.0, 2.9.1, 2.9.2, 2.10.0, 2.10.1, 2.10.2, 2.11.0, and 2.11.1. These versions correspond to the following Cloudera Manager versions: 5.10.0, 5.10.1, 5.10.2, 5.11.0, 5.11.1, 5.11.2, 5.12.0, and 5.12.1.

Eventually you'll want to upgrade away from an affected Navigator version. If you aren't able to upgrade right away, you can ignore this error with the consequences that your logs are noisy and fill up more quickly. See the Knowledge Base article EndPoint2 must not be null or empty for more information.

Affects version: Cloudera Manager 5.12.1, 5.11.1 (Navigator 2.11.1, 2.10.1)

Fixed in version: Cloudera Manager 5.13.x, 5.12.2, 5.11.2 (Navigator 2.12.x, 2.11.2, 2.10.2)

Cloudera Issue: NAV-5661, KB-713

Log includes the error "EndPoint1 must not be null"

The following error may appear in the Navigator Metadata Server log:

2017-10-17 13:00:23,007 ERROR com.cloudera.nav.hive.extractor.AbstractHiveExtractor [CDHExecutor-0-CDHUrlClassLoader@14784b7b]: Unable to parse hive view query *: EndPoint1 must not be null or empty
java.lang.IllegalStateException: EndPoint1 must not be null or empty

This error occurs because the Hive pull extraction for creating a Hive view produces an incorrect lineage relationship for the Hive view. However, Navigator also receives information for the view creation through the push extractor, which correctly produces the lineage relation. You can safely ignore this error. Note that it is distinct from Lineage relations are missing for some views with "Endpoint2 must not be null or empty" error.

Affects version: Cloudera Manager 5.10.x and later (Navigator 2.9.x)

Cloudera Issue: NAV-4224

Log includes the error "EndPoint2 EntityType must not be null unless unlinked"

The following error may appear in the Navigator Metadata Server log:

2017-10-26 15:48:27,440 INFO com.cloudera.nav.hdfs.extractor.HdfsOperationHandler [CDHExecutor-0-CDHUrlClassLoader@3eca0c78]: Unable to process rename for path /hbase/data/default/img/f6572262455416bff42f92fd2b0e75c0/.tmp/1cc1af7b0f784572
b821980f6b4c5adc: can't find source information.
2017-10-26 15:48:27,440 ERROR com.cloudera.nav.hdfs.client.InotifyClient [CDHExecutor-0-CDHUrlClassLoader@3eca0c78]: Error handling event (txid: 2020059072): Renamed /hbase/data/default/img/f6572262455416bff42f92fd2b0e75c0/.tmp/1cc1af7b0f
784572b821980f6b4c5adc to /hbase/data/default/img/f6572262455416bff42f92fd2b0e75c0/D/1cc1af7b0f784572b821980f6b4c5adc at time 1509049550134
java.lang.IllegalStateException: EndPoint2 EntityType must not be null unless unlinked
   at com.google.common.base.Preconditions.checkState(Preconditions.java:149)
   at com.cloudera.nav.core.model.Relation.validate(Relation.java:287)

This error occurs when HDFS files are renamed and the new name is affected by an exclusion filter. The only consequence of the problem is noise in the log. Note that it is distinct from Lineage relations are missing for some views with "Endpoint2 must not be null or empty" error.

This problem prevents new lineage information from being added to existing and new Hive entities.

Affects version: Cloudera Manager 5.7.x and later (Navigator 2.6.x)

Fixed in version: Cloudera Manager 5.14.x (Navigator 2.13.x)

Cloudera Issue: NAV-4654

Purging Deleted Properties shows "Bad Request" error when viewing Hive entity details

After upgrading to Navigator 2.11.0 (Cloudera Manager 5.12.0) and later versions (to Navigator 2.13.0), removing deleted managed metadata properties with the "Purge deleted Properties" command causes a mismatch between the Solr schema and the data stored for entities. (Administration > Managed Metadata).



The result is that the Navigator console shows the error "Bad Request" when displaying details for Hive entities such as databases, tables, or fields.

When this problem occurs, the Navigator Metadata Server log will include a message such as the following:

[CDHExecutor-0-CDHUrlClassLoader@62d310e8]: Internal error while linking. java.lang.IllegalArgumentException: java.lang.ClassCastException@2f145bbc          

To avoid this problem, do not purge deleted properties; allow these properties to stay in the list of properties.

Affects version: Cloudera Manager 5.12.0 and later (Navigator 2.11.0)

Fixed in version: Cloudera Manager 5.15.0, 5.14.2, 5.13.3 (Navigator 2.14.0, 2.13.2, 2.12.3)

Cloudera Issue: NAV-5982

Metadata Server log file and spurious error messages

Certain configurations of OS and database (such as PostgreSQL and Ubuntu Linux) may raise spurious error messages regarding non-existent files. Such messages can be safely disregarded, such as this example:
Error: [main]: PWC6351: In TLD scanning, the
supplied resource file:/usr/share/java/oracle-connector-java.jar does not
exist.

Affects version: Cloudera Manager 5.9.x, 5.8.1 (Navigator 2.8.x, 2.7.1)

Fixed in version: Cloudera Manager 5.10.0 (2.9.0)

Cloudera Issue: NAV-698

Purge

Fail to Purge Hive and Impala Select Operations

In some cases, purge code makes large query requests to solr. In those cases, it can fail with the following error followed by a stack trace:

2018-09-22 23:39:55,736 ERROR com.cloudera.nav.maintenance.purge.hiveandimpala.PurgeHiveOrImpalaSelectOperations [scheduler_Worker-2]: Failed to purge operations for DELETE_HIVE_AND_IMPALA_SELECT_OPERATIONS with error
        org.apache.solr.client.solrj.impl.HttpSolrServer$RemoteSolrException: Expected mime type application/octet-stream but got application/xml.

Work-around: To work around this problem, lower the batch size for Navigator commits to its Solr collection by changing the value of nav.solr.commit_batch_size to be lower.

Affects version: Cloudera Manager 5.12.0 (Navigator 2.11) and later

Cloudera Issue: NAV-6789

Purge operation leaves Kite errors in log

Navigator Metadata Server attempts to handle Kite datasets even though Kite datasets are no longer supported. This problem may affect operations such as metadata purge. If you encounter Kite dataset errors such as the following, you need to manually disable Kite dataset extraction.

2018-09-23 16:10:43,678 ERROR com.cloudera.nav.hdfs.datasets.KiteDatasetReader [hdfs-reader-1]: Dataset expected but not found at /staging/example/.temp/job_1537417909723_0201/mr/job_1537417909723_0201org.kitesdk.data.DatasetNotFoundException: Descriptor location does not exist
To disable Kite dataset extraction, set the following Navigator Metadata Server property in Cloudera Manager, in the Navigator Metadata Server Advanced Configuration Snippet (Safety Valve) for cloudera-navigator.properties:
 nav.extractor.hdfs.datasets.enabled=false

Affects version: Cloudera Manager 5.14.0 (Navigator 2.13) and later

Cloudera Issue: NAV-6965

The flag for disabling purge is ignored in the UI

The Administration > Purge Settings > page may show scheduled purge operations even if the nav.purge.enabled property is set to false. The displayed purge operations will not run.

Affects version: Cloudera Manager 5.14.0, 5.14.1 (Navigator 2.13.1, 2.13.0)

Fixed in version: Cloudera Manager 5.14.2 (Navigator 2.13.2)

Cloudera Issue: NAV-5966

Policy specifications and cluster names affect purge

Policies cannot use cluster names in queries. Cluster name is a derived attribute and cannot be used as-is.

Workaround: When setting move actions for Cloudera Navigator, if there is only one cluster known to the Navigator instance, remove the clusterName clause.

If there is more than one cluster known to the Navigator instance, replace clusterName with sourceId. To get the sourceId, issue a query in this format:
curl '<nav-url>/api/v9/entities/?query=type%3Asource&limit=100&offset=0'
Use the identity of the matching HDFS service for this cluster as the sourceId.

Cloudera Issue: NAV-3288

Spark

Spark Lineage Limitations and Requirements

Spark lineage diagrams are supported in the following releases:

  • Spark 2.3 as of Cloudera Manager 5.14.0 (Navigator 2.13.0)
  • Spark 1.6 as of Cloudera Manager 5.11.0 (Navigator 2.10.0)
Lineage is not available for Spark when Cloudera Manager is running in single user mode. In addition to these requirements, Spark lineage has the following limitations:
  • Lineage is produced only for data that is read/written and processed using the Dataframe and SparkSQL APIs. Lineage is not available for data that is read/written or processed using Spark's RDD APIs.
  • Lineage information is not produced for calls to aggregation functions such as groupBy().
  • The default lineage directory for Spark on Yarn is /var/log/spark/lineage. No process or user should write files to this directory—doing so can cause agent failures. In addition, changing the Spark on Yarn lineage directory has no effect: the default remains /var/log/spark/lineage.

Cloudera Issue: OPSAPS-39589

Navigator doesn't recognize local files in Spark jobs

Spark jobs can use files on the local filesystem as job inputs or outputs. Navigator, however, only supports HFDS, Hive, and S3 assets as job inputs or outputs. When Navigator extracts metadata from Spark and encounters a local source type, the metadata is discarded and the following error appears in the Navigator Metadata Server log:

2018-10-11 12:14:26,192 WARN com.cloudera.nav.api.ApiExceptionMapper [qtp1574898980-23815]: Unexpected exception.
          java.lang.RuntimeException: Source LOCAL isn't supported for Spark Lineage

Affected Versions: Cloudera Manager 5.11 and later (Navigator 2.10)

Fixed in version: Cloudera Manager 5.16.2 (Navigator 2.15.2) Note that the fix does not provide support for extracting metadata for local files for Spark jobs; it prevents the error.

Cloudera Issue: NAV-6811

Spark extractor enabled using safety valve deprecated

The Spark extractor included prior to CDH 5.11 and enabled by setting the safety valve, nav.spark.extraction.enable=true is being deprecated, and could be removed completely in a future release. If you are upgrading to CDH 5.11 from a deployment that had configured this safety valve, be sure to remove the setting when you upgrade.

Upgrade Issues and Limitations

Issues and limitations matrix by release

Before upgrading Cloudera Navigator, review these version-specific release notes:

Upgrade Limitations, requirements, preliminary tasks
From... To...
2.14 (and lower) 2.15.1 (and higher) The default filters set for auditing HDFS changed in Cloudera Manager 5.16.1 (Navigator 2.15.1). If you've updated your HDFS audit filters from previous defaults, the new filters are not applied. Whether or not the filters are changed in your upgrade process, we recommend that you review your HDFS audit filters to ensure you are collecting the audit events that add value in your environment. See Improvements to Audit Filters for more information.
2.10 (and lower) 2.11.0, 2.11.1, 2.12.0

When using Navigator in multi-cluster environments, avoid upgrading to Cloudera Manager deployments of version 5.12.0, 5.12.1, or 5.13.0 due to a known problem where Navigator does not recognize more than one cluster. Instead, upgrade to Cloudera Manager release 5.12.2, 5.13.1, or 5.14.0 (Navigator 2.11.2, 2.12.1, or 2.13.0). See Cloudera Navigator is not supported in installations in multi-cluster environments for Cloudera Manager versions 5.12.0, 5.12.1, and 5.13.0 for more details.

This problem does not affect single cluster deployments or Altus clusters.

2.8 (and lower) 2.9.0, 2.9.1, 2.10.0 Avoid upgrading to Cloudera Manager releases 5.10.0, 5.10.1 and 5.11.0 (Navigator releases 2.9.0, 2.9.1, and 2.10.0) due to the known problem causing the storage directory to fill beyond its capacity. The workaround is to upgrade to Cloudera Manager release 5.10.2, 5.11.1, or 5.12.x (Navigator 2.9.2, 2.10.1, 2.11.x) or later, where this issue is fixed. See Navigator Embedded Solr can reach its limit on number of documents it can store if you are already running one of these releases.
2.10 2.10.1 Upgrading Cloudera Manager 5.11.0 (Cloudera Navigator 2.10) to Cloudera Manager 5.11.1 (Cloudera Navigator 2.10.1) results in failures by Navigator Audit Server to publish to Kafka. See Publishing to Kafka fails after upgrade for details and a workaround.
2.9 (and higher) 2.10 (and higher)

As described later in this table, upgrading across Cloudera Manager 5.10 (Cloudera Navigator 2.9) involved a significant metadata change and therefore the upgrade was very slow. Upgrades between Cloudera Manager 5.10 and later releases do not have a metadata change and the Navigator upgrade is quick.

2.8 (and lower) 2.9 (and higher) Upgrading to Cloudera Navigator 2.9 and higher (Cloudera Manager 5.10 and higher) can take a significant amount of time, depending on the size of the datadir. Before starting to upgrade to Cloudera Manager 5.10 (which automatically starts the upgrade to Cloudera Navigator 2.9), see Upgrading Cloudera Navigator Can be Extremely Slow. Briefly, in this release Solr indexing has been optimized to improve search speed. When the upgrade process completes and Cloudera Navigator services re-start, the Solr indexing upgrade automatically begins. No other actions can be performed until Solr indexing completes (progress message display during this process).
2.6 (and lower) Any The Cloudera Navigator Metadata Server requires an upgrade of data in the storage directory. See Upgrading Cloudera Navigator for details.
2.4.0, 2.4.1 2.4.2 After upgrading, requires manual modification to the HDFS extractor state file to change status of UNDELETE taskTypes from SUCCEEDED to FAILED. This manual process is required only for these specific releases listed. See Issues Fixed in Cloudera Navigator 2.4.2 for details.
Any 2.4 – 2.6 Cloudera Navigator 2.4 – 2.6 do not support JDK 1.6, so upgrade any instances of JDK 1.6 to JDK 1.7 or 1.8 before upgrading to these releases. See Step 2: Install Java Development Kit for details.
2.1 2.2 Cloudera Navigator 2.1 used the beta version of the Navigator Metadata Server policy engine. Policies created using that version are not retained during the upgrade.
2.0 (and lower) 2.1 (and higher) The upgrade wizard for this upgrade path adds the Navigator Metadata Server to the Cloudera Manager cluster. The Navigator Metadata Server is new, and is not the same as the existing Navigator Audit Server database.
1.2 2.0 Cloudera Navigator 1.2 and Cloudera Navigator 2.0 have reached EOL (end of life) status and are no longer supported. There is no wizard for this upgrade path. Cloudera Navigator 2.0 required a clean install. The Navigator 1.2 Navigator Metadata Server was a beta release included with Cloudera Manager 5.0. The 1.2 version of the Navigator Metadata Server role must be removed before the cluster can be upgraded to Cloudera Navigator 2.0 (the roles are not compatible).