Ambari Release Notes
Also available as:

Fixed Issues

The following table lists selected issues resolved in Ambari 2.2:



Apache Jira

HWX Jira

Summary fails on hosts filter for large clusters back to "Customize Services" page from "Review" page on installer/ASW loses all static attributes of properties info written to site configs by UI attribute provision_action=INSTALL_ONLY loses its value after server restart Exception on cancelling server-side custom action and custom views are not working when swebhdfs and Namenode HA is enabled LDAP sync process support for ViprFS filesystem in Ambari Views browser : File preview show first character only when ssl is enabled there is a host with no components, then host page spins Status commands should enforce a timeout < heartbeat interval ability in AMS to filter tracked metrics through a whitelist metric file request execution takes too much time wizard on large clusters is painfully slow while updating stale configuration cache during upgrade JMX protocol values on large cluster takes too long client requests handled by Jetty should not be blocked by JMX property providers View synchronized around entire connection creation; causing deadlock configuration cache expiration caused 10,000's of database hits in large deployments heatmap API call takes over 1 minute to complete on 900-node cluster & mapreduce.jobhistory.webapp.https.address properties are not respected should work in SPNEGO-enabled clusters the ability to perform interactive shell commands from Ambari server-side actions can't be started because of failed webhdfs request server start failed during EU with BindException of the Node Managers fail to start after upgrade host throws 500 Error with unique constraint violation 'Config group' changes YARN Config Group overrides and vice versa Summary screen does not show hostname in tooltip pop-up server actions should not timeout in minutes as specified in configuration wizard loses request id on server restart with many hosts can create alerts with text too large to be stored fails to start nodemanager due to unexpected return code from sudo su command on the pid file Service wizard breaks on large clusters when persisting data in localStorage does not support HBase on HTTPS mode show NA after enabling HTTPS for YARN, NodeManager JVM Metrics on YARN View caching issue causes File does not exist: /user/pig/jobs/job_id/stdout and stderr maps can include dashes and other non-word characters maps for Storm are not being escaped correctly Agent memory leak fix configuration flag to prevent changing of directory permissions[Grafana] Updated dashboards for TopN feature[Grafana] topN implementation for templatized dashboards ability to query all hosts using hostname=% through AMS API TopN does not work when metric name has a wildcard specified TopN queries in AMS sync: use uid and cn in patterns to check a member is a dn or not sync cannot handle if the member attribute value is not DN or id is placed on every host and path is not customizable DataNode Unmounted Alert produces false alerts when file:// URIs are used WebHDFS over SSL in Ambari Views SNMP dispatch behavior if the reusable UDP transport goes away UDP SNMP disptacher port to be configurable ability to ignoreBadMounts on a large cluster with bad disks vs workaround of tens of Config Groups PUT fails on Amazon Linux script on Pig Views encountered an IOExecption as default FS is not supported for any Views ability to pass JPA / EclipseLink properties to the DataSource logging for ActionQueue's retry logic processor should create ConfigGroup even with only one host registered Gateway fails to start with /tmp/.hdfs-nfs warning in logs advisor AMS memory validation seems wrong configs rotate oozie logs that are too small HBase Region servers alert should determine RegionServer liveliness differently Oozie limits.conf for Oozie User as done for HDFS/YARN API does not honor JPA sorting not showing active/standby status for NameNodes when custom ports are used role command order Start/Stop dependency between Metrics Collector and ZooKeeper Server UI does not accept entries with file protocol when enabling 'storage type' feature aggregates counter values as average over the time series (and other issues) HiveServer2 should not change "hadoop.proxyuser.webhcat.hosts" property metrics not written due to permissions on metrics properties file metrics emit interval seems to be 1 minute instead of 10 seconds default a Service results in deleting Config Group mappings (more than 1 CG present) checks and alerts when clusterconfigmapping has multiple selected entries for a config type change hive.metastore.warehouse.dir to S3 with upper-cased symbols was not registered during deploy via blueprints server does not honor the gzip compression correctly are not triggered when ambari agent is down HDFS metrics per RPC port number in AMS and Grafana View: Using multiple 'set' commands in query does not set properties properly view: Visual Explain does not work when there are multiple statements in the query editor jdk version being used by HDP does not work specify ZooKeeper heap size via UI spark.hadoop.cacheConf=false when starting spark thrift server is expecting hadoop client package and configuration, even if no HDFS components are installed on this host resource leak in the metrics collector code modification exception when syncing from ldap Yarn graphs,Heatmaps take longer to load large cluster clicking on 'Install Packages' does not show the 'Installing' progress bar 700 hosts gets stuck without timeout on 8 hosts And c3p0 Queries Can Hang Ambari On Large Queries the default http header size of the Ambari server to 64k, AMBARI-14847BUG-50104Concurrent kinit Commands Cause Alerts To Randomly Trigger kerberos does not remove auth to local rules Server CA should use sha265 as default message digest algorthm logging for when ambari does not have permissions to read the ambari server keytab forcing users to lowercase on ldap sync cluster with RM HA contain incrorrect proxyuser yarn properties API directive to force toggling Kerberos if the cluster's security type has not changed ranger caused proxy settings for user hive in hdfs to get updated user-specified auth-to-local rules fail to render when auto generating auth-to-local rules of a failed component should ask for Kerberos password force_toggle_kerberos directive when retrying enable Kerberos on failure key violation syncing groups when performing ldap-sync with --all option API ignores "fs.permissions.umask-mode" property error message when trying to get defaultRealm in non kerberized environment components failed to start when deploying secured cluster via blueprint overwrites permissions on HDFS directories override trustore when executing "ambari-server setup-ldap" lost trustore when executing "ambari-server setup-ldap" - Managed KDC Credentials is not saving the Admin Principal and Admin Password when Adding Services. RS decommission issue in secure cluster property: hive.server2.authentication.kerberos.keytab Tickets Expire Too Frequently For Alerts cannot enable NN HA in kerberos env re-execution is required after entering the password for LDAP Kerberos: JS error on clicking next button of "Configure Identities page" on Add Service wizard ssl for ambari does not update the tez view url in tez-site.xml Kerberized JournalNode requires HDFS principal to perform 'role edits' task creating WEB URLs for Alerts, if we add http/https then we should not drop the URL path past port number tries to create local users/groups with a Kerberized cluster shows "Corrupted Blocks" but it is "Corrupted replica" HA Wizard should not start if any service is in maintenance mode NameNode-HA properties still in hdfs-site.xml after enabling HA deployed Cluster provisioning is failing due to DataNode start View: "NO DAG Available" on TEZ tab when ambari is running behind proxy Monitor crashes with NoSuchProcess exception on RHEL/CentOS 7.2 breaks ambari-server startup timeout for server-side tasks Server should not use /tmp for extraction cached conflicts with python hdfs lib resulting into monitoring errors Alert Parameters With Shell Characters Are Not Properly Escaped not executing for table mentioned in the editor, but rather what is selected. in ambari server log related tez view monitor fails on restart on TimelineMetricCache in Nodemanager bootstrap/registration fails in multi-byte language environments in hadoop-env for YARN does not require YARN restart in Ambari fails to start: IOException: /ats/active does not exist API fails with multi-value filter kms install fails if log dir does not exist commands temporary break host component state machine't copy fast-hdfs-resource.jar if sys_prepped tag for StandBy name node should not be changed to NameNode if there is an active namenode Service Check fails due to timeout after deployment on manual start of AMS is misleading View : Upload table not handling detached cluster properly service check fails after removing a host 2.3.4 package install through Ambari on SLES11 hosts fails HBase master heap recommendations Scheduler view is not working when using custom configuration - Cannot insert host that belongs to multiple topology_hostgroups NameNode-HA properties still in hdfs-site.xml after enabling HA (scripts refer to non-existent property values) to restart Falcon server upstart script broken in RHEL6 returns error 'unknown job' for the ambari-agent enabling SSL for Resource Manager, getting a constant Ambari Alert: Connection failed to (No JSON object could be decoded) Namenode Move failed on Delete disabled NameNode+ZKFC stage API does not return HDFS RPC metrics based on ports wizard failed on Delete component operation if component is in MM is setting very small limits for kafka causing failures with too many files open error cluster creation using Blueprints the cluster creation request has incorrect COMPLETED state instead of PENDING. initial NN start runs into problem while communicating with JN, restart does not work execution should only be allowed on commands that have auto retry enabled disabling RU/EU PreChecks, Oozie service check failures after upgrade finished icon for service check during rolling upgrade while Skip all Service Check failures enabled ambari-server from 2.0.1 to 2.2.0 sets incorrect value for 'storm.zookeeper.superACL' - cannot "Pause Downgrade" during RU Upgrade: upgrade group status not updated error message when installing registered version which does not have repo defined for the OS Oozie service check failed after downgrade with multiple Oozie servers since rerunning prepare-war removes webapps/oozie folder overwrites auth_to_local rules in core-site.xml Upgrade should not automatically add stack properties that reference localhost like dfs.namenode.secondary.http-address, mark with DONT_ADD_ON_UPGRADE Server upgrade makes unwanted changes to capacity-scheduler.xml Server Upgrade adds unneeded Atlas properties even though Atlas is not needed, causing forced Hive restart has to calculate ranger-env config ranger_privelege_user_jdbc_url can't start if hosts have name in MixedCASE in configs for NameNode, HBASE Master, ResourceManager History server does not start after downgrade due to missing spark-events directory Upgrade reported 'tuple concurrently updated' error during restart of Ranger Admin in HA setup incorrect after Ambari 2.2 upgrade executable not present after upgrade to 2.2.2 on rpm-based systems upgrade fails if there is incorrect version in upgrade table commands throw an error after upgrading to HDP 2.4 from HDP 2.3 issues during EU from HDP to on SLES 11 upgrade/start commands should detect any ambari-server hotfix and fail upgrade/start from proceeding EU from 2.2 -> 2.3 HBase region server started failed with memory error To Ensure That All Components Are On The Same Version Before Upgrading Upgrade: incorrect display_status of upgrade groups EU Some Services Fail To Start Because of Missing JKS Files upgrading from Ambari 2.1.2 to 2.2.1, HDP from 2.3.0 to 2.3.4 , user is getting 'Consistency Check Failed' error while trying to save configuration in Ranger, Hive and HBase HDFS Before Upgrade Finalizing Does Not Supply the rollingUpgrade Flag more warning on Pause button saying no topology changes. calculates stack downgrade as ABORTED under incorrect conditions. UI shows 'Downgrade paused' and button to resume downgrade even when progress is happening RU/EU is paused, services are restarted on the older version. EU is more complex since stopping services should use the original version. jaas files are not generated on a Kafa rolling upgrade if is not set to PLAINTEXTSASL in secure cluster Storm widgets after ugprading to Ambari upgrade fails with MySQL DB and enforce gtid enabled - cannot resume upgrade after pausing on a manual task Some service checks were failed during RU Precheck fails on Ranger Admin with SSL Because Of Missing Truststore many ticks on the graph if large custom time range is specified doesn't load on Summary page layout after new ones are created their are multiple Job History Servers in the cluster, QuickLinks should show the URL for all instances Value in first cell coming up as null for select query names different on host details page and filter Host Check for Ganglia directories"No Data No available data for the time period." instead metrics widget on host details page for 1 year period behavior of period combobox/metrics on Dashboard page after resetting all widgets to default changing the time range for displaying graphs, UI does not show feedback (doesn't update and keeps showing the same graph) when data is not available adds config properties that are not necessary background operations have incorrect order hostnames are not truncated in Alerts dialog restart icon appeared after 5 minutes after reconfig of Ambari DB for postgres says to use wrong .sql script field is displayed instead of Hive default query queue dropdown emits critical exceptions in debug mode only Ambari should support custom stack names  Phoenix command line tool support  ambari agent upstart support Tez Ambari View: Add protocol configuration for YARN ambari-agent upstart script restart triggers the restart of hbase specific JVM processes Inconsistent HIVE CBO setting presentation Make amazon2015 to be part of redhat6 family ambari server upstart support Oozie untar and prepare-war should be only done exclusively in  Remove the dependency between ZK start and Ranger user sync start should user hdfs_path_prefix Add PHD-3.4 and above stack definitions in pluggable stack config Don't copy fast-hdfs-resource.jar if sys_prepped Make AMS HBase initialization check optional through a config  Ambari preupload should skip uploading oozie shared lib for upgrade  Change AMS configs to deal with slow writes to WASB  Decrease the load on ambari database after cluster creation  Fix unreasonable default heap settings for AMS HBase heapsize and xmn size. should pre-create hdfs directories  Ambari's App Timeline Web UI Alert tries make http heavy call timelineserver:httpPort which can also cause OOM  $.timeago for alert instance and alert definitions should use not converted with timezone timestamps  TimelineServer configuration is missing in yarn-env.xml. Blueprints: Stack advisor should consider user provided inputs on ''ONLY_STACK_DEFAULTS_APPLY" strategy  Ambari server throws error when unused configs are present in blueprint  Localhost configuration value for multiple host properties does not work properly  Add role command order Start/Stop dependency between Metrics Collector and ZooKeeper Server  NPE when configuring Kerberos at provisioning secure cluster with Blueprint RU/EU: Upgrading Oozie database fails since new configs are not yet written to /usr/hdp/current/oozie-server/conf Adding a Service results in deleting Config Group mappings (more than 1 CG present) Metrics monitor fails on restart Oozie sharelib missing SQL JDBC driver Ambari missing metrics in UI with Vip settings enabled. Oozie server unable to load JDBC driver Start all services request aborts in the middle and hosts go into heartbeat-lost state Resource Manager fails to start: IOException: /ats/active does not exist Unusable configs after creating override. Config version remains is_current=true even after it's group is deleted  New Alerts Do Not Honor Existing Maintenance Mode Setting RU/EU has to calculate ranger-env config ranger_privelege_user_jdbc_url Failures in Express Upgrade from HDP 2.2 to HDP 2.5 due to missing modules Ambari Schema Version should ignore patch number from version number 'ambari-server start' is failed [import FatalException, NonFatalException] after upgrade from 2.1.2 ... etc  to Express Upgrade Stuck At Manual Prompt Due To HRC Status Calculation Cache Problem  When there are multiple Job History Servers in the cluster, QuickLinks should show the URL for all instances  Tez Ambari View: Improve host configuration logic Convert Storm and Kafka to use enhanced dashboard widgets monitor should use same hostname as ambari agent and server. node https results JournalsStatus not available in the API stack advisor validation for AMS distributed properties changing the hostname on the HA cluster the ZKFC's metadata still uses old hostname. failed to change the host name from Upper to Lower case for the cluster download client configs says config generation failed want to retry when it actually worked View: Config changes(Local <--> Custom) are not reflected unless ambari-server is restarted hdfs metrics do not use active NN view : explain script contains errors on secure cluster ATS show alert at single node cluster via blueprints without slave components config properties get reverted when adding new host or host component (secure cluster) dir for metrics server does not get updated after NN HA is enabled upgrade times out when /app-logs is too large calls to /requests become very slow after installing packages on 900 node cluster Install packages call takes more than 3 minutes on 900 node cluster the case when UI call to 'Install Packages' action times out to start/stop all services are very slow automation logic should use stack advisor when determining configuration updates add new service when kerberos is enabled w/ blueprint specifical properties reverted during secured cluster deploy via blueprint for headless principals with cached keytab files are changed unnecessarily kerberos security after express upgrade has failed at Start service step because of slider client xml permission. KMS install from Ambari - Add ability to skip DB/DB user setup was installed with incorrect permissions on agent host setpgid failed issued when starting ambari server Kerberos deployments fail intermittently due to invalid keytabs View does not honour auth_to_local rules when running queries, AMBARI-14664BUG-48337Blueprints: After Kerberos deployment, UI requires that Yarn service be restarted Requests Via nslcd Take Too Long In Some Organizations region server restart ambari is setting /tmp to hbase:hadoop does not return host component metrics (except for one host) storm fails with a metrics storm sink jar related error sometimes should return 400 (instead of 500) when the hive query cannot be compiled, add PreCheck to prevent starting upgrade if a previous upgrade skipped the step Save DB State agent dir is not recreated if /var/run/<ambari_dir> is deleted on non-root agent advisor: ConcurrentModificationException with blueprint provisioning not showing active/standby status for Resource manager when custom ports are used Service: deploy fails due to duplicate components create Permission is not applied properly while using fast-hdfs-resource.jar the JDBC URL for Oracle contains a non-standard port, some changes cannot be made via Smart Configs UI for Hive MetaData server start fails due to missing package while adding Atlas service after Hive service start fails when time taken to get out of safemode is more than 20 minutes. INSTALL failed Metastore alert timeout Agent should not rely on CWD of user who started it to Open a Saved Query Returns Unexpected Results upgrade to HDP 2.4 has failed on Calculating Yarn Properties for Spark upgrade from 2.3.2 to 2.3.4, HBase fails to start, remove props for local indexing version check doesn't work in case of existing database from older Ambari version error during Hive Server restart while EU from Dal to Dal M10 to Dal M20 error during Hive Metastore restart while EU from 2.2.8 to on SLES ThriftServer Does Not Upgrade shows down after downgrade to 2.2.9 due to schema version mismatch executeHostComponentStateDDLUpdates is not idempotent failure not shown on all levels view disappeared after upgrading from ambari 2.1.2 to START failed post EU add error handler for "Pause upgrade" in non default config group do no get prompted for restart to update stale configs update storm zk quorum from ambari UI zk servers post install did not update the zk servers in other services reports wrong count after manual stack upgrade Oracle DB connection string when not running port 1521 Oozie Servers can be added with Install Wizard and ASW View has incorrect class org.apache.hcatalog.pig.HCatLoader(); being loaded with the pre is no need to put 700 for /user/hive directory, or at least make it configurable ambari-server silent-mode setup does not switch DB - add unit tests 7. Host check does not show warning about THP enabled on hosts. cannot add name-value pair to an empty config View [Visualization] Enhancements script does not work for hosts that are manually registered nimbus.monitor.freq.secs=120 in ambari call to load Host Component links on Service dashboard should be independent of metrics. not log "unable to get latest URL" as ERROR Web should surface HTTP 413 error Hive View should be able to talk HTTP to HS2, AMBARI-13345BUG-46283Views: HDFS API utility needs to work with WASB commands given by Ambari are incorrect when yarn.timeline-service.leveldb-timeline-store.path is non-default links to Ganglia Web UI does not install phoenix if one enables phoenix after cluster install is done point in time queries to JMX endpoints hbase-env config without removing custom changes to set user value for kafka-broker/kafka.metrics.reporters checkpoint window does not appear for hosts operations should expose storm log4j configs special characters in Ranger passwords Dispatched From Ambari Are Not RFC 2822 Compliant And May Be Rejected metrics do not match up with data being processed by storm in our streaming pipeline metrics show 0 numbers on the y axis even though the graph has data View: Enable to preview files in WEBHDFS metrics show no data available - when HA enabled Blueprint sets inconsistent Yarn RM values an upgraded cluster: JDK 1.8 option not available in ambari-server setup option for SNMP Alert Notification needs to be removed Knox service check have started before the knox start action config with hdp.version info on large BIGINT in Ambari Hive View returns incorrect value types, incorrect number of selected=1 View : Upload Table Tab : if selected datatabase is default then data does not show in table does not have all the vertices for outputs after CTAS queries to DB properties are required to be filled even if user doesn't want to use DB auditing, AMBARI-13833BUG-41531Blueprints Cluster configuration task thread should not wait indefinitely page: page hangs for a while on HDFS metrics of config versions works incorrectly for DBs kafka userid ulimit open files to 32k for kafka broker through Ambari between pages in installer too long throwing harmless message ObjectGraphWalker:209 - The configured limit of 1,000 object references was reached server Deadlock on View deploy if Xmn size is oversubscribed Ambari Alert Notifications editing group causes ConcurrentModificationException fast in 'name' filter field of the Hosts table misses trailing characters httpclient library dep from AbastractTimelineMetricsSink to respect timeout settings GC tuning params to Collector daemon similar to HBase / Adopt FIFO compaction policy for AMS high load tables Data LossAMBARI-13368BUG-45345'chmod -R 755' should be executed against Nodemanager local dirs Data LossAMBARI-13880BUG-47974DataNode Data Directories Which Contain Meta Information Tags Trigger Alerts Data LossAMBARI-14030BUG-48302Datanode shows up prematurely as 'Decomissioned' in Ambari UI while it is 'Decomissioning in Progress' in Namenode UI installed with HDP 2.2 doesn't handle alerts when http, https ports value are changed 2.1 with HDP 2.2) Ranger setup in Ambari does not respect custom service users to 'Download CSV' file from the Kerberos Wizard on Safari Wizard: next button on Confirm Configuration doesn't work after page refresh security-related HTTP headers to keep Ambari up to date with best-practices START failed after enabling Kerberos security-related HTTP headers to Views to keep Ambari up to date with best-practices Advisor for with Ranger enabled is not specific enough Alert Fails In Kerberized Environment That Is Not SSL start fails when ranger digital password used is not able to run Oozie Server which configured with different username than oozie Wizard: null in request logs Ambari Server is not part of the cluster, basic operations throw 500 (NPE) when Kerberos is enabled (even if it is Manual Kerberos) 'hive.conf.restricted.list' property under hiveserver2-site when Ranger Hive plugin is enabled upgrade Namenode fails to start when Kerberos is enabled using HDP Alert for verifying if Ranger admin password is in sync between Ranger and Ambari HBase prinicpals for Master and RS need to have the same value Membership not pulled in with FreeIPA/OpenLDAP/RHELIDM Manager (HA) fails to start after enabling Kerberos kerberization, Kafka brokers fail to start enabling kerberos with option 'Manage Kerberos principals and keytabs manually', adding new component buttons will not respond doesn't let user specify Active Directory LDAP configs when using HS2 View job submitted as ambari process owner, not as login user. Usersync Moving From Seconds to Millis On Stack Upgrade cannot login after group-membership pulling in freeIPA Files View ignores alternate HDFS authorization mechanisms gateway fails to restart on Ubuntu 12.04 after system restart using custom pid dir because /usr/hdp/current/knox-server/pids does not point to custom pid dir may have same value is original config HA causes issues AMBARI-12628 additional Kafka broker on node with hostname that comes alphabetically before existing broker fails to start check fails after move NameNode on NN HA cluster for some reason, structured-out-status.json is empty then ambari-server will fail to report any status POST-ing an empty Blueprint document results in a NullPointerException SNMP V3 Alerts in Ambari result in java.lang.NullPointerException when trap is fired view shows completed but does not state if anything went wrong. in /etc/server/conf is not updated on webhcat server installation of Storm/Atlas, Ambari overwrites and warns about wrong dependent configuration leak in ambari-agent caused by logger being created on every task failing Yarn service check even when error code is 0[Hive View] UI issue with Hive View in old Firefox browser agent specific metrics not showing up as graphs HDFS API utility needs to work with WASB start fails with bad permissions on state directory. Service wizard: Browser refresh on Assign Master page loses all installed service components node went in critical state on Ambari : message on UI : Connection failed: [Errno 111] Connection refused to balancer via ambari fails to run once HA is enabled balancer via ambari fails to run after Ambari upgrade with NN HA enabled, AMBARI-13861BUG-45961hdfs balancer via ambari fails to run after HDP upgrade with NN HA enabled Checkpoint window does not appear for restart all reconfigured operation on host API query should be made part of request body, AMBARI-13517BUG-46121Ambari Server JVM crashed after several clicks in Web UI to navigate graphs File caching does not work because of ulimit not being passed into the start command for datanode. Metastore Upgrade Retry Fails Because Of Driver Copy Issue "log4j.rootLogger" should have ROLLINGFILE instead of CONSOLE Ambari Repository gives Key Too Long error messages for hostcomponentdesiredstate table. there are more than one ambari-server*jar in classpath we should fail to start HA wizard gets stuck when JournalNodes fail after getting started HDFS HA deployment and namenode move, Accumulo failed to come up. in collector logs for JOIN queries metastore HA enables compactor initiator on all metastores. Timeline Server unexpectedly turn down [rarely reproduced] View query error, should use GetOperationStatus thrift api to verify if hive query is completed successfully Server Start times out due to ranger http calls is always disabled if Start all services are failed when enabling RM HA Pointer exception warning message at org.apache.ambari.server.controller.metrics.timeline.cache.TimelineMetricCacheEntryFactory.updateTimelineMetricsInCache service check failing after reassigning host with custom service user error when adding service to a host with 'HEARTBEAT LOST' status has issues after enabling wire encryption and reconfiguring DATANODEs JDK: need to switch to 8uXX by default when choosing Oracle JDK Option 1 during setup call to delete service fails with error: ORA-02292 integrity constraint violated - child record found and Storm Ambari metrics are not working ability to retry operations after database exception for blueprint deploy code that caches stale entity references quorum hosts for AMS should point to localhost for distributed mode. Agent cannot talk to server if "test -w" command hangs status of RS process for AMS in embedded mode error logs to script alerts when there are failures log file not getting created HDFS tag.HAState JMX Metric Causes JavaScript Errors points to Incorrect Repository version [After upgraded from ambari 1.7.0] Component Install conf-select Shell Uses Current Working Directory That Doesn't Exist Yet or component install fails when a non-ambari apt-get command is running code that breaks transaction chaining a Warning When Jetty Pool Size Is Too Low And Increase It"Application Tracking URL" in Tez View broken due to RM HA changes in Ambari views framework, HDFS-7931BUG-47113Upgrade Ambari Views to Hadoop 2.7.1 jar which has the fix for HDFS-7931 2.7.9 or later fails to establish connection between agent and server. view fails at runtime to instantiate ProtobufRpcEngine class, AMBARI-13325BUG-45594Ambari-server upgrade and RU update hbase bucket cache setting incorrectly not found exception while running step "ambari-server upgradestack HDP-2.3" for upgrade path 2.2 -> 2.3.2 Amabri DB is Oracle Metastore start error after stack upgrade tooltips for status icon in Upgrade Wizard Perform RU If Ambari Server Is Not A Part of the Cluster package command during RU should honor excluded packages throwing ganglia errors after upgrade Fails After Manual Upgrade Because of Cluster Version State Set To UPGRADING Precheck call does not work for both upgrade types Upgrade: Upgrade START call cannot find upgrade pack to use should check for INSTALL_FAILED components before starting 1.7 to 2.1.x upgrade with existing Kerberos, keytab files fail to be distributed to some hosts START failed after upgrade ambari from 1.7.0 to 2.1.2 HA with both NN's being passive path on ORA 1.6.1 -> 2.0.1 -> 2.1.1 - Downgrade from 2.3 to 2.2 may fail if configs were not back'ed up any snapshots/backups to be skippable when doing RU. Progress Popup: even though there is a failure during service check it has a green tick mark Metrics Collector failed to start in Ambari [upgraded from ambari 1.7.0] Tez is failed(after upgrade from 1.6.1 to PreCheck for consistency of password between Ranger and Ambari ams-site resultset limit property to upgrade stack_upgrade 2.2 to 2.3 hook script fails due to missing /configurations/tez-site/ value when enabling security during Update Configurations stage (after upgrade from 1.6.1 to Upgrade: Configuration Groups not applied for upgrades across stack versions Upgrade: Kafka to call migration script when upgrading across HDP upgrade request is un-affected if there exists a stage with no tasks"Falcon - Oozie Integration" property are populated and marked Mandatory After rolling Upgrade Intra-Stack Upgrade Web Client Still Uses Old Stack Version In Requests issue with user button filter works incorrectly on Hosts page Versions list isn't updated while user is on the configs page after adding/removing HiveServer2 size 1024 clips tooltips does not redirect the user to the tez ui if log in is required mistake in the help message for max capacity in capacity scheduler Tez View page is displayed if ResourceManagers are down disk usage metric confusing Hive View is not default with hive service stack definition for Spark has incorrect value for SPARK_CONF_DIR View Data Visualization - Change of Tab Order to compare config Versions when chosen from 'Show more' to read keysdir cannot handle space in the front dir for metrics server does not get updated after NN HA is enabled editing alert notifications, changes to severity are not reflected HDFS files (View) appends several "non-printable" characters at the end of the file. statement queries with errors after the first statement cause orphaned AMs does not have the ability to set heapsize for hive components about kafka ranger plugin state does not appeared after adding kafka Values/units for HiveServer2 Heap Size overlapping on slider does not configure hbase.coprocessor.regionserver.classes CS View "Ambari Cluster URL" setting regex is too restrictive View : Upload Table tab : Distorted alignment of form fields View : In Database Explorer the scroll bar come over the show table button Should Return Correct System Level Run Code