This site contains release notes for MapR Version 5.0 and below.  You can also refer to the release notes for the latest release.

Skip to end of metadata
Go to start of metadata

General Information

The page Hadoop Compatibility in Version 2.0 could not be found.

We have released Hbase 94.1 and Oozie 3.2.0 separately, in the following location: http://package.mapr.com/releases/ecosystem-all

New in This Release

This is a maintenance release with no new features.

Resolved Issues

The following issues have been resolved in this release.

MapReduce

(Issue 6132)

JobTracker HA detects JT process being terminated and fails over to a standby JT. However if JT process is there, but is not responsive, no failover occurs. With the fix, if JT process is unresponsive for a configurable duration (default 10 min), its jstack is logged and it is restarted automatically.

(Issue 7628)

TaskTracker won't run if node hostname resolves to IP not covered by MAPR_SUBNETS.

(Issue 8105)

Potential deadlock while cleaning up Job Metrics data.

(Issue 8115)

Split calculation does not handle Mapr-FS chunk size 0 and client JVM spins at 100% CPU utilization while submitting Map/Reduce job.

Filesystem

(Issue 6717)

Potential memleak in flushing dirty inodes in file system.

(Issue 7894)

Mirroring no longer causes an assert failure by losing track of the most recent replica of a blank source container.

(Issue 7925)

A lot of containers may get created on a volume when a node goes down.

(Issue 7931)

IOMgr trying to access the cache pages after failing the IO can cause an assert in the filesystem.

(Issue 7984)

Name container for large volume stuck under replicated for 2 days due to message timeout likely because of network issues.

(Issue 8071)

CLDB shows containers has master but containers are stale due to larger than 64K message not being sent.

(Issue 8092)

fsck doesn't work on storage pools > 6TB.

(Issue 8139)

No CLDB master for ~20 minutes after upgrading.

(Issue 8174)

Excessive number of containers in mapred local volume.

(Issue 8179)

The "disk online" operation may crash with EIO error.

(Issue 8189)

The parameter cldb.balancer.disk.threshold.percentage not used.

(Issue 8262)

When the topology of a local volume changes, containers for the local volume are not re-used, resulting in a large number of container creation operations.

Logging

(Issue 7677)

Can't set debug logging for maprcli (or any option from command line).

(Issue 7793)

Jobtracker log shows garbage characters in "NetworkTopology: Adding a new node:" entry.

(Issue 8144)

Fileclient should have option to add the IP address of the node to log message along with the container-id.

MCS & CLI

(Issue 6507)

Spurious volume alarms may get raised.

(Issue 6552)

Support for anonymous authentication with SMTP.

(Issue 7892)

Nodes that register early as a CLDB node becomes the master CLDB node sometimes incorrectly appear twice in the balancer classification.

(Issue 7989)

Add start time to maprcli dump rereplicationinfo -json output.

(Issue 8047)

Maintenance mode may not work correctly - node under maintenance is reported with status Critical.

(Issue 8134)

2.0 MCS keeps giving the error "Cannot use both alarmednodes and nfsnodes together" after switch from nfsnodes report to alarmed node report.

Services

(Issue 8032)

Warden and zookeeper fail to start due to requiretty being enabled by default in 1.7.4p5 version of sudo.

(Issue 8280)

When MapR is run as non-root user, stop warden may not stop JT/TT due to incorrect permissions on the pid file.

NFS

(Issue 8260)

NFS client hung accessing files with chunksize set to 0.

(Issue 8136)

Buffer overflow when there are more entries to export than 1024 buffer.

Security

(Issue 8137)

mount fails due to mapr user not having root privileges.

(Issue 10598)

When a node's IP number is reassigned to another node, remote procedure calls are misdirected to the wrong node. This can result in widespread errors affecting any system that use the RPC binding layer, including container resync operations that never finish or the creation of several superfluous containers. Restart the cluster to clear this condition.

Known Issues

(Issue 11389)

If you use the latest version of any browser to view the MCS, the CLDB, HBase, JobTracker, and job log pages do not load when you click on the respective links. As a workaround, use the following instructions to disable mixed-content-blocking in your browser:
Internet Explorer
Scroll to the bottom of the screen, and click Show all content. The page refreshes and displays mixed content. 

Mozilla Firefox

  1. At the top of the page, to the left of the address bar, click  (the shield icon).
  2. In the pop-up window that appears, click the down arrow next to "Keep Blocking", and select Disable Protection on This Page. The page refreshes and displays mixed content. The shield icon is replaced with a warning symbol to indicate that the page is displaying both secure and non-secure items.

Google Chrome

  1. Click  (the shield icon) on the right side of the address bar.
  2. In the icon dialog box, click Load anyway. The page refreshes and displays mixed content. The URL in the address bar shows https crossed out to indicate that the page is displaying both secure and non-secure items.
  • No labels