Ryan Williams created SPARK-9195:
------------------------------------
Summary: RDD/Storage metrics don't update cached partition counts
when executors are removed/lost
Key: SPARK-9195
URL: https://issues.apache.org/jira/browse/SPARK-9195
Project: Spark
Issue Type: Bug
Components: Web UI
Affects Versions: 1.4.1
Reporter: Ryan Williams
Priority: Minor
I have an RDD that was cached, and then i lost most of my executors (in this
case it was due to the executors having been dynamically-allocated and timing
out, but this applies when the executors are lost in general).
The RDD page correctly shows that most of the RDD is no longer cached:
!http://f.cl.ly/items/1i2C1v1H2o0i243f453P/Screen%20Shot%202015-07-20%20at%2011.46.11%20AM.png!
(However, the "Cached Partitions" count at the top is incorrect).
Additionally, the storage page still shows it as 100% cached:
!http://f.cl.ly/items/3P3l440u0R3L3R0U1a2E/Screen%20Shot%202015-07-20%20at%2011.45.57%20AM.png!
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
