Spark charm issues

[Cory Johns]

You may have run into some issues deploying the Apache Spark charm (
https://jujucharms.com/apache-spark/) or related bundles (e.g.,
https://jujucharms.com/realtime-syslog-analytics/) yesterday and today.  It
was due to the binaries for Spark 1.3.0 being archived from dist.apache.org.

We fixed the issue and improved the deployment speed of the Spark charm by
mirroring the Spark binaries on the S3 bucket that we use for the other big
data charm binaries.  This will future-proof the charms and gives much
better download speeds for the binaries.

We are also looking in to moving up to a more recent version of Spark
(currently at 1.5.1) but ran into compatibility issues with Apache Zeppelin
that we haven't had a chance yet to work out.

Let us know if anyone runs into any more issues with those (or any of the
big data charms)!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/bigdata/attachments/20151021/f5534bbf/attachment.html>