Hi All,
As Ambari back now, let’s start discussing roadmap. I will be closely tracking
for Jira, GitHub(with Jenkins) and wiki to work normal.
Based on the conclusion of this thread will update wiki[1].
As 2.7.6 is released recently, we need to release 2.7.7
1. Stack Management. There are two approaches for this. Open to discuss on
following..
* Entirely new stack
i. Define
new stack name. (BDP: BigDataPlatform, BDS:BigDataStack,
OSS:OpenSourceStack,TDP: TuskerDataPlatform…)
ii. Define
stack versions (zk-3.5.9,Hadoop-3.2.3,hive-3.1.3,tez-0.10.1,spark-3.2..)
iii. Based
above two we need to have select in bigtop code.
* Using the existing for current release.
i. Plugin
the new stack into HDP self
ii. Still, we
can have hdp select
I prefer the planA but planB takes less time. Users can easily adopt it upgrade
their stack.
Note: Ideally bigtop mpack can be long term approach as stack id decoupled now,
but currently it’s not ready for latest stack (upgrades can be risk). We’ll
work in trunk for same. If this works shortly, we can choose.
1. Upgrade :
* Most of users will be deployed with HDP versions(HDP-2.6.5,HDP-3.1.0),
we might need to support the upgrade to 2.7.7
2. Other work
* Any critical/blocker CVE’s Fixes
* CVEs Fixes : we’ll try to target some CVE fixes
* Review the open PR’s and Jira’s
Trunk Code : Parallel we can work for the trunk code, this we might give as 2.8.
* Removal of python2 support
* Removal of hard dependency on HDP
* Working on bigtop mpack
Can add more based on your inputs…
Any thoughts on this..? Please feel free to add anything which I missed.
Reference:
1.https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=30755705
