- Cloudera Manager for administering the Hadoop cluster
- Cloudera specific solutions like Impala
- Extensive documentation
- Good user community
Implementing a Hadoop cluster has become relatively straight-forward using CDH. Administering it is also less complex. As a result, efforts spent in these areas are less than anticipated.
We have been using it for the last two years.
Following a single path for installation becomes confusing due to multiple recommended approaches e.g. parcels vs packages.
Flume seems unstable and has to be restarted quite often.
None as such
We are mostly using Cloudera Express so we did not use their technical support. However, the Cloudera community is an active place and Cloudera representatives participate actively in understanding and resolving issues.
Cloudera is a prominent player in the Hadoop space and we did not have a need to adopt a different solution. However, we are also looking to work on Hadoop and MapR
Following a single path for installation was initially confusing due to multiple recommended approaches e.g. parcels vs. packages. However, after a while, we managed to master it. However, knoweldge of Cloudera Manager and Hadoop architecture is a must.
We have our own team of consultants who are proficient in implementing it. The high level steps about the implementation remain the same; however, it is the environment specific issues which are challenging.
We haven't really measured ROI.
Licensing price on per node basis for Cloudera seems to be pretty steep (based on the inputs we have received from Cloudera).
It is user friendly and installation is pretty straightforward. Cloudera Manager is a good tool to administer it. However, configuration for specific requirements is sometimes pretty complex.
You should have a team which is knowledgeable in Hadoop. Do keep in mind that the product is still maturing so there are good chances that you will come across unexpected issues now and then.