Cloudera Distribution for Hadoop Reviews

We are dealing with data from the telecom industry. We were using an Oracle system but our volume has increased. We now have a lot of real-time data that needs to be transformed so that it can be made available and used.

The most valuable feature is Impala, the querying engine, which is very fast. We have been able to work with one terabyte of data in less than 20 minutes. The speed makes it easy for us to process all of the data that comes in, in time.

The support is very good.

All of the data has automatic triple replication in order to secure integrity.

There is a maximum of a one-gigabyte block size, which is an area of storage that can be improved upon.

When we are upgrading CDH, there are many things that need to be upgraded and it would be helpful if it were bundled. As it is now, we have to upgrade many different things separately.

I have been working with the Cloudera Distribution for Hadoop for around two years.

It is a stable solution.

The scalability is good and it works on commodity hardware. One of the problems we have right now is that there is a lot of data and we're moving it from our Oracle solution. This means that there is a double cost, in terms of storage, during our transition to working with big data.

We are using a data lake that is a store for all of the data in our organization. There are more than25 projects, with between 25 and 30 people in each one, for a total of almost 1,000 people. All of them are dependent on this solution.

Most of our users are technicians who have problems to solve using the data available to them. A couple of them are data scientists and the remainder are upper management, who do the analysis.

The technical support is very good. Whenever we open a ticket, we get support right away.

We did use another solution prior to this one but it could not keep up with our increase in data.

This suitability of this solution depends on the size of the data that you are going to be working with. If you have going to be working with a huge dataset that contains many gigabytes of data then this is a good solution. For smaller datasets, you should also consider other technologies.

My advice for anybody who is implementing this solution is to take some time to learn it. Beyond that, be sure to contact support if you have any problems because they are very helpful.

I would rate this solution a seven out of ten.

I primarily use CDH for data storage and regular dashboard reports.

The most valuable feature is that I can use CDH for almost all use cases across all industries, including the financial sector, public sector, private retailers, and so on.

Cloudera's prices are too high and are not competitive with other solutions. They could also improve the Data Science Workbench and add some more features, like wizard activities.

CDH is stable.

CDH is scalable, but it's expensive to do it.

Cloudera's support is extremely bad and cannot be relied on.

I wouldn't recommend CDH to others because of its high cost.

I would rate CDH as eight out of ten.

The Cloudera Manager administrator webpage simplifies the administration tasks and helps to maintain a global overview of the cluster performance.

We are moving from an standard SQL environment (Oracle DataWarehouse) to a Big Data environment, and the Hadoop cluster will be the key of our new organization. It will allow to scale in an easy namer.

We found some difficulties when importing Hive tables from another Cluster.

I want to point the fact that we encounter many problems related to the cloud storage and how resources are managed. Our learning has been that, although it is quite simple to deploy single machines on the cloud, deploying clusters of machines is much more complex as many factors need to be considered: individual machines, connectivity across machines, storage.

I've used it for three months.

We found some issues but were related with the hardware provider. For the moment I have not detected any problem from the Cloudera software point of view.

Technical support is really efficient.

We chose this product as it is considered a market standard and due to its wide documentation on the web. I evaluated other options but the fact that now it is becoming an standard for many companies helped me to choose this option.

In the cloud environment where we deployed (Azure Resource Manager) there was a ready-to-deploy template which simplified a lot the initial set-up.

We implemented with an in-house team. Our initial idea was to stop the cluster during the weekends and when there was no usage. However, we found strong difficulties and we were not able to start programmatically the whole cluster, so finally we left the cluster working all the time.

This issues were mainly related with the cloud provider and how this provider manages the resources for the cluster machines.

From our point of view it is a long-time investment. We hope to get the ROI in the following years.

I am very comfortable with this product. The combination of Cloudera Manager administrator server, which allows the management of the Hadoop Cluster, and the Hue server, which simplifies the use make this product a current standard on the market. Perhaps it lacks a full integration of all its components.

• Cluster rolling restarts 
• Cluster wide configuration management

For the clusters using CM, we are able to more tightly control and manage the configuration of all nodes in the clusters. 

We are currently running six production clusters totaling 900+ nodes, and are building three more clusters. Knowing that if someone has some custom configuration on a node that they haven’t communicated out, and that I can ignore that configuration and bring that node into line with where we’ve decided to run the cluster, is very beneficial.

HBase 1.0 stability issues and processing speed is a major area for improvement. Right now, our Cloudera 5 clusters run four to seven times slower than our Cloudera 4 clusters using our storm and kafka topologies, which causes real-time processing to be a major challenge.

CM’s API is very limited and difficult when used on multiple clusters in the same CM instance

We've used it for approximately two years. We also use Cloudera Manager, which is 6/10.

No issues encountered.

Cloudera 5 is currently very unstable. Between two Cloudera 5 clusters, we have an incident at least twice a week due to what are now outstanding bugs.

It's very easy to deploy and scale as large as you want. Once created on the CM management cluster, is difficult to scale up as needed, as you add more clusters to the same CM instance.

No previous solution was used.

We were already running one production cluster with approximately 75 nodes when I joined, so I’m not familiar with what was needed to get the initial production cluster up. Once I joined, I assisted in standing up the additional nodes and clusters using our chef automation.

In house via chef automation. Chef, or similar systems, makes it much simpler to stand up large scale clusters. That said, I have not used or evaluated vendor team implementation methods.

Faster runtime for batch jobs.

Improved Business Intelligence reporting from daily to every two hours satisfying the business stakeholders who would favour transactional systems to draft reports because it had the latest data. 

The issue that arises using transactional systems with multiple version of truths across the enterprise. With faster turn-around time business stakeholders are now adopting the BI systems designed to give a cohesive view of the performance metrics important to them.

Full Support for all Spark SQL features, support for SparkR, compatibility with Hive for DataFrame saved tables.

Cloudera CDH5.5.x does not support SparkR. SparkR, the integration of R models in API would be a great addition since this will enable fast near real-time analytical integration of R models with data feed.

The functionality in SparkSQL to save a DataFrame as a table in HIVE produces a table not compatible with HIVE. There is a workaround for this in creating the HIVE table first and then doing inserts.

Cloudera CDH5.5.x is a great product, but the adoption of additional features not currently supported will make the product even better but by no means subtract from its desirability.

Do thorough research and ensure your use-cases or scale does not conflict with the system requirements and that those features that would make a difference are supported.

• 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.

• Some of the UI features seem confusing e.g. charts on the CM Services page
• Sometimes it gets confusing to follow a single path for installation due to multiple recommended approaches e.g. parcels vs packages

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.

Our core product is an insurance product and the actuarial module is quite complex. SMEs so far collect data from various sources into Excel sheets and through macros do the analytics which is a very crude form of doing the analysis. So we thought to use big data for such analysis.

That is still in PUC stage, as I have mentioned our analyst used to do the actuarial on a spreadsheet but after Hadoop  implementation they are getting confidence that now analysis is more appropriate and fast. Now exploring cloud implementation as well.

Keeping multi copies of the file and tools of map reduce like PIG, HIVE due to their flexibility it is easy to develop the application with less or almost no knowledge of Java and Sql. And capability to handle huge data size.

As such in the product side, I don't have much to comment. But like other upcoming technologies like RPA, AI, GO etc they have ample training materials with variety of USE Cases, which users can understand and aligned with their current requirements. On same ground I didn't see much training materials from Cloudera.

Seems quite stable, as such didn't face any issue.

It is very stable, didn't face any performance issue.

No when we were heard of Hadoop, we tried on that only. I mean tried to migrate from spreadsheets to Hadoop.

Very straight forward. Typical Windows type installation...Next, next, next clicks.

In-house.

Other department handles all these so I can't comment on that.

Not really.