Category: Hadoop

Garima Dosi discusses an architecture using ZooKeeper to introduce some limited protections for concurrent access in HDFS:

The ZooKeeper nodes topology as per the design looks like this. ZooKeeper works like a filesystem starting with a root directory followed with several nodes (analogous to folders) and finally the data nodes (analogous to files). The circles in the image represent the name of a property/folder that we are trying to maintain and the rounded boxes are the values/files for those properties/folders.

So, the image above shows that the “global version” is 100 and there are 10 & 20 read requests being executed on versions 98 and 99 respectively and since there is a write request in progress, no other write request would be taken up until it completes.

This feels a little overly complicated to me.

Jon Morisi has notes on downloading and preparing the Hortonworks sandbox VM via VirtualBox:

Lessons learned:

Use wget for downloads that repeatedly drop the connection.
Don’t try to run VirtualBox inside a VM.
Enable VTx in the BIOS to run VirtualBox.

This post is mostly about troubles in getting the VM and preparing the software.  Definitely take his advice on enabling VTx in your BIOS.

Ben Davis discusses considerations when rolling your own Hadoop cluster:

5. Security hardening
I find it is easier to deploy Hadoop in a fairly low security configuration. This is because there are a range of ports that Hadoop talks on and having an incorrectly configured firewall can cause you problems. So after deployment, set aside time to identify how to customise your firewalls, user and group settings, Kerberos and ssl settings.

I think the article makes some good points.  DIY is great for a proof of concept or for playing around with a technology, but if you don’t already have a good amount of experience with a technology, you’ll probably make costly mistakes in development and administration.  This is not Hadoop-specific:  I’ve seen companies do terrible things to SQL Server because they didn’t know the correct way to do it but needed to get work done.  As part of a proof of concept, do all the terrible things you’d like; they’re how you’ll learn.  But if this is going to production, it’s a good idea to have people who know what they’re doing involved.

Jon Morisi is looking at Hadoop:

From here I think I’ll start playing around on a sandbox.  Each of the distributions offers a way to spin up a VM or log into a cloud based environment.  There are also docker images out there. (search hadoop, cloudera, or hortonworks).  Most of these docker images look fairly new, so don’t cut yourself.

I’m looking at the HortonWorks distro, so I’ll probably setup a Hortonworks Sandbox.

Jon includes some resources he’s used to learn a bit about the topic.  I think he’s going down the right path with videos and mailing lists—the ecosystem changes too quickly for books to have much long-term value, and mailing lists & forums tend to be better for keeping up to date.  My biggest suggestion is to get case studies and play around.  Check out studies on e-mail ingest, real-time data analysis, and analyzing fantasy sports for starters.

Skumar T explains container sizes in Yarn:

So jobs on yarn cluster runs in individual containers which is allocated by Node Manager which in turn gets permissions from Resource Manager.

So few configuration parameters of node manager those are important in context of jobs running in the containers.

–>yarn.nodemanager.resource.memory-mb  8192(value)

Amount of physical memory, in MB, that can be allocated for containers.

–>yarn.nodemanager.pmem-check-enabled  true(value)

Whether physical memory limits will be enforced for containers.

The bottom half of the article goes into an extended example.

Jordan Volz and Stefan Salandy show how to feed e-mails into Hadoop for almost-immediate analysis:

In particular, compliance-related use cases centered on electronic forms of communication, such as archiving, supervision, and e-discovery, are extremely important in financial services and related industries where being “out of compliance” can result in hefty fines. For example, financial institutions are under regulatory pressure to archive all forms of e-communication (email, IM, social media, proprietary communication tools, and so on) for a set period of time. Once data has grown past its retention period, it can then be permanently removed; in the meantime, such data is subject to e-discovery requests and legal holds. Even outside of compliance use cases, most large organizations that are subject to litigation have some form of archive in place for purposes of e-discovery.

Traditional solutions in this area comprise various moving parts and can be quite costly and complex to implement, maintain, and upgrade. By using the Hadoop stack to take advantage of cost-efficient distributed computing, companies can expect significant cost savings and performance benefits.

In this post, as a simple example of this use case, I’ll describe how to set up an open source, real-time ingestion pipeline from the leading source of electronic communication, Microsoft Exchange.

Most of this post is about setting up the interconnections between Exchange and Apache James, and feeding data in.  It looks like this will be part 1 of a multi-part series.