Hi im trying to learn every side of mapr, Is it necessary to build a cluster in order to learn mapr?
Please help me to have a step by step learning procedure.
In addition to what Mufeed Usman says, it is absolutely important to build out a cluster.
The sandbox allows you to build a very small vm as a single node cluster. The downside is that it limits what you can do due to its size (disk, memory and horsepower). However its upside is that it allows you to see how things work and to test things out in a limited space and with limited funds. You can go a step further and build out a single node cluster on a small format PC like the Intel NUC platforms. This gives you a little bit more insight on how to manage an actual cluster, albeit a single node. It also provides more system resources over a VM and the cost isn't that great. (If you have an older machine that you can run Centos... that too will work.)
You can also go to AWS or another cloud provider and you can then build a small cluster using EC2 machines. Cheaper than building your own dedicated hardware and it will teach you more about the maintenance and monitoring of distributed processing.
There are really two main learning paths. Standing up and maintaining a cluster, or Data Engineering / Data Science on a cluster.
While you can watch the video lessons, its boring and it doesn't really give you a feel for what to do. You need to go hands on and play.
I would also recommend that you also join a local user group or meetup. It will allow you to find others to bounce ideas off of and to help answer questions and learn new things.
A single node cluster (like a sandbox) will be an ideal environment to learn the basics. You could get started by referring MapR Sandbox for Hadoop. Please feel free to post your queries here so that the community can assist you in your journey.
Hi ! Does the single node installation is quite similar with the installation of the whole cluster. Is there any instruction on installing mapr on a single nide?
The primary difference will only be in the selection of packages according to your service distribution needs across the nodes in the cluster. If only a single node, all the requisite packages will be installed on the same node.
Hope that helps.
There are actually free classes at MapR Academy that include hands-on lab exercises that will get you started (they're not actually boring, and they do give you a feel for what to do). They will walk you through installing a cluster, as well as configuring and using most of the key features. They're a great starting point. Check out ADM 200 through ADM 203 on MapR Academy (learn.mapr.com).
Yes I agree. My point was watching videos is boring. :-)
Having a local cluster so you can do "what if ... " type of tests.
While many build VMs on their laptops or PCs, I find that they can be a little too limiting. So going out and buying a dedicated PC makes sense. If you want, you could build a 'Franken Cluster' out of old hardware. (Note: Steve Hoffman coined that term when Orbitz built their first cluster)
Intel NUCs aren't too terribly expensive and have more horsepower than the VMs. Or you could set up an AWS account and play there.
I appreciate every answers in this community. Thank you very much.
I feel so pressured in my job right now because I need to learn it fast (so now I have requested a 3 node cluster which I am playing).
Mark Dendrix Garcia Do feel free to tag me (as @mufeed_usman) in your queries in the future so that I get notified immediately via our internal mail. Let me highlight the fact that I work more closely on MapR core components. So, if I am unable to provide a satisfactory answer on other non-related areas, at the best I can try and reel-in folk(s) in to the discussion to try and assist you.
Of course, you'll also have our Community Manager keeping an eye out as well, in case we lose sight of some of the queries sent our way.
So, rest assured; no pressures. We're here to help . And those are not empty words.
Both Michael and Mufeed provided great suggestions for you to build a MapR cluster. Do you have more questions before I close this discussion?
Nah, Ill open a new thread after I successfully install it thanks
Look forward to that .
Retrieving data ...