I'm trying to choose a topology for apache spark cluster across different sites. Does spark has it's own awareness capabilities?
For example assume a cluster with workers in Oregon and Penang.
Now when submitting an application, which loads data from Oregon processes it and saves it back to Oregon. Will Oregon workers will be preferred (if they are free)?
Haven't found documentation regarding this subject.
As described here https://jaceklaskowski.gitbooks.io/mastering-apache-spark/content/spark-data-locality.html
Spark relies on data locality, aka data placement or proximity to data
source, that makes Spark jobs sensitive to where the data is located.
It is therefore important to have Spark running on Hadoop YARN cluster
if the data comes from HDFS.
The data system may itself be geo-aware e.g. cassandra: Does Spark use data locality? http://www.slideshare.net/RussellSpitzer/spark-cassandralocality
Related
I have a DSE cluster where every node in the cluster has both spark and Cassandra running.
When I load data from Cassandra to spark Rdd and do some action on the rdd, i know the data would be distributed into multi nodes. In my case, I want to write these rdds from every node to its local Cassandra dB table directly, is there anyway to do it.
If I do normal rdd collect, all data from spark nodes would be merged and go back to node with driver.
I do not want this to happen as the data flow from nodes back to driver node may take Long time, I want the data been save to local node directly to avoid the data movement across the spark nodes.
When Spark executor is reading data from Cassandra it's sending request to the "best node" that is selected based on the different factors:
When Spark is collocated with Cassandra, then Spark is trying to pull data from the same node
When Spark is on different node, then it's using token-aware routing, and read data from multiple nodes in parallel, as it's defined by the partition ranges.
When it's comes to the writing, and you have multiple executors, then each executor is opening multiple connections to each node, and writing the data using the token-aware routing, meaning that data is sent directly to one of the replicas. Also, Spark is trying to batch multiple rows that are belonging to the same partition into an UNLOGGED BATCH as it's more performant. Even if the Spark partition is colocated with the Cassandra partition, writing could involve an additional network overhead as SCC is writing using the consistency level TWO.
You can get colocated data if you re-partitioned the data to match Cassandra's partitioning), but such re-partition may induce Spark shuffle that could be much more heavyweight compared to the writing data from executor to another node.
P.S. You can find a lot of additional information about Spark Cassandra Connector in the Russell Spitzer's blog.
A word of warning: i only use Cassandra and Spark as separate open source projects, i do not have expertise with DSE.
I am afraid the data need to hit the network to replicate, even when every spark node talks to its local cassandra node.
Without replication and running a Spark job to make sure all data is hashed and preshuffled to the corresponding Cassandra node, it should be possible to use 127.0.0.1:9042 and avoid the network.
I have to design a setup to read incoming data from twitter (streaming). I have decided to use Apache Kafka with Spark streaming for real time processing. It is required to show analytics in a dashboard.
Now, being a newbie is this domain, My assumed data rate will be 10 Mb/sec maximum. I have decided to use 1 machine for Kafka of 12 cores and 16 GB memory. *Zookeeper will also be on same machine. Now, I am confused about Spark, it will have to perform streaming job analysis only. Later, analyzed data output is pushed to DB and dashboard.
Confused list:
Should I run Spark on Hadoop cluster or local file system ?
Is standalone mode of Spark can fulfill my requirements ?
Is my approach is appropriate or what should be best in this case ?
Try answer:
Should I run Spark on Hadoop cluster or local file system ?
recommend use hdfs,it can can save more data, ensure High availability.
Is standalone mode of Spark can fulfill my requirements ?
Standalone mode is the easiest to set up and will provide almost all the same features as the other cluster managers if you are only running Spark.
YARN allows you to dynamically share and centrally configure the same pool of cluster resources between all frameworks that run on YARN.
YARN doesn’t need to run a separate ZooKeeper Failover Controller.
YARN will likely be preinstalled in many Hadoop distributions.such as CDH HADOOP.
so recommend use
YARN doesn’t need to run a separate ZooKeeper Failover Controller.
so recommend yarn
Useful links:
spark yarn doc
spark standalone doc
other wonderful answer
Is my approach is appropriate or what should be best in this case ?
If you data not more than 10 million ,I think can use use local cluster to do it.
local mode avoid many nodes shuffle. shuffles between processes are faster than shuffles between nodes.
else recommend use greater than or equal 3 nodes,That is real Hadoop cluster.
As a spark elementary players,this is my understand. I hope ace corrects me.
As Apache Spark is a suggested distributed processing engine for Cassandra, I know that there is a possibility to run Spark executors along with Cassandra nodes.
My question is if the driver and Spark connector are smart enough to understand partitioning and shard allocation so data are processed in a hyper-converged manner.
Simply, does the executors read data stored from partitions that are hosted on nodes where an executor is running so no unnecessary data are transferred across the network as Spark does when it's run over HDFS?
Yes, Spark Cassandra Connector is able to do this. From the source code:
The getPreferredLocations method tells Spark the preferred nodes to fetch a partition from, so that the data for the partition are at the same node the task was sent to. If Cassandra nodes are collocated with Spark nodes, the queries are always sent to the Cassandra process running on the same node as the Spark Executor process, hence data are not transferred between nodes. If a Cassandra node fails or gets overloaded during read, the queries are retried to a different node.
Theoretically yes. Same for HDFS too. Howevet practically I have seen less of it on the cloud where separate nodes are used for spark and Cassandra when their cloud services are used. If you use IAsAS and setup your own Cassandra and Spark then you can achieve it.
I would like to add to Alex's answer:
Yes, Spark Cassandra Connector is able to do this. From the source
code:
The getPreferredLocations method tells Spark the preferred nodes to
fetch a partition from, so that the data for the partition are at the
same node the task was sent to. If Cassandra nodes are collocated with
Spark nodes, the queries are always sent to the Cassandra process
running on the same node as the Spark Executor process, hence data are
not transferred between nodes. If a Cassandra node fails or gets
overloaded during read, the queries are retried to a different node.
That this is a bad behavior.
In Cassandra when you ask to get the data of a particular partition, only one node is accessed. Spark can actually access 3 nodes thanks to the replication. So without shuffeling you have 3 nodes participating in the job.
In Hadoop however, when you ask to get the data of a particular partition, usually all nodes in the cluster are accessed and then Spark uses all nodes in the cluster as executors.
So in case you have a 100 nodes: In Cassandra, Spark will take advantage of 3 nodes. In Hadoop, Spark will take advantage of a 100 nodes.
Cassandra is optimized for real-time operational systems, and therefore not optimized for analytics like data lakes.
While exploring various tools like [Nifi, Gobblin etc.], I have observed that Databricks is now promoting for using Spark for data ingestion/on-boarding.
We have a spark[scala] based application running on YARN. So far we are working on a hadoop and spark cluster where we manually place required data files in HDFS first and then run our spark jobs later.
Now when we are planning to make our application available for the client we are expecting any type and number of files [mainly csv, jason, xml etc.] from any data source [ftp, sftp, any relational and nosql database] of huge size [ranging from GB to PB].
Keeping this in mind we are looking for options which could be used for data on-boarding and data sanity before pushing data into HDFS.
Options which we are looking for based on priority:
1) Spark for data ingestion and sanity: As our application is written and is running on spark cluster, we are planning to use the same for data ingestion and sanity task as well.
We are bit worried about Spark's support for many datasources/file types/etc. Also, we are not sure if we try to copy data from let's say any FTP/SFTP then will all workers will write data on HDFS in parallel? Is there any limitation while using it? Is there any Audit trail maintained by Spark while this data copy?
2) Nifi in clustered mode: How good Nifi would be for this purpose? Can it be used for any datasource and for any size of file? Will be maintain the Audit trail? Would Nifi we able to handle such large files? How large cluster would be required in case we try to copy GB - PB of data and perform certain sanity on top of that data before pushing it to HDFS?
3) Gobblin in clustered mode: Would like to hear similar answers as that for Nifi?
4) If at all there is any other good option available for this purpose with lesser infra/cost involved and better performance?
Any guidance/pointers/comparisions for above mentioned tools and technologies would be appreciated.
Best Regards,
Bhupesh
After doing certain R&D and considering the fact that using NIFI or goblin will demand for more infrastructure cost. I have started testing Spark for data on-boarding.
SO far I have tried using Spark job for importing data [present at a remote staging area/node] into my HDFS and I am able to do that by mounting that remote location with all my spark cluster worker nodes. Doing this made that location local to those workers, hence spark job ran properly and data is on-boarded to my HDFS.
Since my whole project is going to be on Spark, hence keeping data on-boarding part on spark would not cost anything extra to me. So far I am going good. Hence I would suggest to others as well, if you already have spark cluster and hadoop cluster up and running then instead of adding extra cost [where cost could be a major constraint] go for spark job for data on-boarding.
I am looking for directions:
I have a cassandra database with latitude & longitude data. I need to search for data within a radius or a box coordinates around a point. I am using golang(gocql) client to query Cassandra.
I need some understanding regarding Spark and Cassandra as this seams like the way to go.
Is the following assumptions correct; I have 2 Cassandra nodes(the data in a replica of 2).
Should I then install an extra node and install Spark on this and then connect it to the other two existing Cassandra nodes containing the data(With the Spark Connector from DataStax).
And do the two existing Cassandra nodes need to have Spark workers installed on them to work with Spark Master node?
When the Spark setup is in place, do you query(Scala) the existing data and then save the data onto the Spark node and then query this with the gaoling(gocql) client?
Any directions is welcome
Thanks in advance
Geospatial Searching is a pretty deep topic. If it's just doing searches that you're after (not batch/analytics), I can tell you that you probably don't want to use Spark. Spark isn't very good at 'searching' for data - even when it's geospatial. The main reason is that Spark doesn't index data for efficient searches and you'd have to create a job/context (unless using job server) every time you'd want to do a search. That takes forever when you're thinking in terms of user facing application time.
Solr, Elastic Search, and DataStax Enterprise Search (Disclaimer I work for DataStax) are all capable of box and radius searches on Cassandra data and do so in near real time.
To answer your original question though, if the bulk of your analytics in general come from Cassandra data, it may be good idea to run Spark on the same nodes as Cassandra for data locality. The nice thing is that Spark scales quite nicely, so if you find Spark taking too many resources from Cassandra, you can simply scale out (both Cassandra and Spark).
Should I then install an extra node and install Spark on this and then
connect it to the other two existing Cassandra nodes containing the
data(With the Spark Connector from DataStax).
Spark is a cluster compute engine so it needs a cluster of nodes to work well. You'll need to install it on all nodes if you want it to be as efficient as possible.
And do the two existing Cassandra nodes need to have Spark workers
installed on them to work with Spark Master node?
I don't think they 'have' to have them, but it's a good idea for locality. There's a really good video on academy.datastax.com that shows how the spark cassandra connector reads data from Cassandra to Spark. I think it will clear a lot of things up for you: https://academy.datastax.com/demos/how-spark-cassandra-connector-reads-data
When the Spark setup is in place, do you query(Scala) the existing
data and then save the data onto the Spark node and then query this
with the gaoling(gocql) client?
The Spark-Cassandra connector can communicate to both Cassandra and Spark. There are methods, saveToCassandra(), for example, that will write data back to Cassandra your jobs are processed. Then you can use your client as you normally would.
There are some really good free Spark + Cassandra tutorials at academy.datastax.com. This is also a good place to start: http://rustyrazorblade.com/2015/01/introduction-to-spark-cassandra/