I am writing Spark application (Single client) and dealing with lots of small files upon whom I want to run an algorithm. The same algorithm for each one of them. But the files cannot be loaded into the same RDD for the algorithm to work, Because it should sort data within one file boundary.
Today I work on a file at a time, As a result I have poor resource utilization (Small amount of data each action, lots of overhead)
Is there any way to perform the same action/transformation on multiple RDD's simultaneously (And only using one driver program)? Or should I look for another platform? Because such mode of operation isn't classic for Spark.
If you use SparkContext.wholeTextFiles, then you could read the files into one RDD and each partition of the RDD would have the content of a single file. Then, you could work on each partition separately using SparkContext.mapPartitions(sort_file), where sort_file is the sorting function that you want to apply on each file. This would use concurrency better than your current solution, as long as your files are small enough that they can be processed in a single partition.
Related
1.) I understand that "Spark's operators spills data to disk if it does not fit memory allowing it to run well on any sized data".
If this is true, why do we ever get OOM (Out of Memory) errors?
2.) Increasing the no. of executor cores increases parallelism. Would that also increase the chances of OOM, because the same memory is now divided into smaller parts for each core?
3.) Spark is much more susceptible to OOM because it performs operations in memory as compared to Hive, which repeatedly reads, writes into disk. Is that correct?
There is one angle that you need to consider there. You may get memory leaks if the data is not properly distributed. That means that you need to distribute your data evenly (if possible) on the Tasks so that you reduce shuffling as much as possible and make those Tasks to manage their own data. So if you need to perform a join, if data is distributed randomly, every Task (and therefore executor) will have to:
See what data they have
Send data to other executors (and tasks) to provide the same keys they need
Request the data that is needed by that task to the others
All that data exchange may cause network bottlenecks if you have a large dataset and also will make every Task to hold their data in memory plus whatever has been sent and temporary objects. All of those will blow up memory.
So to prevent that situation you can:
Load the data already repartitioned. By that I mean, if you are loading from a DB, try Spark stride as defined here. Please refer to the partitionColumn, lowerBound, upperBound attributes. That way you will create a number of partitions on the dataframe that will set the data on different tasks based on the criteria you need. If you are going to use a join of two dataframes, try similar approach on them so that partitions are similar (for not to say same) and that will prevent shuffling over network.
When you define partitions, try to make those values as evenly distributed among tasks as possible
The size of each partition should fit on memory. Although there could be spill to disk, that would slow down performance
If you don't have a column that make the data evenly distributed, try to create one that would have n number of different values, depending on the n number of tasks that you have
If you are reading from a csv, that would make it harder to create partitions, but still it's possible. You can either split the data (csv) on multiple files and create multiple dataframes (performing a union after they are loaded) or you can read that big csv and apply a repartition on the column you need. That will create shuffling as well, but it will be done once if you cache the dataframe already repartitioned
Reading from parquet it's possible that you may have multiple files but if they are not evenly distributed (because the previous process that generated didn't do it well) you may end up on OOM errors. To prevent that situation, you can load and apply repartition on the dataframe too
Or another trick valid for csv, parquet files, orc, etc. is to create a Hive table on top of that and run a query from Spark running a distribute by clause on the data, so that you can make Hive to redistribute, instead of Spark
To your question about Hive and Spark, I think you are right up to some point. Depending on the execute engine that Hive uses in your case (map/reduce, Tez, Hive on Spark, LLAP) you can have different behaviours. With map/reduce, as they are mostly disk operations, the chance to have a OOM is much lower than on Spark. Actually from Memory point of view, map/reduce is not that affected because of a skewed data distribution. But (IMHO) your goal should be to find always the best data distribution for the Spark job you are running and that will prevent that problem
Another consideration is if you are testing in a dev environment that doesn't have same data as in a prod environment. I suppose the data distribution should be similar although volumes may differ a lot (I am talking from experience ;)). In that case, when you assign Spark tuning parameters on the spark-submit command, they may be different in prod. So you need to invest some time on finding the best approach on dev and fine tune in prod
Huge majority of OOM in Spark are on the driver, not executors. This is usually a result of running .collect or similar actions on a dataset that won't fit in the driver memory.
Spark does a lot of work under the hood to parallelize the work, when using structured APIs (in contrast to RDDs) the chances of causing OOM on executor are really slim. Some combinations of cluster configuration and jobs can cause memory pressure that will impact performance and cause lots of garbage collection to happen so you need to address it, however spark should be able to handle low memory without explicit exception.
Not really - as above, Spark should be able to recover from memory issues when using structured APIs, however it may need intervention if you see garbage collection and performance impact.
Before I write dataframe into hdfs, I coalesce(1) to make it write only one file, so it is easily to handle thing manually when copying thing around, get from hdfs, ...
I would code like this to write output.
outputData.coalesce(1).write.parquet(outputPath)
(outputData is org.apache.spark.sql.DataFrame)
I would like to ask if their are any impact on performance vs not coalesce
outputData.write.parquet(outputPath)
Yes, it will write with 1 worker.
So, even through you give 10 CPU core, it will write with 1 worker (single partition).
Problem if your file very big (10 gb or more). But recommend if you have small file (100 mb)
I would not recommend doing that. The whole purpose of distributed computing is to have data and processing sitting on multiple machine and capitalize the benefits of CPU/Memory of many machines (worker nodes).
In your case, you are trying to put everything in one place. Why do you need a distributed file system if you want to write into a single file with just one partition? Performance can be an issue but it can be only assessed after you check before/after using Coalesce function on huge amount of data that is spread across multiple nodes on cluster.
Though really not suggested when dealing with huge data, Using coalesce(1) can be handy when there are too many small partition files in the _temporary and the file movement is taking quite a bit of time to move them into the proper directories.
In CouchDB and system designs like Incoop, there's a concept called "Incremental MapReduce" where results from previous executions of a MapReduce algorithm are saved and used to skip over sections of input data that haven't been changed.
Say I have 1 million rows divided into 20 partitions. If I run a simple MapReduce over this data, I could cache/store the result of reducing each separate partition, before they're combined and reduced again to produce the final result. If I only change data in the 19th partition then I only need to run the map & reduce steps on the changed section of the data, and then combine the new result with the saved reduce results from the unchanged partitions to get an updated result. Using this sort of catching I'd be able to skip almost 95% of the work for re-running a MapReduce job on this hypothetical dataset.
Is there any good way to apply this pattern to Spark? I know I could write my own tool for splitting up input data into partitions, checking if I've already processed those partitions before, loading them from a cache if I have, and then running the final reduce to join all the partitions together. However, I suspect that there's an easier way to approach this.
I've experimented with checkpointing in Spark Streaming, and that is able to store results between restarts, which is almost what I'm looking for, but I want to do this outside of a streaming job.
RDD caching/persisting/checkpointing almost looks like something I could build off of - it makes it easy to keep intermediate computations around and reference them later, but I think cached RDDs are always removed once the SparkContext is stopped, even if they're persisted to disk. So caching wouldn't work for storing results between restarts. Also, I'm not sure if/how checkpointed RDDs are supposed to be loaded when a new SparkContext is started... They seem to be stored under a UUID in the checkpoint directory that's specific to a single instance of the SparkContext.
Both use cases suggested by the article (incremental logs processing and incremental query processing) can be generally solved by Spark Streaming.
The idea is that you have incremental updates coming in using DStreams abstraction. Then, you can process new data, and join it with previous calculation either using time window based processing or using arbitrary stateful operations as part of Structured Stream Processing. Results of the calculation can be later dumped to some sort of external sink like database or file system, or they can be exposed as an SQL table.
If you're not building an online data processing system, regular Spark can be used as well. It's just a matter of how incremental updates get into the process, and how intermediate state is saved. For example, incremental updates can appear under some path on a distributed file system, while intermediate state containing previous computation joined with new data computation can be dumped, again, to the same file system.
I'm trying to benchmark a few approaches to putting an image processing algorithm into apache spark. For one step in this algorithm, a computation on a pixel in the image will depend on an unknown amount of surrounding data, so we can't partition the image with guaranteed sufficient overlap a priori.
One solution to that problem I need to benchmark is for a worker node to ask the master node for more data when it encounters a pixel with insufficient surrounding data. I'm not convinced this is the way to do things, but I need to benchmark it anyway because of reasons.
Unfortunately, after a bunch of googling and reading docs I can't find any way for a processingFunc called as part of sc.parallelize(partitions).map(processingFunc) to query the master node for more data from a different partition mid-computation.
Does a way for a worker node to ask the master for more data exist in spark, or will I need to hack something together that kind of goes around spark?
Master Node in Spark is for allocating the resources to a particular job and once the resources are allocated, the Driver ships the complete code with all its dependencies to the various executors.
The first step in every code is to load the data to the Spark cluster. You can read the data from any underlying data repository like Database, filesystem, webservices etc.
Once data is loaded it is wrapped into an RDD which is partitioned across the nodes in the cluster and further stored in the workers/ Executors Memory. Though you can control the number of partitions by leveraging various RDD API's but you should do it only when you have valid reasons to do so.
Now all operations are performed over RDD's using its various methods/ Operations exposed by RDD API. RDD keep tracks of partitions and partitioned data and depending upon the need or request it automatically query the appropriate partition.
In nutshell, you do not have to worry about the way data is partitioned by RDD or which partition stores which data and how they communicate with each other but if you do care, then you can write your own custom partitioner, instructing Spark of how to partition your data.
Secondly if your data cannot be partitioned then I do not think Spark would be an ideal choice because that will result in processing of everything in 1 single machine which itself is contrary to the idea of distributed computing.
Not sure what is exactly your use case but there are people who have been leveraging Spark for Image processing. see here for the comments from Databricks
I have a dataset in a csv file that occupies two blocks in HDFS and replicated on two nodes, A and B. Each node has a copy of the dataset.
When Spark starts processing the data, I have seen two ways how Spark loads the dataset as input. It either loads the entire dataset into memory on one node and perform most of the tasks on it or loads the dataset into two nodes and spill the tasks on both nodes (based on what I observed on the history server). For both cases, there is sufficient capacity to keep the whole datasets in memory.
I repeated the same experiment multiple times and Spark seemed to alternate between these two ways. Supposedly Spark inherits the input split location as in a MapReduce job. From my understanding, MapReduce should be able to take advantage of two nodes. I don't understand why Spark or MapReduce will alternate between the two cases.
When only one node is used for processing, the performance is worse.
When your loading the data in Spark you can specify the minimum number of splits and this will force Spark to load the data on multiple machines (with the textFile api you would add minPartitions=2 to your call.