I am trying to join two large spark dataframes and keep running into this error:
Container killed by YARN for exceeding memory limits. 24 GB of 22 GB physical memory used. Consider boosting spark.yarn.executor.memoryOverhead.
This seems like a common issue among spark users, but I can't seem to find any solid descriptions of what spark.yarn.executor.memoryOverheard is. In some cases it sounds like it's a kind of memory buffer before YARN kills the container (e.g. 10GB was requested, but YARN won't kill the container until it uses 10.2GB). In other cases it sounds like it's being used to to do some kind of data accounting tasks that are completely separate from the analysis that I want to perform. My questions are:
What is the spark.yarn.executor.memoryOverhead being using for?
What is the benefit of increasing this kind of memory instead of
executor memory (or the number of executors)?
In general, are there things steps I can take to reduce my
spark.yarn.executor.memoryOverhead usage (e.g. particular
datastructures, limiting the width of the dataframes, using fewer executors with more memory, etc)?
Overhead options are nicely explained in the configuration document:
This is memory that accounts for things like VM overheads, interned strings, other native overheads, etc. This tends to grow with the executor size (typically 6-10%).
This also includes user objects if you use one of the non-JVM guest languages (Python, R, etc...).
Related
I am new to spark so am following this amazing tutorial from sparkbyexamples.com and while reading I found this section:
Shuffle partition size & Performance
Based on your dataset size, a number of cores and memory PySpark
shuffling can benefit or harm your jobs. When you dealing with less
amount of data, you should typically reduce the shuffle partitions
otherwise you will end up with many partitioned files with less number
of records in each partition. which results in running many tasks with
lesser data to process.
On other hand, when you have too much of data and having less number
of partitions results in fewer longer running tasks and some times you
may also get out of memory error.
Getting the right size of the shuffle partition is always tricky and
takes many runs with different values to achieve the optimized number.
This is one of the key properties to look for when you have
performance issues on PySpark jobs.
Can someone help me understand how do you determine how many shuffle partitions you will need for your job?
As you quoted, it’s tricky, but this is my strategy:
If you’re using “static allocation”, means you tell Spark how many executors you want to allocate for the job, then it’s easy, number of partitions could be executors * cores per executor * factor. factor = 1 means each executor will handle 1 job, factor = 2 means each executor will handle 2 jobs, and so on
If you’re using “dynamic allocation”, then it’s trickier. You can read the long description here https://databricks.com/blog/2021/03/17/advertising-fraud-detection-at-scale-at-t-mobile.html. The general idea is you need to answer many questions like what’s the size if your data (how big in terms of gigabytes), how its structure looks like (how many files, how many folders, how many rows etc), how would you read it (from hdfs or from hive or from jdbc), how much resources do you have (cores, executors, memory), … Then you run and benchmark over and over to find the sweet spot that is “just right” for your circumstances.
Update #1:
So what is the general industry practice, will a company simply use first tactic and allocate more hardware or they will use dynamic allocation?
Usually, if you have an on-premise Hadoop environment, you can choose between static (default mode) and dynamic allocation (advanced mode). Also, I often start with dynamic because I have no idea how big the data and its transformation is, so stick with dynamic give me flexibility to expand my work without thinking too much about Spark configuration. But you also can start with static if you want to, nothing preventing you to do so.
Then eventually, when it came to productionize process, you also can choose between static (very stable but consumes more resources) vs dynamic (less stable, i.e fail sometimes due to resources allocation, but save resources.
Finally, most Hadoop cloud solution (like Databricks) come with dynamic allocation by default, which is is less costly.
The RDDs that are cached (in total 8) are not big, only around 30G, however, on Hadoop UI, it shows that the Spark application is taking lots of memory (no active jobs are running), i.e. 1.4T, why so much?
Why it shows around 100 executors (here, i.e. vCores) even when there's no active jobs running?
Also, if cached RDDs are stored across 100 executors, are those executors preserved and no more other Spark apps can use them for running tasks any more? To rephrase the question: will preserving a little memory resource (.cache) in executors prevents other Spark app from leveraging the idle computing resource of them?
Is there any potential Spark config / zeppelin config that can cause this phenomenon?
UPDATE 1
After checking the Spark conf (zeppelin), it seems there's the default (configured by administrator by default) setting for spark.executor.memory=10G, which is probably the reason why.
However, here's a new question: Is it possible to keep only the memory needed for the cached RDDs in each executors and release the rest, instead of holding always the initially set memory spark.executor.memory=10G?
Spark configuration
Perhaps you can try to repartition(n) your RDD to a fewer n < 100 partitions before caching. A ~30GB RDD would probably fit into storage memory of ten 10GB executors. A good overview of Spark memory management can be found here. This way, only those executors that hold cached blocks will be "pinned" to your application, while the rest can be reclaimed by YARN via Spark dynamic allocation after spark.dynamicAllocation.executorIdleTimeout (default 60s).
Q: Is it possible to keep only the memory needed for the cached RDDs in each executors and release the rest, instead of holding always the initially set memory spark.executor.memory=10G?
When Spark uses YARN as its execution engine, YARN allocates the containers of a specified (by application) size -- at least spark.executor.memory+spark.executor.memoryOverhead, but may be even bigger in case of pyspark -- for all the executors. How much memory Spark actually uses inside a container becomes irrelevant, since the resources allocated to a container will be considered off-limits to other YARN applications.
Spark assumes that your data is equally distributed on all the executors and tasks. That's the reason why you set memory per task. So to make Spark to consume less memory, your data has to be evenly distributed:
If you are reading from Parquet files or CSVs, make sure that they have similar sizes. Running repartition() causes shuffling, which if the data is so skewed may cause other problems if executors don't have enough resources
Cache won't help to release memory on the executors because it doesn't redistribute the data
Can you please see under "Event Timeline" on the Stages "how big are the green bars?" Normally that's tied to the data distribution, so that's a way to see how much data is loaded (proportionally) on every task and how much they are doing. As all tasks have same memory assigned, you can see graphically if resources are wasted (in case there are mostly tiny bars and few big bars). A sample of wasted resources can be seen on the image below
There are different ways to create evenly distributed files for your process. I mention some possibilities, but for sure there are more:
Using Hive and DISTRIBUTE BY clause: you need to use a field that is equally balanced in order to create as many files (and with proper size) as expected
If the process creating those files is a Spark process reading from a DB, try to create as many connections as files you need and use a proper field to populate Spark partitions. That is achieved, as explained here and here with partitionColumn, lowerBound, upperBound and numPartitions properties
Repartition may work, but see if coalesce also make sense in your process or in the previous one generating the files you are reading from
I'm relatively new to spark and I have a few questions related to the tuning optimizations with respect to the spark submit command.
I have followed : How to tune spark executor number, cores and executor memory?
and I understand how to utilise maximum resources out of my spark cluster.
However, I was recently asked how to define the number of cores, memory and cores when I have a relatively smaller operation to do as if I give maximum resources, it is going to be underutilised .
For instance,
if I have to just do a merge job (read files from hdfs and write one single huge file back to hdfs using coalesce) for about 60-70 GB (assume each file is of 128 mb in size which is the block size of HDFS) of data(in avro format without compression), what would be the ideal memory, no of executor and cores required for this?
Assume I have the configurations of my nodes same as the one mentioned in the link above.
I can't understand the concept of how much memory will be used up by the entire job provided there are no joins, aggregations etc.
The amount of memory you will need depends on what you run before the write operation. If all you're doing is reading data combining it and writing it out, then you will need very little memory per cpu because the dataset is never fully materialized before writing it out. If you're doing joins/group-by/other aggregate operations all of those will require much ore memory. The exception to this rule is that spark isn't really tuned for large files and generally is much more performant when dealing with sets of reasonably sized files. Ultimately the best way to get your answers is to run your job with the default parameters and see what blows up.
I've noticed strange behavior when running a pyspark application with spark 2.0. In the first step in my script involving a reduceByKey (and thus shuffle) operation, I observe that the amount the shuffle writes is roughly in line with my expectations, but that much more spills occur than I had expected. I tried to avoid these spills by increasing the amount of memory assigned per executor up to 8x the original amount, but see basically no difference in the amount spilled. Strangely, I also see that while this stage is running, hardly any of the assigned storage memory is used (as reported in the executors tab in the spark web UI).
I saw this earlier question, which led me to believe that increasing executor memory might help avoid the spills: How to optimize shuffle spill in Apache Spark application
. This leads me to believe that some hard limit is leading to the spills, and not the spark.shuffle.memoryFraction parameter. Does such a hard limit exist, possibly among HDFS parameters? Otherwise, what could be done to avoid spills besides increasing executor memory?
Many thanks, R
Spilling behavior in PySpark is controlled using spark.python.worker.memory:
Amount of memory to use per python worker process during aggregation, in the same format as JVM memory strings (e.g. 512m, 2g). If the memory used during aggregation goes above this amount, it will spill the data into disks.
which is by default set to 512MB. Moreover PySpark uses its own reducing mechanism with External(GroupBy|Sorter|Merger) and exhibits slightly different behavior than its native counterpart.
I am talking about the standalone mode of spark.
Lets say the value of SPARK_WORKER_CORES=200 and there are only 4 cores available on the node where I am trying to start the worker. Will the worker get 4 cores and continue or will the worker not start at all ?
A similar case, If I set SPARK_WORKER_MEMORY=32g and there is only 2g of memory actually available on that node ?
"Cores" in Spark is sort of a misnomer. "Cores" actually corresponds to the number of threads created to process data. So, you could have an arbitrarily large number of cores without an explicit failure. That being said, overcommitting by 50x will likely lead to incredibly poor performance due to context switching and overhead costs. This means that for both workers and executors you can arbitrarily increase this number. In practice in Spark Standalone, I've generally seen this overcommitted no more than 2-3x the number of logical cores.
When it comes to specifying worker memory, once again, you can in theory increase it to an arbitrarily large number. This is because, for a worker, the memory amount specifies how much it is allowed to allocate for executors, but it doesn't explicitly allocate that amount when you start the worker. Therefore you can make this value much larger than physical memory.
The caveat here is that when you start up an executor, if you set the executor memory to be greater than the amount of physical memory, your executors will fail to start. This is because executor memory directly corresponds to the -Xmx setting of the java process for that executor.