Skip to main content

HBase: MSLAB and CMS vs. ParallelGC

Listen:
Tuning Java opts for HBase, for example, are necessary steps to get the best performance and stability in large installations. The optimal recommendation looks like:
HBASE_OPTS="-XX:+UseParNewGC -XX:+UseConcMarkSweepGC -XX:CMSInitiatingOccupancyFraction=70 -XX:+CMSParallelRemarkEnabled"

But you can also achieve great success with:
HBASE_OPTS="-server -XX:+UseParallelGC XX:+UseParallelOldGC -XX:ParallelGCThreads=8"

What are the differences between ParallelGC and CMS? 

CMS uses more CPU, but runs concurrently. If a thread is failing, CMS falls back to a non-parallel mode and stops the VM for the entire time it's collecting. But this risk can be minimized by using MSLAB in your HBase configuration.
ParallelGC have a better throughput and longer pause times, and stop the VM on every collection. Means for HBase, you'll have a pause (around 1 sec per GB), which can lead on high loaded clusters to outages in a non acceptable time range.

MSLAB (MemStore-Local Allocation Buffers)

The most GC pauses are caused by old-gen fragmentation, and CMS can't defragment without pause the VM (Juliet pause). MSLAB now moves the memstore allocations into the configured chunks into the old generation. When you start or upgrade into HBase 0.92x, MSLAB is enabled per default (http://hbase.apache.org/book/upgrade0.92.html). 

hbase.hregion.memstore.mslab.enabled=true
hbase.hregion.memstore.mslab.chunksize=2MB (2MB per default)
hbase.hregion.memstore.mslab.max.allocation=256KB (256KB per default)


More about MSLAB you can find in Todd's presentation: http://www.slideshare.net/cloudera/hbase-hug-presentation


Comments

Popular posts from this blog

Deal with corrupted messages in Apache Kafka

Under some strange circumstances, it can happen that a message in a Kafka topic is corrupted. This often happens when using 3rd party frameworks with Kafka. In addition, Kafka < 0.9 does not have a lock on Log.read() at the consumer read level, but does have a lock on Log.write(). This can lead to a rare race condition as described in KAKFA-2477 [1]. A likely log entry looks like this: ERROR Error processing message, stopping consumer: (kafka.tools.ConsoleConsumer$) kafka.message.InvalidMessageException: Message is corrupt (stored crc = xxxxxxxxxx, computed crc = yyyyyyyyyy Kafka-Tools Kafka stores the offset of each consumer in Zookeeper. To read the offsets, Kafka provides handy tools [2]. But you can also use zkCli.sh, at least to display the consumer and the stored offsets. First we need to find the consumer for a topic (> Kafka 0.9): bin/kafka-consumer-groups.sh --zookeeper management01:2181 --describe --group test Prior to Kafka 0.9, the only way to get this inform

Hive query shows ERROR "too many counters"

A hive job face the odd " Too many counters:"  like Ended Job = job_xxxxxx with exception 'org.apache.hadoop.mapreduce.counters.LimitExceededException(Too many counters: 201 max=200)' FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.MapRedTask Intercepting System.exit(1) These happens when operators are used in queries ( Hive Operators ). Hive creates 4 counters per operator, max upto 1000, plus a few additional counters like file read/write, partitions and tables. Hence the number of counter required is going to be dependent upon the query.  To avoid such exception, configure " mapreduce.job.counters.max " in mapreduce-site.xml to a value above 1000. Hive will fail when he is hitting the 1k counts, but other MR jobs not. A number around 1120 should be a good choice. Using " EXPLAIN EXTENDED " and " grep -ri operators | wc -l " print out the used numbers of operators. Use this value to tweak the MR s

AI's False Reality: Understanding Hallucination

Artificial Intelligence (AI) has leapfrogged to the poster child of technological innovation, on track to transform industries in a scale similar to the Industrial Revolution of the 1800s. But in this case, as cutting-edge technology, AI presents its own unique challenge, exploiting our human behavior of "love to trust", we as humans face a challenge: AI hallucinations. This phenomenon, where AI models generate outputs that are factually incorrect, misleading, or entirely fabricated, raises complex questions about the reliability and trust of AI models and larger systems. The tendency for AI to hallucinate comes from several interrelated factors. Overfitting – a condition where models become overly specialized to their training data – can lead to confident but wildly inaccurate responses when presented with novel scenarios (Guo et al., 2017). Moreover, biases embedded within datasets shape the models' understanding of the world; if these datasets are flawed or unreprese