| <?xml version="1.0"?> |
| <!-- |
| Licensed to the Apache Software Foundation (ASF) under one or more |
| contributor license agreements. See the NOTICE file distributed with |
| this work for additional information regarding copyright ownership. |
| The ASF licenses this file to You under the Apache License, Version 2.0 |
| (the "License"); you may not use this file except in compliance with |
| the License. You may obtain a copy of the License at |
| |
| http://www.apache.org/licenses/LICENSE-2.0 |
| |
| Unless required by applicable law or agreed to in writing, software |
| distributed under the License is distributed on an "AS IS" BASIS, |
| WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| See the License for the specific language governing permissions and |
| limitations under the License. |
| --> |
| <!DOCTYPE document PUBLIC "-//APACHE//DTD Documentation V2.0//EN" "http://forrest.apache.org/dtd/document-v20.dtd"> |
| <document> |
| <header> |
| <title>GridMix</title> |
| </header> |
| <body> |
| <section id="overview"> |
| <title>Overview</title> |
| <p>GridMix is a benchmark for Hadoop clusters. It submits a mix of |
| synthetic jobs, modeling a profile mined from production loads.</p> |
| <p>There exist three versions of the GridMix tool. This document |
| discusses the third (checked into <code>src/contrib</code>), distinct |
| from the two checked into the <code>src/benchmarks</code> sub-directory. |
| While the first two versions of the tool included stripped-down versions |
| of common jobs, both were principally saturation tools for stressing the |
| framework at scale. In support of a broader range of deployments and |
| finer-tuned job mixes, this version of the tool will attempt to model |
| the resource profiles of production jobs to identify bottlenecks, guide |
| development, and serve as a replacement for the existing GridMix |
| benchmarks.</p> |
| <p>To run GridMix, you need a MapReduce job trace describing the job mix |
| for a given cluster. Such traces are typically generated by Rumen (see |
| Rumen documentation). GridMix also requires input data from which the |
| synthetic jobs will be reading bytes. The input data need not be in any |
| particular format, as the synthetic jobs are currently binary readers. |
| If you are running on a new cluster, an optional step generating input |
| data may precede the run.</p> |
| <p>In order to emulate the load of production jobs from a given cluster |
| on the same or another cluster, follow these steps:</p> |
| <ol> |
| <li>Locate the job history files on the production cluster. This |
| location is specified by the |
| <code>mapreduce.jobtracker.jobhistory.completed.location</code> |
| configuration property of the cluster.</li> |
| <li>Run Rumen to build a job trace in JSON format for all or select |
| jobs.</li> |
| <li><em>(Optional)</em> Use Rumen to fold this job trace to scale |
| the load.</li> |
| <li>Use GridMix with the job trace on the benchmark cluster.</li> |
| </ol> |
| <p>Jobs submitted by GridMix have names of the form |
| "<code>GRIDMIXnnnnn</code>", where |
| "<code>nnnnn</code>" is a sequence number padded with leading |
| zeroes.</p> |
| </section> |
| <section id="usage"> |
| <title>Usage</title> |
| <p>Basic command-line usage without configuration parameters:</p> |
| <source> |
| org.apache.hadoop.mapred.gridmix.Gridmix [-generate <size>] [-users <users-list>] <iopath> <trace> |
| </source> |
| <p>Basic command-line usage with configuration parameters:</p> |
| <source> |
| org.apache.hadoop.mapred.gridmix.Gridmix \ |
| -Dgridmix.client.submit.threads=10 -Dgridmix.output.directory=foo \ |
| [-generate <size>] [-users <users-list>] <iopath> <trace> |
| </source> |
| <note> |
| Configuration parameters like |
| <code>-Dgridmix.client.submit.threads=10</code> and |
| <code>-Dgridmix.output.directory=foo</code> as given above should |
| be used <em>before</em> other GridMix parameters. |
| </note> |
| <p>The <code>-generate</code> option is used to generate input data |
| for the synthetic jobs. It accepts size suffixes, e.g. |
| <code>100g</code> will generate 100 * 2<sup>30</sup> bytes.</p> |
| <p>The <code>-users</code> option is used to point to a users-list |
| file (see <a href="#usersqueues">Emulating Users and Queues</a>).</p> |
| <p>The <code><iopath></code> parameter is the destination |
| directory for generated output and/or the directory from which input |
| data will be read. Note that this can either be on the local file-system |
| or on HDFS, but it is highly recommended that it be the same as that for |
| the original job mix so that GridMix puts the same load on the local |
| file-system and HDFS respectively.</p> |
| <p>The <code><trace></code> parameter is a path to a job trace |
| generated by Rumen. This trace can be compressed (it must be readable |
| using one of the compression codecs supported by the cluster) or |
| uncompressed. Use "-" as the value of this parameter if you |
| want to pass an <em>uncompressed</em> trace via the standard |
| input-stream of GridMix.</p> |
| <p>The class <code>org.apache.hadoop.mapred.gridmix.Gridmix</code> can |
| be found in the JAR |
| <code>contrib/gridmix/hadoop-$VERSION-gridmix.jar</code> inside your |
| Hadoop installation, where <code>$VERSION</code> corresponds to the |
| version of Hadoop installed. A simple way of ensuring that this class |
| and all its dependencies are loaded correctly is to use the |
| <code>hadoop</code> wrapper script in Hadoop:</p> |
| <source> |
| hadoop jar <gridmix-jar> org.apache.hadoop.mapred.gridmix.Gridmix \ |
| [-generate <size>] [-users <users-list>] <iopath> <trace> |
| </source> |
| <p>The supported configuration parameters are explained in the |
| following sections.</p> |
| </section> |
| <section id="cfgparams"> |
| <title>General Configuration Parameters</title> |
| <p/> |
| <table> |
| <tr> |
| <th>Parameter</th> |
| <th>Description</th> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.output.directory</code> |
| </td> |
| <td>The directory into which output will be written. If specified, |
| <code>iopath</code> will be relative to this parameter. The |
| submitting user must have read/write access to this directory. The |
| user should also be mindful of any quota issues that may arise |
| during a run. The default is "<code>gridmix</code>".</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.client.submit.threads</code> |
| </td> |
| <td>The number of threads submitting jobs to the cluster. This |
| also controls how many splits will be loaded into memory at a given |
| time, pending the submit time in the trace. Splits are pre-generated |
| to hit submission deadlines, so particularly dense traces may want |
| more submitting threads. However, storing splits in memory is |
| reasonably expensive, so you should raise this cautiously. The |
| default is 1 for the SERIAL job-submission policy (see |
| <a href="#policies">Job Submission Policies</a>) and one more than |
| the number of processors on the client machine for the other |
| policies.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.submit.multiplier</code> |
| </td> |
| <td>The multiplier to accelerate or decelerate the submission of |
| jobs. The time separating two jobs is multiplied by this factor. |
| The default value is 1.0. This is a crude mechanism to size |
| a job trace to a cluster.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.client.pending.queue.depth</code> |
| </td> |
| <td>The depth of the queue of job descriptions awaiting split |
| generation. The jobs read from the trace occupy a queue of this |
| depth before being processed by the submission threads. It is |
| unusual to configure this. The default is 5.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.gen.blocksize</code> |
| </td> |
| <td>The block-size of generated data. The default value is 256 |
| MiB.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.gen.bytes.per.file</code> |
| </td> |
| <td>The maximum bytes written per file. The default value is 1 |
| GiB.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.min.file.size</code> |
| </td> |
| <td>The minimum size of the input files. The default limit is 128 |
| MiB. Tweak this parameter if you see an error-message like |
| "Found no satisfactory file" while testing GridMix with |
| a relatively-small input data-set.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.max.total.scan</code> |
| </td> |
| <td>The maximum size of the input files. The default limit is 100 |
| TiB.</td> |
| </tr> |
| </table> |
| </section> |
| <section id="jobtypes"> |
| <title>Job Types</title> |
| <p>GridMix takes as input a job trace, essentially a stream of |
| JSON-encoded job descriptions. For each job description, the submission |
| client obtains the original job submission time and for each task in |
| that job, the byte and record counts read and written. Given this data, |
| it constructs a synthetic job with the same byte and record patterns as |
| recorded in the trace. It constructs jobs of two types:</p> |
| <table> |
| <tr> |
| <th>Job Type</th> |
| <th>Description</th> |
| </tr> |
| <tr> |
| <td> |
| <code>LOADJOB</code> |
| </td> |
| <td>A synthetic job that emulates the workload mentioned in Rumen |
| trace. In the current version we are supporting I/O. It reproduces |
| the I/O workload on the benchmark cluster. It does so by embedding |
| the detailed I/O information for every map and reduce task, such as |
| the number of bytes and records read and written, into each |
| job's input splits. The map tasks further relay the I/O patterns of |
| reduce tasks through the intermediate map output data.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>SLEEPJOB</code> |
| </td> |
| <td>A synthetic job where each task does <em>nothing</em> but sleep |
| for a certain duration as observed in the production trace. The |
| scalability of the Job Tracker is often limited by how many |
| heartbeats it can handle every second. (Heartbeats are periodic |
| messages sent from Task Trackers to update their status and grab new |
| tasks from the Job Tracker.) Since a benchmark cluster is typically |
| a fraction in size of a production cluster, the heartbeat traffic |
| generated by the slave nodes is well below the level of the |
| production cluster. One possible solution is to run multiple Task |
| Trackers on each slave node. This leads to the obvious problem that |
| the I/O workload generated by the synthetic jobs would thrash the |
| slave nodes. Hence the need for such a job.</td> |
| </tr> |
| </table> |
| <p>The following configuration parameters affect the job type:</p> |
| <table> |
| <tr> |
| <th>Parameter</th> |
| <th>Description</th> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.job.type</code> |
| </td> |
| <td>The value for this key can be one of LOADJOB or SLEEPJOB. The |
| default value is LOADJOB.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.key.fraction</code> |
| </td> |
| <td>For a LOADJOB type of job, the fraction of a record used for |
| the data for the key. The default value is 0.1.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.sleep.maptask-only</code> |
| </td> |
| <td>For a SLEEPJOB type of job, whether to ignore the reduce |
| tasks for the job. The default is <code>false</code>.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.sleep.fake-locations</code> |
| </td> |
| <td>For a SLEEPJOB type of job, the number of fake locations |
| for map tasks for the job. The default is 0.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.sleep.max-map-time</code> |
| </td> |
| <td>For a SLEEPJOB type of job, the maximum runtime for map |
| tasks for the job in milliseconds. The default is unlimited.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.sleep.max-reduce-time</code> |
| </td> |
| <td>For a SLEEPJOB type of job, the maximum runtime for reduce |
| tasks for the job in milliseconds. The default is unlimited.</td> |
| </tr> |
| </table> |
| </section> |
| <section id="policies"> |
| <title>Job Submission Policies</title> |
| <p>GridMix controls the rate of job submission. This control can be |
| based on the trace information or can be based on statistics it gathers |
| from the Job Tracker. Based on the submission policies users define, |
| GridMix uses the respective algorithm to control the job submission. |
| There are currently three types of policies:</p> |
| <table> |
| <tr> |
| <th>Job Submission Policy</th> |
| <th>Description</th> |
| </tr> |
| <tr> |
| <td> |
| <code>STRESS</code> |
| </td> |
| <td>Keep submitting jobs so that the cluster remains under stress. |
| In this mode we control the rate of job submission by monitoring |
| the real-time load of the cluster so that we can maintain a stable |
| stress level of workload on the cluster. Based on the statistics we |
| gather we define if a cluster is <em>underloaded</em> or |
| <em>overloaded</em>. We consider a cluster <em>underloaded</em> if |
| and only if the following three conditions are true: |
| <ol> |
| <li>the number of pending and running jobs are under a threshold |
| TJ</li> |
| <li>the number of pending and running maps are under threshold |
| TM</li> |
| <li>the number of pending and running reduces are under threshold |
| TR</li> |
| </ol> |
| The thresholds TJ, TM and TR are proportional to the size of the |
| cluster and map, reduce slots capacities respectively. In case of a |
| cluster being <em>overloaded</em>, we throttle the job submission. |
| In the actual calculation we also weigh each running task with its |
| remaining work - namely, a 90% complete task is only counted as 0.1 |
| in calculation. Finally, to avoid a very large job blocking other |
| jobs, we limit the number of pending/waiting tasks each job can |
| contribute.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>REPLAY</code> |
| </td> |
| <td>In this mode we replay the job traces faithfully. This mode |
| exactly follows the time-intervals given in the actual job |
| trace.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>SERIAL</code> |
| </td> |
| <td>In this mode we submit the next job only once the job submitted |
| earlier is completed.</td> |
| </tr> |
| </table> |
| <p>The following configuration parameters affect the job submission |
| policy:</p> |
| <table> |
| <tr> |
| <th>Parameter</th> |
| <th>Description</th> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.job-submission.policy</code> |
| </td> |
| <td>The value for this key would one of the three: STRESS, REPLAY or |
| SERIAL. In most of the cases the value of key would be STRESS or |
| REPLAY. The default value is STRESS.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.throttle.jobs-to-tracker-ratio</code> |
| </td> |
| <td>In STRESS mode, the minimum ratio of running jobs to Task |
| Trackers in a cluster for the cluster to be considered |
| <em>overloaded</em>. This is the threshold TJ referred to earlier. |
| The default is 1.0.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.throttle.maps.task-to-slot-ratio</code> |
| </td> |
| <td>In STRESS mode, the minimum ratio of pending and running map |
| tasks (i.e. incomplete map tasks) to the number of map slots for |
| a cluster for the cluster to be considered <em>overloaded</em>. |
| This is the threshold TM referred to earlier. Running map tasks are |
| counted partially. For example, a 40% complete map task is counted |
| as 0.6 map tasks. The default is 2.0.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.throttle.reduces.task-to-slot-ratio</code> |
| </td> |
| <td>In STRESS mode, the minimum ratio of pending and running reduce |
| tasks (i.e. incomplete reduce tasks) to the number of reduce slots |
| for a cluster for the cluster to be considered <em>overloaded</em>. |
| This is the threshold TR referred to earlier. Running reduce tasks |
| are counted partially. For example, a 30% complete reduce task is |
| counted as 0.7 reduce tasks. The default is 2.5.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.throttle.maps.max-slot-share-per-job</code> |
| </td> |
| <td>In STRESS mode, the maximum share of a cluster's map-slots |
| capacity that can be counted toward a job's incomplete map tasks in |
| overload calculation. The default is 0.1.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.throttle.reducess.max-slot-share-per-job</code> |
| </td> |
| <td>In STRESS mode, the maximum share of a cluster's reduce-slots |
| capacity that can be counted toward a job's incomplete reduce tasks |
| in overload calculation. The default is 0.1.</td> |
| </tr> |
| </table> |
| </section> |
| <section id="usersqueues"> |
| <title>Emulating Users and Queues</title> |
| <p>Typical production clusters are often shared with different users and |
| the cluster capacity is divided among different departments through job |
| queues. Ensuring fairness among jobs from all users, honoring queue |
| capacity allocation policies and avoiding an ill-behaving job from |
| taking over the cluster adds significant complexity in Hadoop software. |
| To be able to sufficiently test and discover bugs in these areas, |
| GridMix must emulate the contentions of jobs from different users and/or |
| submitted to different queues.</p> |
| <p>Emulating multiple queues is easy - we simply set up the benchmark |
| cluster with the same queue configuration as the production cluster and |
| we configure synthetic jobs so that they get submitted to the same queue |
| as recorded in the trace. However, not all users shown in the trace have |
| accounts on the benchmark cluster. Instead, we set up a number of testing |
| user accounts and associate each unique user in the trace to testing |
| users in a round-robin fashion.</p> |
| <p>The following configuration parameters affect the emulation of users |
| and queues:</p> |
| <table> |
| <tr> |
| <th>Parameter</th> |
| <th>Description</th> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.job-submission.use-queue-in-trace</code> |
| </td> |
| <td>When set to <code>true</code> it uses exactly the same set of |
| queues as those mentioned in the trace. The default value is |
| <code>false</code>.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.job-submission.default-queue</code> |
| </td> |
| <td>Specifies the default queue to which all the jobs would be |
| submitted. If this parameter is not specified, GridMix uses the |
| default queue defined for the submitting user on the cluster.</td> |
| </tr> |
| <tr> |
| <td> |
| <code>gridmix.user.resolve.class</code> |
| </td> |
| <td>Specifies which <code>UserResolver</code> implementation to use. |
| We currently have three implementations: |
| <ol> |
| <li><code>org.apache.hadoop.mapred.gridmix.EchoUserResolver</code> |
| - submits a job as the user who submitted the original job. All |
| the users of the production cluster identified in the job trace |
| must also have accounts on the benchmark cluster in this case.</li> |
| <li><code>org.apache.hadoop.mapred.gridmix.SubmitterUserResolver</code> |
| - submits all the jobs as current GridMix user. In this case we |
| simply map all the users in the trace to the current GridMix user |
| and submit the job.</li> |
| <li><code>org.apache.hadoop.mapred.gridmix.RoundRobinUserResolver</code> |
| - maps trace users to test users in a round-robin fashion. In |
| this case we set up a number of testing user accounts and |
| associate each unique user in the trace to testing users in a |
| round-robin fashion.</li> |
| </ol> |
| The default is |
| <code>org.apache.hadoop.mapred.gridmix.SubmitterUserResolver</code>.</td> |
| </tr> |
| </table> |
| <p>If the parameter <code>gridmix.user.resolve.class</code> is set to |
| <code>org.apache.hadoop.mapred.gridmix.RoundRobinUserResolver</code>, |
| we need to define a users-list file with a list of test users. |
| This is specified using the <code>-users</code> option to GridMix.</p> |
| <note> |
| Specifying a users-list file using the <code>-users</code> option is |
| mandatory when using the round-robin user-resolver. Other user-resolvers |
| ignore this option. |
| </note> |
| <p>A users-list file has one user per line, each line of the format:</p> |
| <source> |
| <username> |
| </source> |
| <p>For example:</p> |
| <source> |
| user1 |
| user2 |
| user3 |
| </source> |
| <p>In the above example we have defined three users <code>user1</code>, |
| <code>user2</code> and <code>user3</code>. |
| Now we would associate each unique user in the trace to the above users |
| defined in round-robin fashion. For example, if trace's users are |
| <code>tuser1</code>, <code>tuser2</code>, <code>tuser3</code>, |
| <code>tuser4</code> and <code>tuser5</code>, then the mappings would |
| be:</p> |
| <source> |
| tuser1 -> user1 |
| tuser2 -> user2 |
| tuser3 -> user3 |
| tuser4 -> user1 |
| tuser5 -> user2 |
| </source> |
| <p>For backward compatibility reasons, each line of users-list file can |
| contain username followed by groupnames in the form username[,group]*. |
| The groupnames will be ignored by Gridmix. |
| </p> |
| </section> |
| <section id="assumptions"> |
| <title>Simplifying Assumptions</title> |
| <p>GridMix will be developed in stages, incorporating feedback and |
| patches from the community. Currently its intent is to evaluate |
| MapReduce and HDFS performance and not the layers on top of them (i.e. |
| the extensive lib and sub-project space). Given these two limitations, |
| the following characteristics of job load are not currently captured in |
| job traces and cannot be accurately reproduced in GridMix:</p> |
| <ul> |
| <li><em>CPU Usage</em> - We have no data for per-task CPU usage, so we |
| cannot even attempt an approximation. GridMix tasks are never |
| CPU-bound independent of I/O, though this surely happens in |
| practice.</li> |
| <li><em>Filesystem Properties</em> - No attempt is made to match block |
| sizes, namespace hierarchies, or any property of input, intermediate |
| or output data other than the bytes/records consumed and emitted from |
| a given task. This implies that some of the most heavily-used parts of |
| the system - the compression libraries, text processing, streaming, |
| etc. - cannot be meaningfully tested with the current |
| implementation.</li> |
| <li><em>I/O Rates</em> - The rate at which records are |
| consumed/emitted is assumed to be limited only by the speed of the |
| reader/writer and constant throughout the task.</li> |
| <li><em>Memory Profile</em> - No data on tasks' memory usage over time |
| is available, though the max heap-size is retained.</li> |
| <li><em>Skew</em> - The records consumed and emitted to/from a given |
| task are assumed to follow observed averages, i.e. records will be |
| more regular than may be seen in the wild. Each map also generates |
| a proportional percentage of data for each reduce, so a job with |
| unbalanced input will be flattened.</li> |
| <li><em>Job Failure</em> - User code is assumed to be correct.</li> |
| <li><em>Job Independence</em> - The output or outcome of one job does |
| not affect when or whether a subsequent job will run.</li> |
| </ul> |
| </section> |
| <section id="appendix"> |
| <title>Appendix</title> |
| <p>Issues tracking the original implementations of <a |
| href="https://issues.apache.org/jira/browse/HADOOP-2369">GridMix1</a>, |
| <a href="https://issues.apache.org/jira/browse/HADOOP-3770">GridMix2</a>, |
| and <a |
| href="https://issues.apache.org/jira/browse/MAPREDUCE-776">GridMix3</a> |
| can be found on the Apache Hadoop MapReduce JIRA. Other issues tracking |
| the current development of GridMix can be found by searching <a |
| href="https://issues.apache.org/jira/browse/MAPREDUCE/component/12313086">the |
| Apache Hadoop MapReduce JIRA</a></p> |
| </section> |
| </body> |
| </document> |