[jira] [Commented] (YARN-802) APPLICATION_INIT is never sent to AuxServices other than the builtin ShuffleHandler
[ https://issues.apache.org/jira/browse/YARN-802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13685110#comment-13685110 ] Siddharth Seth commented on YARN-802: - bq. If I understand correctly, in order to simultaneously support multiple jobs of multiple users that each can contact different Shuffle provider we must have all providers in the air in parallel. Multiple providers can be run by the NodeManager in parallel. An application chooses which provider(s) it wants to use when it starts a container on a NodeManager. bq. This data for this map arrive from the APPLICATION_INIT event. Hence, all AuxServices that serve as ShuffleProviders need to get APPLICATION_INIT events. The data in the APPLICAITON_INIT event is from the startContainer request (the serviceData in the ContainerLaunchConetxt). If the application wants the INIT event to go to multiple providers, it can set the service data accordingly. The MapReduce AM hardcodes this to the default SHUFFLE_PROVIDER which is why only that one gets the init event. There may be auxillary services which are not responsible for shuffle, or are in general incompatible with the shuffle consumer configured by the job. I don't think they need to get an INIT event. > APPLICATION_INIT is never sent to AuxServices other than the builtin > ShuffleHandler > --- > > Key: YARN-802 > URL: https://issues.apache.org/jira/browse/YARN-802 > Project: Hadoop YARN > Issue Type: Bug > Components: applications, nodemanager >Affects Versions: 2.0.4-alpha >Reporter: Avner BenHanoch > > APPLICATION_INIT is never sent to AuxServices other than the built-in > ShuffleHandler. This means that 3rd party ShuffleProvider(s) will not be > able to function, because APPLICATION_INIT enables the AuxiliaryService to > map jobId->userId. This is needed for properly finding the MOFs of a job per > reducers' requests. > NOTE: The built-in ShuffleHandler does get APPLICATION_INIT events due to > hard-coded expression in hadoop code. The current TaskAttemptImpl.java code > explicitly call: serviceData.put (ShuffleHandler.MAPREDUCE_SHUFFLE_SERVICEID, > ...) and ignores any additional AuxiliaryService. As a result, only the > built-in ShuffleHandler will get APPLICATION_INIT events. Any 3rd party > AuxillaryService will never get APPLICATION_INIT events. > I think a solution can be in one of two ways: > 1. Change TaskAttemptImpl.java to loop on all Auxiliary Services and register > each of them, by calling serviceData.put (…) in loop. > 2. Change AuxServices.java similar to the fix in: MAPREDUCE-2668 > "APPLICATION_STOP is never sent to AuxServices". This means that in case the > 'handle' method gets APPLICATION_INIT event it will demultiplex it to all Aux > Services regardless of the value in event.getServiceID(). > I prefer the 2nd solution. I am welcoming any ideas. I can provide the > needed patch for any option that people like. > See [Pluggable Shuffle in Hadoop > documentation|http://hadoop.apache.org/docs/current/hadoop-mapreduce-client/hadoop-mapreduce-client-core/PluggableShuffleAndPluggableSort.html] -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (YARN-802) APPLICATION_INIT is never sent to AuxServices other than the builtin ShuffleHandler
[ https://issues.apache.org/jira/browse/YARN-802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13684624#comment-13684624 ] Avner BenHanoch commented on YARN-802: -- Hi Siddharth In addition to what I wrote, I just noticed that perhaps we have a misunderstanding. See "Implementing a Custom Shuffle" from [Hadoop documentation about Pluggable Shuffle|http://hadoop.apache.org/docs/current/hadoop-mapreduce-client/hadoop-mapreduce-client-core/PluggableShuffleAndPluggableSort.html] {quote} A custom shuffle implementation requires a org.apache.hadoop.yarn.server.nodemanager.containermanager.AuxServices.AuxiliaryServiceimplementation class running in the NodeManagers and a org.apache.hadoop.mapred.ShuffleConsumerPlugin implementation class running in the Reducer tasks. The default implementations provided by Hadoop can be used as references: • org.apache.hadoop.mapred.ShuffleHandler • org.apache.hadoop.mapreduce.task.reduce.Shuffle {quote} In this issue we are talking about the *provider side of the Shuffle*, we want to have additional ShuffleProvider(s) in addition to the default ShuffleHandler. All shuffle providers are AuxServices running in the NodeManager. I see them as daemons like ShuffleHandler, they are not applications or jobs. If I understand correctly, in order to simultaneously support multiple jobs of multiple users that each can contact different Shuffle provider we must have all providers in the air in parallel. With this, the ShuffleConsumer in each reducer will be able to request MOFs from its desired provider. However, the provider must know where in the disks are the MOFs of a particular job. The way to know that (see ShuffleHandler) is based on the mapping: jobId -> userId. This data for this map arrive from the APPLICATION_INIT event. Hence, all AuxServices that serve as ShuffleProviders need to get APPLICATION_INIT events. > APPLICATION_INIT is never sent to AuxServices other than the builtin > ShuffleHandler > --- > > Key: YARN-802 > URL: https://issues.apache.org/jira/browse/YARN-802 > Project: Hadoop YARN > Issue Type: Bug > Components: applications, nodemanager >Affects Versions: 2.0.4-alpha >Reporter: Avner BenHanoch > > APPLICATION_INIT is never sent to AuxServices other than the built-in > ShuffleHandler. This means that 3rd party ShuffleProvider(s) will not be > able to function, because APPLICATION_INIT enables the AuxiliaryService to > map jobId->userId. This is needed for properly finding the MOFs of a job per > reducers' requests. > NOTE: The built-in ShuffleHandler does get APPLICATION_INIT events due to > hard-coded expression in hadoop code. The current TaskAttemptImpl.java code > explicitly call: serviceData.put (ShuffleHandler.MAPREDUCE_SHUFFLE_SERVICEID, > ...) and ignores any additional AuxiliaryService. As a result, only the > built-in ShuffleHandler will get APPLICATION_INIT events. Any 3rd party > AuxillaryService will never get APPLICATION_INIT events. > I think a solution can be in one of two ways: > 1. Change TaskAttemptImpl.java to loop on all Auxiliary Services and register > each of them, by calling serviceData.put (…) in loop. > 2. Change AuxServices.java similar to the fix in: MAPREDUCE-2668 > "APPLICATION_STOP is never sent to AuxServices". This means that in case the > 'handle' method gets APPLICATION_INIT event it will demultiplex it to all Aux > Services regardless of the value in event.getServiceID(). > I prefer the 2nd solution. I am welcoming any ideas. I can provide the > needed patch for any option that people like. > See about Pluggable Shuffle in this URL: > http://hadoop.apache.org/docs/current/hadoop-mapreduce-client/hadoop-mapreduce-client-core/PluggableShuffleAndPluggableSort.html -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (YARN-802) APPLICATION_INIT is never sent to AuxServices other than the builtin ShuffleHandler
[ https://issues.apache.org/jira/browse/YARN-802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13684258#comment-13684258 ] Avner BenHanoch commented on YARN-802: -- Thanks for the explanation about YARN. Still this is not enough, for two reasons: 1. It is true that *usually* "A shuffle consumer instance will only contact one of the shuffle providers". Still, as written in the quote I pasted , "it should be possible to fallback to another shuffle on the fly". This means that one consumer can load another consumer and serve as proxy to the real consumer that will contact another provider. 2. In a single job there are multiple reducers each with its own shuffle consumer instance; hence, we have *multiple shuffle consumers per job*. It should be possible for each consumer to choose its preffered provider based on memory/network/... condition on its machine regardless of other consumers in the same job. > APPLICATION_INIT is never sent to AuxServices other than the builtin > ShuffleHandler > --- > > Key: YARN-802 > URL: https://issues.apache.org/jira/browse/YARN-802 > Project: Hadoop YARN > Issue Type: Bug > Components: applications, nodemanager >Affects Versions: 2.0.4-alpha >Reporter: Avner BenHanoch > > APPLICATION_INIT is never sent to AuxServices other than the built-in > ShuffleHandler. This means that 3rd party ShuffleProvider(s) will not be > able to function, because APPLICATION_INIT enables the AuxiliaryService to > map jobId->userId. This is needed for properly finding the MOFs of a job per > reducers' requests. > NOTE: The built-in ShuffleHandler does get APPLICATION_INIT events due to > hard-coded expression in hadoop code. The current TaskAttemptImpl.java code > explicitly call: serviceData.put (ShuffleHandler.MAPREDUCE_SHUFFLE_SERVICEID, > ...) and ignores any additional AuxiliaryService. As a result, only the > built-in ShuffleHandler will get APPLICATION_INIT events. Any 3rd party > AuxillaryService will never get APPLICATION_INIT events. > I think a solution can be in one of two ways: > 1. Change TaskAttemptImpl.java to loop on all Auxiliary Services and register > each of them, by calling serviceData.put (…) in loop. > 2. Change AuxServices.java similar to the fix in: MAPREDUCE-2668 > "APPLICATION_STOP is never sent to AuxServices". This means that in case the > 'handle' method gets APPLICATION_INIT event it will demultiplex it to all Aux > Services regardless of the value in event.getServiceID(). > I prefer the 2nd solution. I am welcoming any ideas. I can provide the > needed patch for any option that people like. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (YARN-802) APPLICATION_INIT is never sent to AuxServices other than the builtin ShuffleHandler
[ https://issues.apache.org/jira/browse/YARN-802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13683579#comment-13683579 ] Siddharth Seth commented on YARN-802: - With YARN, a new AM (Application) is started per job. The initApp in the NM is per app - so each job/app can choose which shuffle provider it wants to use. The shuffle service configured for an AM will be specific to a single job only. >From MAPREDUCE-4049 bq. A shuffle consumer instance will only contact one of the shuffle providers and will request its desired files only from from this provider. I'm assuming a single job will only use one shuffle provider - or do you see a situation where multiple shuffle providers can serve data to a single job ? In case of multiple jobs being run by a single AM - this gets more complicated, and we may need to initialize multiple providers. > APPLICATION_INIT is never sent to AuxServices other than the builtin > ShuffleHandler > --- > > Key: YARN-802 > URL: https://issues.apache.org/jira/browse/YARN-802 > Project: Hadoop YARN > Issue Type: Bug > Components: applications, nodemanager >Affects Versions: 2.0.4-alpha >Reporter: Avner BenHanoch > > APPLICATION_INIT is never sent to AuxServices other than the built-in > ShuffleHandler. This means that 3rd party ShuffleProvider(s) will not be > able to function, because APPLICATION_INIT enables the AuxiliaryService to > map jobId->userId. This is needed for properly finding the MOFs of a job per > reducers' requests. > NOTE: The built-in ShuffleHandler does get APPLICATION_INIT events due to > hard-coded expression in hadoop code. The current TaskAttemptImpl.java code > explicitly call: serviceData.put (ShuffleHandler.MAPREDUCE_SHUFFLE_SERVICEID, > ...) and ignores any additional AuxiliaryService. As a result, only the > built-in ShuffleHandler will get APPLICATION_INIT events. Any 3rd party > AuxillaryService will never get APPLICATION_INIT events. > I think a solution can be in one of two ways: > 1. Change TaskAttemptImpl.java to loop on all Auxiliary Services and register > each of them, by calling serviceData.put (…) in loop. > 2. Change AuxServices.java similar to the fix in: MAPREDUCE-2668 > "APPLICATION_STOP is never sent to AuxServices". This means that in case the > 'handle' method gets APPLICATION_INIT event it will demultiplex it to all Aux > Services regardless of the value in event.getServiceID(). > I prefer the 2nd solution. I am welcoming any ideas. I can provide the > needed patch for any option that people like. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (YARN-802) APPLICATION_INIT is never sent to AuxServices other than the builtin ShuffleHandler
[ https://issues.apache.org/jira/browse/YARN-802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13683102#comment-13683102 ] Avner BenHanoch commented on YARN-802: -- This idea will force us to restart mapred daemons any time we want to switch shuffle service and will limit us to having just 1 ShuffleProvider at a time. I already discussed this idea in MAPREDUCE-4049 (look for the word multiple there). I am pasting here one paragraph from that discussion: {quote} It could be that a ShuffleConsumerX will be ideal for jobs of one type, while ShuffleConsumerY will be ideal for jobs of other type (for example Grep vs. TeraSort). Hence, multiple Shuffle-Consumer plugins may run in parallel in multiple jobs. Each of the consumers will contact its desired shuffle provider. Hence, all providers should be available in parallel (also, one shuffle service can be sensitive to type of network problems that doesn't disturb other shuffle services, hence, it should be possible to fallback to another shuffle on the fly). {quote} > APPLICATION_INIT is never sent to AuxServices other than the builtin > ShuffleHandler > --- > > Key: YARN-802 > URL: https://issues.apache.org/jira/browse/YARN-802 > Project: Hadoop YARN > Issue Type: Bug > Components: applications, nodemanager >Affects Versions: 2.0.4-alpha >Reporter: Avner BenHanoch > > APPLICATION_INIT is never sent to AuxServices other than the built-in > ShuffleHandler. This means that 3rd party ShuffleProvider(s) will not be > able to function, because APPLICATION_INIT enables the AuxiliaryService to > map jobId->userId. This is needed for properly finding the MOFs of a job per > reducers' requests. > NOTE: The built-in ShuffleHandler does get APPLICATION_INIT events due to > hard-coded expression in hadoop code. The current TaskAttemptImpl.java code > explicitly call: serviceData.put (ShuffleHandler.MAPREDUCE_SHUFFLE_SERVICEID, > ...) and ignores any additional AuxiliaryService. As a result, only the > built-in ShuffleHandler will get APPLICATION_INIT events. Any 3rd party > AuxillaryService will never get APPLICATION_INIT events. > I think a solution can be in one of two ways: > 1. Change TaskAttemptImpl.java to loop on all Auxiliary Services and register > each of them, by calling serviceData.put (…) in loop. > 2. Change AuxServices.java similar to the fix in: MAPREDUCE-2668 > "APPLICATION_STOP is never sent to AuxServices". This means that in case the > 'handle' method gets APPLICATION_INIT event it will demultiplex it to all Aux > Services regardless of the value in event.getServiceID(). > I prefer the 2nd solution. I am welcoming any ideas. I can provide the > needed patch for any option that people like. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (YARN-802) APPLICATION_INIT is never sent to AuxServices other than the builtin ShuffleHandler
[ https://issues.apache.org/jira/browse/YARN-802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13682724#comment-13682724 ] Siddharth Seth commented on YARN-802: - It should be possible to configure the MR AM with the shuffle service that needs to be used, in which case the MR AM sets up the service id correctly (in TaskAttemptImpl), and the NodeManager can send the init event to the correct service. We should probably change the stop to behave the same way. > APPLICATION_INIT is never sent to AuxServices other than the builtin > ShuffleHandler > --- > > Key: YARN-802 > URL: https://issues.apache.org/jira/browse/YARN-802 > Project: Hadoop YARN > Issue Type: Bug > Components: applications, nodemanager >Affects Versions: 2.0.4-alpha >Reporter: Avner BenHanoch > > APPLICATION_INIT is never sent to AuxServices other than the built-in > ShuffleHandler. This means that 3rd party ShuffleProvider(s) will not be > able to function, because APPLICATION_INIT enables the AuxiliaryService to > map jobId->userId. This is needed for properly finding the MOFs of a job per > reducers' requests. > NOTE: The built-in ShuffleHandler does get APPLICATION_INIT events due to > hard-coded expression in hadoop code. The current TaskAttemptImpl.java code > explicitly call: serviceData.put (ShuffleHandler.MAPREDUCE_SHUFFLE_SERVICEID, > ...) and ignores any additional AuxiliaryService. As a result, only the > built-in ShuffleHandler will get APPLICATION_INIT events. Any 3rd party > AuxillaryService will never get APPLICATION_INIT events. > I think a solution can be in one of two ways: > 1. Change TaskAttemptImpl.java to loop on all Auxiliary Services and register > each of them, by calling serviceData.put (…) in loop. > 2. Change AuxServices.java similar to the fix in: MAPREDUCE-2668 > "APPLICATION_STOP is never sent to AuxServices". This means that in case the > 'handle' method gets APPLICATION_INIT event it will demultiplex it to all Aux > Services regardless of the value in event.getServiceID(). > I prefer the 2nd solution. I am welcoming any ideas. I can provide the > needed patch for any option that people like. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (YARN-802) APPLICATION_INIT is never sent to AuxServices other than the builtin ShuffleHandler
[ https://issues.apache.org/jira/browse/YARN-802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13682539#comment-13682539 ] Avner BenHanoch commented on YARN-802: -- Hi Siddharth, I am not sure I understand the question. Do you suggest that we'll have a configuration with sub-list of the AuxServices and only members of the new list will get the APPLICATION_INIT event? - This is possible, however, it will not match the current behavior with APPLICATION_STOP event, since the last event is being sent to ALL AuxServices. Please elaborate. > APPLICATION_INIT is never sent to AuxServices other than the builtin > ShuffleHandler > --- > > Key: YARN-802 > URL: https://issues.apache.org/jira/browse/YARN-802 > Project: Hadoop YARN > Issue Type: Bug > Components: applications, nodemanager >Affects Versions: 2.0.4-alpha >Reporter: Avner BenHanoch > > APPLICATION_INIT is never sent to AuxServices other than the built-in > ShuffleHandler. This means that 3rd party ShuffleProvider(s) will not be > able to function, because APPLICATION_INIT enables the AuxiliaryService to > map jobId->userId. This is needed for properly finding the MOFs of a job per > reducers' requests. > NOTE: The built-in ShuffleHandler does get APPLICATION_INIT events due to > hard-coded expression in hadoop code. The current TaskAttemptImpl.java code > explicitly call: serviceData.put (ShuffleHandler.MAPREDUCE_SHUFFLE_SERVICEID, > ...) and ignores any additional AuxiliaryService. As a result, only the > built-in ShuffleHandler will get APPLICATION_INIT events. Any 3rd party > AuxillaryService will never get APPLICATION_INIT events. > I think a solution can be in one of two ways: > 1. Change TaskAttemptImpl.java to loop on all Auxiliary Services and register > each of them, by calling serviceData.put (…) in loop. > 2. Change AuxServices.java similar to the fix in: MAPREDUCE-2668 > "APPLICATION_STOP is never sent to AuxServices". This means that in case the > 'handle' method gets APPLICATION_INIT event it will demultiplex it to all Aux > Services regardless of the value in event.getServiceID(). > I prefer the 2nd solution. I am welcoming any ideas. I can provide the > needed patch for any option that people like. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (YARN-802) APPLICATION_INIT is never sent to AuxServices other than the builtin ShuffleHandler
[ https://issues.apache.org/jira/browse/YARN-802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13682488#comment-13682488 ] Siddharth Seth commented on YARN-802: - Can the MR AM specify the service to be used via configuration, and set the service data accordingly. > APPLICATION_INIT is never sent to AuxServices other than the builtin > ShuffleHandler > --- > > Key: YARN-802 > URL: https://issues.apache.org/jira/browse/YARN-802 > Project: Hadoop YARN > Issue Type: Bug > Components: applications, nodemanager >Affects Versions: 2.0.4-alpha >Reporter: Avner BenHanoch > > APPLICATION_INIT is never sent to AuxServices other than the built-in > ShuffleHandler. This means that 3rd party ShuffleProvider(s) will not be > able to function, because APPLICATION_INIT enables the AuxiliaryService to > map jobId->userId. This is needed for properly finding the MOFs of a job per > reducers' requests. > NOTE: The built-in ShuffleHandler does get APPLICATION_INIT events due to > hard-coded expression in hadoop code. The current TaskAttemptImpl.java code > explicitly call: serviceData.put (ShuffleHandler.MAPREDUCE_SHUFFLE_SERVICEID, > ...) and ignores any additional AuxiliaryService. As a result, only the > built-in ShuffleHandler will get APPLICATION_INIT events. Any 3rd party > AuxillaryService will never get APPLICATION_INIT events. > I think a solution can be in one of two ways: > 1. Change TaskAttemptImpl.java to loop on all Auxiliary Services and register > each of them, by calling serviceData.put (…) in loop. > 2. Change AuxServices.java similar to the fix in: MAPREDUCE-2668 > "APPLICATION_STOP is never sent to AuxServices". This means that in case the > 'handle' method gets APPLICATION_INIT event it will demultiplex it to all Aux > Services regardless of the value in event.getServiceID(). > I prefer the 2nd solution. I am welcoming any ideas. I can provide the > needed patch for any option that people like. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira