[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16744537#comment-16744537 ] Aihua Xu commented on YARN-9116: Checked the UT failure but don't see explicit test case failures but the build indicates a timeout. [~cheersyang] Can you tell which one is causing the timeout from the log? Otherwise, I can attach a new patch to test out. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch, YARN-9116.2.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16744457#comment-16744457 ] Aihua Xu commented on YARN-9116: I will take a look at UT failure [~cheersyang] I was debating if I should make it backward compatible for that fairly new feature. I will add that. Another thing: in this patch, I just silently call {{maximumAllocation = Resources.componentwiseMin(queueMax, clusterMax);}} to get queue level maximum-allocation. Should we fail with some exception if queueMax > clusterMax and let the admin fix the configuration? That seems to be what it was. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch, YARN-9116.2.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16743648#comment-16743648 ] Weiwei Yang commented on YARN-9116: --- Hi [~aihuaxu] Thanks for the patch, the UT seems related, could you pls take a look? One concern of the patch is it replaces configuration "maximum-allocation-mb/maximum-allocation-vcores" with "maximum-allocation". This is an incompatible change. Can we make sure old configuration can still work for the compatibility? Thanks > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch, YARN-9116.2.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16742699#comment-16742699 ] Hadoop QA commented on YARN-9116: - | (x) *{color:red}-1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || | {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m 14s{color} | {color:blue} Docker mode activated. {color} | || || || || {color:brown} Prechecks {color} || | {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m 0s{color} | {color:green} The patch does not contain any @author tags. {color} | | {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m 0s{color} | {color:green} The patch appears to include 5 new or modified test files. {color} | || || || || {color:brown} trunk Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 38s{color} | {color:blue} Maven dependency ordering for branch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 19m 4s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 7m 59s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 1m 29s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m 4s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} shadedclient {color} | {color:green} 16m 28s{color} | {color:green} branch has no errors when building and testing our client artifacts. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 4m 57s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m 33s{color} | {color:green} trunk passed {color} | || || || || {color:brown} Patch Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 14s{color} | {color:blue} Maven dependency ordering for patch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m 27s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 7m 20s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javac {color} | {color:green} 7m 20s{color} | {color:green} the patch passed {color} | | {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange} 1m 28s{color} | {color:orange} hadoop-yarn-project/hadoop-yarn: The patch generated 34 new + 468 unchanged - 2 fixed = 502 total (was 470) {color} | | {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m 4s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m 0s{color} | {color:green} The patch has no whitespace issues. {color} | | {color:green}+1{color} | {color:green} shadedclient {color} | {color:green} 12m 39s{color} | {color:green} patch has no errors when building and testing our client artifacts. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m 32s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m 32s{color} | {color:green} the patch passed {color} | || || || || {color:brown} Other Tests {color} || | {color:green}+1{color} | {color:green} unit {color} | {color:green} 0m 51s{color} | {color:green} hadoop-yarn-api in the patch passed. {color} | | {color:green}+1{color} | {color:green} unit {color} | {color:green} 3m 43s{color} | {color:green} hadoop-yarn-common in the patch passed. {color} | | {color:red}-1{color} | {color:red} unit {color} | {color:red} 88m 58s{color} | {color:red} hadoop-yarn-server-resourcemanager in the patch failed. {color} | | {color:green}+1{color} | {color:green} unit {color} | {color:green} 0m 40s{color} | {color:green} hadoop-yarn-submarine in the patch passed. {color} | | {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m 30s{color} | {color:green} The patch does not generate ASF License warnings. {color} | | {color:black}{color} | {color:black} {color} | {color:black}184m 26s{color} | {color:black} {color} | \\ \\ || Subsystem || Report/Notes || | Docker | Client=17.05.0-ce Server=17.05.0-ce Image:yetus/hadoop:8f97d6f | | JIRA Issue | YARN-9116 | | JIRA Patch URL | https://issues.apache.org/jira/secure/attachment/12954897/YARN-9116.2.patch | | Optional Tests | dupname asflicense compile javac javadoc mvninstall mvnsite unit shadedclient findbugs checkstyle | | uname | Linux 3df1392bb8c0 4.4.0-139-generic #165-Ubuntu SMP Wed Oct 24 10:58:50 UTC 2018 x86_64 x86_64 x86_64
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16740998#comment-16740998 ] Hadoop QA commented on YARN-9116: - | (x) *{color:red}-1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || | {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m 22s{color} | {color:blue} Docker mode activated. {color} | || || || || {color:brown} Prechecks {color} || | {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m 0s{color} | {color:green} The patch does not contain any @author tags. {color} | | {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m 0s{color} | {color:green} The patch appears to include 5 new or modified test files. {color} | || || || || {color:brown} trunk Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 16s{color} | {color:blue} Maven dependency ordering for branch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 20m 16s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 8m 26s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 1m 36s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 3m 4s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} shadedclient {color} | {color:green} 17m 13s{color} | {color:green} branch has no errors when building and testing our client artifacts. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 4m 59s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m 24s{color} | {color:green} trunk passed {color} | || || || || {color:brown} Patch Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 14s{color} | {color:blue} Maven dependency ordering for patch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 2m 27s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 7m 34s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javac {color} | {color:green} 7m 34s{color} | {color:green} the patch passed {color} | | {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange} 1m 34s{color} | {color:orange} hadoop-yarn-project/hadoop-yarn: The patch generated 30 new + 467 unchanged - 2 fixed = 497 total (was 469) {color} | | {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m 53s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m 0s{color} | {color:green} The patch has no whitespace issues. {color} | | {color:green}+1{color} | {color:green} shadedclient {color} | {color:green} 13m 15s{color} | {color:green} patch has no errors when building and testing our client artifacts. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 5m 26s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 2m 19s{color} | {color:green} the patch passed {color} | || || || || {color:brown} Other Tests {color} || | {color:green}+1{color} | {color:green} unit {color} | {color:green} 0m 47s{color} | {color:green} hadoop-yarn-api in the patch passed. {color} | | {color:green}+1{color} | {color:green} unit {color} | {color:green} 3m 27s{color} | {color:green} hadoop-yarn-common in the patch passed. {color} | | {color:red}-1{color} | {color:red} unit {color} | {color:red} 92m 20s{color} | {color:red} hadoop-yarn-server-resourcemanager in the patch failed. {color} | | {color:green}+1{color} | {color:green} unit {color} | {color:green} 0m 43s{color} | {color:green} hadoop-yarn-submarine in the patch passed. {color} | | {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m 37s{color} | {color:green} The patch does not generate ASF License warnings. {color} | | {color:black}{color} | {color:black} {color} | {color:black}189m 51s{color} | {color:black} {color} | \\ \\ || Reason || Tests || | Failed junit tests | hadoop.yarn.server.resourcemanager.scheduler.capacity.TestContainerAllocation | | | hadoop.yarn.server.resourcemanager.scheduler.capacity.TestCapacityScheduler | \\ \\ || Subsystem || Report/Notes || | Docker | Client=17.05.0-ce Server=17.05.0-ce Image:yetus/hadoop:8f97d6f | | JIRA Issue | YARN-9116 | | JIRA Patch URL | https://issues.apache.org/jira/secure/attachment/12954664/YARN-9116.2.patch | | Optional Tests | dupname
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16739558#comment-16739558 ] Weiwei Yang commented on YARN-9116: --- Hi [~aihuaxu] {quote}Just to clarify the behavior: ... {quote} Exactly, thanks. {quote}I will work on the support of maximum-allocation-mb/vcores to parent queues first and have a follow up on the general resource type. {quote} Yes, it makes sense to track with separate tasks. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16738441#comment-16738441 ] Aihua Xu commented on YARN-9116: Thanks [~leftnoteasy] and [~cheersyang] for the suggestion. Just to clarify the behavior: the queue's maximum-allocation will not exceed the global setting (yarn.scheduler.capacity.maximum-allocation-mb) to keep the compatibility; among the queues, the child inherits the setting from the parent and the child queue can override the parent queue with larger or smaller setting but still respecting the global setting. That sounds reasonable to me. I will work on the support of maximum-allocation-mb/vcores to parent queues first and have a follow up on the general resource type. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16737856#comment-16737856 ] Wangda Tan commented on YARN-9116: -- Maybe we can support maximum-allocation-mb/vcores to parent queues first, and once YARN-9161 is done, we can file a separate Jira to support customer resource types of maximum-allocation. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16737855#comment-16737855 ] Wangda Tan commented on YARN-9116: -- [~cheersyang], [~aihuaxu], The proposal from Weiwei sounds good to me. We have several other fields in CS have similar inheritable behaviors. To me it is not an incompatible change since we don't expect user set maximum-allocation of parent queues. And also, the old config only support setting maximum allocation for mem and vcores, can we add a field to support customized resources like GPU/FPGA? I think we can refer to Absolute Resource Specification in Capacity Scheduler. See: [https://hadoop.apache.org/docs/r3.1.1/hadoop-yarn/hadoop-yarn-site/CapacityScheduler.html], search for "Resource Allocation using Absolute Resources configuration". I think we should either deprecated the old maximum-allocation-mb/vcores or at least make maximum-allocation can override maximum-allocation-mb/vcores. There's an on-going work of YARN-9161 to add support of Absolute Resource for customized resource types to Capacity Scheduler. We should make sure consistent format is being used. Thoughts? > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16737807#comment-16737807 ] Weiwei Yang commented on YARN-9116: --- Hi [~aihuaxu], [~tgraves] Absolutely we should always respect the cluster level maximum, all queue-level max should be smaller than that. What I was suggesting is to make a leaf queue can inherit queue-level max from its parent. IMO this model is more flexible. Please see the example here. {quote}As I understand currently you can only set maximum-allocation-mb on the leaf queue, not intermediate parent queues. {quote} That's the case for now, I was suggesting to improve this be inheritable. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16737392#comment-16737392 ] Thomas Graves commented on YARN-9116: - Yes so you want to keep the behavior that the cluster level maximum is the absolute maximum and no child queues can be larger then that, otherwise it breaks backwards compatibility. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16737351#comment-16737351 ] Aihua Xu commented on YARN-9116: [~cheersyang] As I understand currently you can only set maximum-allocation-mb on the leaf queue, not intermediate parent queues. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16736552#comment-16736552 ] Weiwei Yang commented on YARN-9116: --- Hi [~aihuaxu] {quote}we are still introducing the incompatibility change since the root is set to 16G {{yarn.scheduler.capacity.root.maximum-allocation-mb=16G}} while the child queue root.large is set to 80G (larger value). {quote} Is that a different behavior than today? If I remember correctly, this is same as today which sets queue max separately. I have listed the incompatibility in my last comment, {quote}this way might introduce a different behavior because the max-allocation user sets for a queue will be passing to all its child. {quote} that's the thing I am concerned with. {quote}What you are proposing is: the child can override the parent's value (larger or smaller) but won't exceed the global value, correct? {quote} Yes, this is my suggestion. But it would be nice to hear from [~tgraves] as he initiated added this, also from [~leftnoteasy]. Thanks > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16736083#comment-16736083 ] Aihua Xu commented on YARN-9116: [~cheersyang] In the example above, we are still introducing the incompatibility change since the root is set to 16G {{yarn.scheduler.capacity.root.maximum-allocation-mb=16G}} while the child queue root.large is set to 80G (larger value). Do you think it's OK change? What you are proposing is: the child can override the parent's value (larger or smaller) but won't exceed the global value, correct? > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16734918#comment-16734918 ] Weiwei Yang commented on YARN-9116: --- Hi [~aihuaxu] Thanks for pointing me to YARN-9055. But what I meant was not exactly the same, we should still honor the cluster max as before. Let me add the cluster max into this example {code:java} yarn.scheduler.maximum-allocation-mb=120G yarn.scheduler.capacity.root.maximum-allocation-mb=16G yarn.scheduler.capacity.root.small.maximum-allocation-mb=8G yarn.scheduler.capacity.root.large.maximum-allocation-mb=80G {code} then the hierarchy looks like {code:java} // global (120G) -- root (16G) a (16G) b (16G) c (16G) small (8G) large (80G) {code} this is just like the idea you are proposing in this idea. But without adding a new property for queue default, I am suggesting to define it as the top level queue max. Then all sub queues can inherent this value. At the same time, any queue is allowed to override this value, to be bigger (e.g root.large) or smaller (e.g root.small) values. For the compatibility, this way might introduce a different behavior because the max-allocation user sets for a queue will be passing to all its child. So for example if user only sets a max for {{root.a}} on the cluster like following {code:java} // global (120G) -- root (120G) root.a (16G) root.a.b (120G) {code} then after this change, the behavior changes to {code:java} // global (120G) -- root (120G) root.a (16G) root.a.b (16G) // this is changed.. {code} but personally I think this is better. Maybe we can list as incompatible change and giving user a note. Adding a property is OK, but two main counter reason: 1) conf is already too complex; 2) still not flex enough. Thanks > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16734369#comment-16734369 ] Aihua Xu commented on YARN-9116: [~cheersyang] That was my initial idea (see YARN-9055) that we can override the parent setting, but it introduces incompatibility since it's always assumed that the child queue can't have larger settings than the parents. Some clients such as spark will check the top settings and fail immediately if the resource request can't be satisfied. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16733952#comment-16733952 ] Weiwei Yang commented on YARN-9116: --- Hi [~aihuaxu] {quote}Actually that would introduce many queue level configuration if we don't introduce new property even with such inheritance. Even after we implement inheritance mechanism, we have to set the global to be 120G/256vCores (the maximum value allowed in the cluster) and then override all the top queues to be 16G/16vCores and set the larger container top queue to 120G/256vCores. {quote} I was thinking if we can setup something like following (use memory as example): {code:java} yarn.scheduler.capacity.root.maximum-allocation-mb=16G yarn.scheduler.capacity.root.large.maximum-allocation-mb=120G {code} so if the queue structure is like {code:java} -- root (16G) a (16G) b (16G) c (16G) large (120G) {code} This implicits the queue level max-allocation is inherited from its parent and it can be overwritten at the same time. Would this work? Thanks > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16732283#comment-16732283 ] Aihua Xu commented on YARN-9116: Thanks [~cheersyang] for the comment. Happy new year. So you are suggesting the following, is that correct? Actually that would introduce many queue level configuration if we don't introduce new property even with such inheritance. Even after we implement inheritance mechanism, we have to set the global to be 120G/256vCores (the maximum value allowed in the cluster) and then override all the top queues to be 16G/16vCores and set the larger container top queue to 120G/256vCores. I feel the current approach is simpler and straightforward. Let me know if you think the inheritance implementation is still needed, but seems we do need to add additional configuration. {noformat} Queue level max inherits the value from its parent if it is not explicitly set If queue level max is set explicitly, then it is honored without considering its parents {noformat} > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16727775#comment-16727775 ] Weiwei Yang commented on YARN-9116: --- Hi [~aihuaxu] Thanks for the patch. I am +0 on adding this new property, as I previously suggested. [~leftnoteasy], are you OK with adding this? For the patch you uploaded, some comments # CapacitySchedulerConfiguration: line 32: do not use wild char "*" in the import # Can we move {{createResourceFromString}}, {{parseResourcesString}}, \{{validateResourceTypes}} to \{{ResourceUtils}} class? Which is more appropriate for the utility methods. # Do we allow user to set more than memory/cpu in this property? E.g \{{memory=1024,vcore=1,gpu=1}}. Follow the code in \{{CapacitySchedulerConfiguration#getMaximumAllocationPerQueue}}, it only accepts memory/cpu which seems would cause \{{validateResourceTypes}} to fail. # If #3 is not supported, we need an appropriate check and a test case to cover this scenario. Thanks > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > Attachments: YARN-9116.1.patch > > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16725560#comment-16725560 ] Hadoop QA commented on YARN-9116: - | (x) *{color:red}-1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || | {color:blue}0{color} | {color:blue} reexec {color} | {color:blue} 0m 13s{color} | {color:blue} Docker mode activated. {color} | || || || || {color:brown} Prechecks {color} || | {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m 0s{color} | {color:green} The patch does not contain any @author tags. {color} | | {color:green}+1{color} | {color:green} test4tests {color} | {color:green} 0m 0s{color} | {color:green} The patch appears to include 2 new or modified test files. {color} | || || || || {color:brown} trunk Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 35s{color} | {color:blue} Maven dependency ordering for branch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 18m 32s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 7m 45s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 1m 13s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 2m 1s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} shadedclient {color} | {color:green} 13m 48s{color} | {color:green} branch has no errors when building and testing our client artifacts. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 3m 12s{color} | {color:green} trunk passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 33s{color} | {color:green} trunk passed {color} | || || || || {color:brown} Patch Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 12s{color} | {color:blue} Maven dependency ordering for patch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m 44s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 7m 8s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javac {color} | {color:green} 7m 8s{color} | {color:green} the patch passed {color} | | {color:orange}-0{color} | {color:orange} checkstyle {color} | {color:orange} 1m 12s{color} | {color:orange} hadoop-yarn-project/hadoop-yarn: The patch generated 10 new + 327 unchanged - 1 fixed = 337 total (was 328) {color} | | {color:green}+1{color} | {color:green} mvnsite {color} | {color:green} 1m 57s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m 0s{color} | {color:green} The patch has no whitespace issues. {color} | | {color:green}+1{color} | {color:green} shadedclient {color} | {color:green} 11m 24s{color} | {color:green} patch has no errors when building and testing our client artifacts. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 3m 57s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 56s{color} | {color:green} the patch passed {color} | || || || || {color:brown} Other Tests {color} || | {color:green}+1{color} | {color:green} unit {color} | {color:green} 3m 42s{color} | {color:green} hadoop-yarn-common in the patch passed. {color} | | {color:red}-1{color} | {color:red} unit {color} | {color:red} 90m 1s{color} | {color:red} hadoop-yarn-server-resourcemanager in the patch failed. {color} | | {color:green}+1{color} | {color:green} unit {color} | {color:green} 0m 49s{color} | {color:green} hadoop-yarn-submarine in the patch passed. {color} | | {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m 40s{color} | {color:green} The patch does not generate ASF License warnings. {color} | | {color:black}{color} | {color:black} {color} | {color:black}172m 34s{color} | {color:black} {color} | \\ \\ || Subsystem || Report/Notes || | Docker | Client=17.05.0-ce Server=17.05.0-ce Image:yetus/hadoop:8f97d6f | | JIRA Issue | YARN-9116 | | JIRA Patch URL | https://issues.apache.org/jira/secure/attachment/12952434/YARN-9116.1.patch | | Optional Tests | dupname asflicense compile javac javadoc mvninstall mvnsite unit shadedclient findbugs checkstyle | | uname | Linux 7d0bf6844083 4.4.0-138-generic #164-Ubuntu SMP Tue Oct 2 17:16:02 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux | | Build tool | maven | | Personality | /testptch/patchprocess/precommit/personality/provided.sh | | git revision | trunk / e815fd9 | |
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16721603#comment-16721603 ] Aihua Xu commented on YARN-9116: [~cheersyang] Yes. Agree with you. I will implement toward such goal. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16721066#comment-16721066 ] Weiwei Yang commented on YARN-9116: --- The logic to inherit should not be a problem. I mean we could still hornor global value if it is set (the old logic not changed). If global value is not set, then we just enforce the check with the queue level configs, and that allows a leaf queue has its own max. Would that make our configuration simpler? I have no problem to add a new config if we run out of options, given we already have a lot of configurations. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16720919#comment-16720919 ] Aihua Xu commented on YARN-9116: [~cheersyang] As I understand from the code, the logic to inherit from the parent is not implemented yet. We need to set such properties on the leaf queues. That is another approach I'm also thinking of but need more work. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16720845#comment-16720845 ] Weiwei Yang commented on YARN-9116: --- Hi [~aihuaxu] I don't see why this needs 2 options. If global max is not set, just to configure queue max, with following convention # Queue level max inherits the value from its parent if it is not explicitly set # If queue level max is set explicitly, then it is honored without considering its parents For the case you mentioned, # Configure root queue max to 16G/16vcore # Configure larger container queue to 120G/256vcore would that work? > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16720436#comment-16720436 ] Aihua Xu commented on YARN-9116: Thanks [~leftnoteasy]. I will look into what you mentioned above. Sounds like a good implementation to handle different resource types. Originally I was only thinking of memory and vCores. [~cheersyang] What I'm trying to achieve is to configure a larger container queue. To my understanding from the implementation of YARN-1582, we have to do the following steps: # Configure the global maximum-allocation to 120G/256vCores # Configure regular queues to 16G/16vCores or desired values # Configure larger container queue to 120G/256vCores The default queue-default I'm talking about is just to set to 16G/16vCores in this case. Without such default value, you have to set for all the queues. This is just the default value and you can set the desired one if the queue need a different value. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16720289#comment-16720289 ] Weiwei Yang commented on YARN-9116: --- Hi [~aihuaxu] I see why you are proposing this. But when you have a lot of queues, how you come up with a reasonable queue-default-max, given each queue has different capacity that serves different sizes of apps. If I understand correctly, you want to use queue-default-max to simplify the configuration, but I think it's a bit tricky to set such a value for all queues. > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
[jira] [Commented] (YARN-9116) Capacity Scheduler: add the default maximum-allocation-mb and maximum-allocation-vcores for the queues
[ https://issues.apache.org/jira/browse/YARN-9116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16719676#comment-16719676 ] Wangda Tan commented on YARN-9116: -- [~aihuaxu], This sounds like a plan, but existing maximum memory, etc. is defined inside yarn-site.xml. And given we will have multiple resource types support like GPU. I suggest to make the default maximum capacity definition use absolute values for all types. The name of the config could be: {code:java} yarn.scheduler.capacity.default-queue-maximum-capacity memory=20G,vcores=20,gpu=3 {code} You can reuse implementation org.apache.hadoop.yarn.submarine.client.cli.CliUtils#createResourceFromString from trunk (and move it to common place if you think it is required). Thoughts? + [~sunilg], [~cheersyang] > Capacity Scheduler: add the default maximum-allocation-mb and > maximum-allocation-vcores for the queues > -- > > Key: YARN-9116 > URL: https://issues.apache.org/jira/browse/YARN-9116 > Project: Hadoop YARN > Issue Type: Sub-task > Components: capacity scheduler >Affects Versions: 2.7.0 >Reporter: Aihua Xu >Assignee: Aihua Xu >Priority: Major > > YARN-1582 adds the support of maximum-allocation-mb configuration per queue > which is targeting to support larger container features on dedicated queues > (larger maximum-allocation-mb/maximum-allocation-vcores for such queue) . > While to achieve larger container configuration, we need to increase the > global maximum-allocation-mb/maximum-allocation-vcores (e.g. 120G/256) and > then override those configurations with desired values on the queues since > queue configuration can't be larger than cluster configuration. There are > many queues in the system and if we forget to configure such values when > adding a new queue, then such queue gets default 120G/256 which typically is > not what we want. > We can come up with a queue-default configuration (set to normal queue > configuration like 16G/8), so the leaf queues gets such values by default. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org