[
https://issues.apache.org/jira/browse/IMPALA-8469?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16830782#comment-16830782
]
ASF subversion and git services commented on IMPALA-8469:
---------------------------------------------------------
Commit d820952d86d34ba887c55a09e58b735cbef866c2 in impala's branch
refs/heads/master from Tim Armstrong
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=d820952 ]
IMPALA-8469: admit_mem_limit for dedicated coordinator
Refactored to avoid the code duplication that resulted in this bug:
* admit_mem_limit is calculated once in ExecEnv
* The local backend descriptor is always constructed with
a static helper: Scheduler::BuildLocalBackendDescriptor()
I chose to factor it in this way, in part, to avoid invasive
changes to scheduler-test, which currently doesn't depend on
ExecEnv or ImpalaServer.
Testing:
Added basic test that reproduces the bug.
Change-Id: Iaceb21b753b9b021bedc4187c0d44aaa6a626521
Reviewed-on: http://gerrit.cloudera.org:8080/13180
Reviewed-by: Tim Armstrong <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
> Admit memory not set in backend descriptor for coordinator-only nodes
> ---------------------------------------------------------------------
>
> Key: IMPALA-8469
> URL: https://issues.apache.org/jira/browse/IMPALA-8469
> Project: IMPALA
> Issue Type: Bug
> Components: Backend
> Affects Versions: Impala 3.2.0
> Reporter: Ian Buss
> Assignee: Tim Armstrong
> Priority: Critical
> Labels: admission-control, resource-management
>
> When configuring admission control with dedicated coordinator daemons,
> queries in pools with memory limits fail with the admission rejections like
> the following:
> {noformat}
> Rejected query from pool root.default: request memory needed 3.00 GB per node
> is greater than memory available for admission 0 of coord1.example.com:22000.
> Use the MEM_LIMIT query option to indicate how much memory is required per
> node.{noformat}
> Tracing this in the code leads us to line 576 of {{admission-controller.cc}}
> and therefore to suspect that the local {{TBackendDescriptor}}
> ({{local_backend_descriptor_}}) for the coordinator node in {{scheduler.cc}}
> never has {{admit_mem_limit}} set, and thus ends up with the default value of
> 0.
> The issue goes away if NO_SPECIALIZATION is used instead of COORDINATOR_ONLY.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
