You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Vadim Opolski (JIRA)" <ji...@apache.org> on 2017/09/04 14:52:02 UTC

[jira] [Comment Edited] (IGNITE-4756) Print info about partition distribution to log

    [ https://issues.apache.org/jira/browse/IGNITE-4756?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16152677#comment-16152677 ] 

Vadim Opolski edited comment on IGNITE-4756 at 9/4/17 2:51 PM:
---------------------------------------------------------------

[~tledkov-gridgain] [~avinogradov] In my opinion tests - [TC #2557|https://ci.ignite.apache.org/viewType.html?buildTypeId=Ignite20Tests_RunAll&branch_Ignite20Tests=pull%2F2557%2Fhead&tab=buildTypeStatusDiv] look good.


was (Author: javaller):
[~tledkov-gridgain] [~avinogradov] In my opinion tests look good. [TC #2557|https://ci.ignite.apache.org/viewType.html?buildTypeId=Ignite20Tests_RunAll&branch_Ignite20Tests=pull%2F2557%2Fhead&tab=buildTypeStatusDiv]

> Print info about partition distribution to log 
> -----------------------------------------------
>
>                 Key: IGNITE-4756
>                 URL: https://issues.apache.org/jira/browse/IGNITE-4756
>             Project: Ignite
>          Issue Type: Improvement
>          Components: general
>            Reporter: Taras Ledkov
>            Assignee: Vadim Opolski
>            Priority: Minor
>              Labels: newbie
>             Fix For: 2.3
>
>
> Summarize discussions:
> Add log message in case partitions distribution is not close to even distribution:
> # Add system property IGNITE_PART_DISTRIBUTION_WARN_THRESHOLD with default value 0.1 to print warn message only when nodes count differs more then threshold;
> # The statistic is calculated and printed only for the local node;
> # Statistic is placed at the {{GridAffinityAssignmentCache#calculate}} and calculated for new {{idealAssignment}}.
> # Message format is
> {noformat}
> Local node affinity assignment distribution is not ideal [cache=<cache_name>, expectedPrimary=<ideal_count_of_primary_partitions>, exectedBackups=<ideal_count_of_backups>, primary=<count_of_primary_partitions>, backups=<count_of_backup_partitions>].
> {noformat}
> e.g. for cache with name "test", 2 backups, 4 partition, 3 nodes:
> {noformat}
> Local node affinity assignment distribution is not ideal [cache=test, expectedPrimary=1.33 (33.3%), exectedBackups=2.66 (66.66%), primary=1 (25%), backups=3(75%)].
> {noformat}



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)