ignite-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Dmitry Sherstobitov (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (IGNITE-7449) Rebalancing metrics doesn't display actual information about current rebalance state
Date Wed, 17 Jan 2018 11:55:00 GMT

     [ https://issues.apache.org/jira/browse/IGNITE-7449?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Dmitry Sherstobitov updated IGNITE-7449:
----------------------------------------
    Description: 
In case of shutdown node with clear LFS following metrics doesn't display correct information
about rebalance:

RebalancingKeysRate

RebalancingBytesRate

KeysToRebalanceLeft

 

Otherwise RebalancingPartitionsCount metric displays information correctly

Steps to reproduce:
 1. Cluster with enabled statistics
 2. Shutdown node and clear LFS
 3. Run node
 4. Start node and ask node for current rebalance state through JMX:

Current result:

1 tick

RebalancingKeysRate 0

RebalancingBytesRate 0

KeysToRebalanceLeft 0

RebalancingPartitionsCount 342

2 tick

RebalancingKeysRate 0

RebalancingBytesRate 0

KeysToRebalanceLeft 0

RebalancingPartitionsCount 80

Expected:

2 tick

RebalancingKeysRate SOME_NON_ZERO_VALUE

RebalancingBytesRate SOME_NON_ZERO_VALUE

KeysToRebalanceLeft SOME_NON_ZERO_VALUE

RebalancingPartitionsCount 80

 

UPD: -DIGNITE_REBALANCE_STATISTICS_TIME_INTERVAL=1000 doesn't affect results

> Rebalancing metrics doesn't display actual information about current rebalance state
> ------------------------------------------------------------------------------------
>
>                 Key: IGNITE-7449
>                 URL: https://issues.apache.org/jira/browse/IGNITE-7449
>             Project: Ignite
>          Issue Type: Bug
>         Environment: In case of shutdown node with clear LFS following metrics doesn't
display correct information about rebalance:
> RebalancingKeysRate
> RebalancingBytesRate
> KeysToRebalanceLeft
>  
> Otherwise RebalancingPartitionsCount metric displays information correctly
> Steps to reproduce:
> 1. Cluster with enabled statistics
> 2. Shutdown node and clear LFS
> 3. Run node
> 4. Start node and ask node for current rebalance state through JMX:
> Current result:
> 1 tick
> RebalancingKeysRate 0
> RebalancingBytesRate 0
> KeysToRebalanceLeft 0
> RebalancingPartitionsCount 342
> 2 tick
> RebalancingKeysRate 0
> RebalancingBytesRate 0
> KeysToRebalanceLeft 0
> RebalancingPartitionsCount 80
> Expected:
> 2 tick
> RebalancingKeysRate SOME_NON_ZERO_VALUE
> RebalancingBytesRate SOME_NON_ZERO_VALUE
> KeysToRebalanceLeft SOME_NON_ZERO_VALUE
> RebalancingPartitionsCount 80
>  
> UPD: -DIGNITE_REBALANCE_STATISTICS_TIME_INTERVAL=1000 doesn't affect results
>            Reporter: Dmitry Sherstobitov
>            Priority: Major
>
> In case of shutdown node with clear LFS following metrics doesn't display correct information
about rebalance:
> RebalancingKeysRate
> RebalancingBytesRate
> KeysToRebalanceLeft
>  
> Otherwise RebalancingPartitionsCount metric displays information correctly
> Steps to reproduce:
>  1. Cluster with enabled statistics
>  2. Shutdown node and clear LFS
>  3. Run node
>  4. Start node and ask node for current rebalance state through JMX:
> Current result:
> 1 tick
> RebalancingKeysRate 0
> RebalancingBytesRate 0
> KeysToRebalanceLeft 0
> RebalancingPartitionsCount 342
> 2 tick
> RebalancingKeysRate 0
> RebalancingBytesRate 0
> KeysToRebalanceLeft 0
> RebalancingPartitionsCount 80
> Expected:
> 2 tick
> RebalancingKeysRate SOME_NON_ZERO_VALUE
> RebalancingBytesRate SOME_NON_ZERO_VALUE
> KeysToRebalanceLeft SOME_NON_ZERO_VALUE
> RebalancingPartitionsCount 80
>  
> UPD: -DIGNITE_REBALANCE_STATISTICS_TIME_INTERVAL=1000 doesn't affect results



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message