[
https://issues.apache.org/jira/browse/IGNITE-3018?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15964431#comment-15964431
]
Taras Ledkov edited comment on IGNITE-3018 at 4/11/17 6:33 PM:
---------------------------------------------------------------
Tables below contains average count of partitions to rebalance (new assignment
contains a node that doesn't exist in the previous assignment)
(2 backups, 1024 partitions)
|| Nodes || Fair (total) || Fair (parts on stable nodes)
|| Rendezvous (total) ||Rendezvous (parts on stable nodes) ||
| *3* | 988.3|397.2 | 729.3 | 459.3 |
| *4* | 790.7|563.1 | 638.3 | 392.3 |
| *6* | 408.3|243.8 | 447.0 | 257.9 |
| *8* | 322.8|195.8 | 380.2 | 235.7 |
| *10* | 270.8|156.2 | 314.1 | 194.7 |
| *64* | 47.2 |31.2 | 135.1 | 149.9 |
(2 backups, 10240 partitions)
|| Nodes || Fair (total) || Fair (parts on stable nodes)
|| Rendezvous (total) ||Rendezvous (parts on stable nodes) ||
| *100* | 324.7| 184.4| 570.5|425.1|
| *200* | 157.6| 96.6 | 547.2|692.6|
| *300* | 114.7| 62.3 | 935.7|285.9|
| *400* | 79.2 | 44.4 | 336.3|318.4|
| *500* | 63.0 | 34.9 | 277.3|256.7|
| *600* | 54.0 | 30.1 | 175.8|142.7|
was (Author: tledkov-gridgain):
Tables below contains average count of partitions to rebalance (new assignment
contains a node that doesn't exist in the previous assignment)
(2 backups, 1024 partitions)
|| Nodes || Fair || Rendezvous (balanced) ||
| *3* | 988.3 | 729.3 |
| *4* | 790.7 | 638.3 |
| *6* | 408.3 | 447.0 |
| *8* | 322.8 | 380.2 |
| *10* | 270.8 | 314.1 |
| *64* | 47.2 | 135.1 |
(2 backups, 10240 partitions)
|| Nodes || Fair || Rendezvous (balanced) ||
| *100* | 324.7| 570.5|
| *200* | 157.6| 547.2|
| *300* | 114.7| 935.7|
| *400* | 79.2 | 336.3|
| *500* | 63.0 | 277.3|
| *600* | 54.0 | 175.8|
> Cache affinity calculation is slow with large nodes number
> ----------------------------------------------------------
>
> Key: IGNITE-3018
> URL: https://issues.apache.org/jira/browse/IGNITE-3018
> Project: Ignite
> Issue Type: Bug
> Components: cache
> Reporter: Semen Boikov
> Assignee: Yakov Zhdanov
> Labels: important
> Fix For: 2.0
>
> Attachments: 003.png, 004.png, 008.png, 016.png, 064.png, 100.png,
> 128.png, 200.png, 256.png, 400.png, 600.png, balanced.003.png,
> balanced.004.png, balanced.008.png, balanced.016.png, balanced.064.png,
> balanced.100.png, balanced.128.png, balanced.200.png, balanced.256.png,
> balanced.400.png, balanced.600.png
>
>
> With large number of cache server nodes (> 200) RendezvousAffinityFunction
> and FairAffinityFunction work pretty slow .
> For RendezvousAffinityFunction.assignPartitions can take hundredes of
> milliseconds, for FairAffinityFunction it can take seconds.
> For RendezvousAffinityFunction most time is spent in MD5 hash calculation and
> nodes list sorting. As optimization we can try to cache {partion, node} MD5
> hash or try another hash function. Also several minor optimizations are
> possible (avoid unncecessary allocations, only one thread local 'get', etc).
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
