You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Vladimir Ozerov (JIRA)" <ji...@apache.org> on 2018/11/21 09:59:00 UTC
[jira] [Comment Edited] (IGNITE-9777) Gathering local node IO
statistics with cache and index dimensions
[ https://issues.apache.org/jira/browse/IGNITE-9777?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16693201#comment-16693201 ]
Vladimir Ozerov edited comment on IGNITE-9777 at 11/21/18 9:58 AM:
-------------------------------------------------------------------
[~jooger], my comments:
1) IoStatisticsQueryHelper - unused methods, should be removed?
2) IoStatisticsHolderIndex.HASH_PK_IDX_NAME - "PK" as name might be confusing because we also have PK index in SQL. I would rename it to "HASH_PK" instead.
3) IoStatisticsHolderIndex#toString - why not using S.toString() instead? IoStatisticsHolderCache#toString - same thing. IoStatisticsHolderQuery - but here we do not have "toString" at all, why?
4) IoStatisticsManager.statsSince - why local and not UTC time?
5) IoStatisticsMetricsMXBean: "getStartGatheringStatistics" -> "getStartTime". Please make sure that LocalDateTime displayed correctly in JConsole or any other JMX tool
6) IoStatisticsMetricsMXBean: "resetStatistics" -> "reset"
7) IoStatisticsMetricsMXBean: "getCacheStatisticsFormatted" -> "getCacheGroupStatistics", "getCachePhysicalReadsStatistics" -> "getCacheGroupPhysicalReads", "getCacheLogicalReadsStatistics" -> "getCacheGroupLogicalReads"
8) Same as p.7 but for indexes.
9) "getStatHashIndexesNames", "getStatSortedIndexesNames", "getStatCachesNames" - I do not think we need this at all. It is better to simply list available constants in docs.
10) IoStatisticsMetricsMXBean.getCachePhysicalReadsStatistics, getIndexPhysicalReadsStatistics, getIndexLeafPhysicalReadsStatistics, getIndexInnerPhysicalReadsStatistics -> looks like these methods are never tested.
was (Author: vozerov):
[~jooger], my comments:
1) IoStatisticsQueryHelper - unused methods, should be removed?
2) IoStatisticsHolderIndex.HASH_PK_IDX_NAME - "PK" as name might be confusing because we also have PK index in SQL. I would rename it to "HASH_PK" instead.
3) IoStatisticsHolderIndex#toString - why not using S.toString() instead? IoStatisticsHolderCache#toString - same thing. IoStatisticsHolderQuery - but here we do not have "toString" at all, why?
4) IoStatisticsManager.statsSince - why local and not UTF time?
5) IoStatisticsMetricsMXBean: "getStartGatheringStatistics" -> "getStartTime". Please make sure that LocalDateTime displayed correctly in JConsole or any other JMX tool
6) IoStatisticsMetricsMXBean: "resetStatistics" -> "reset"
7) IoStatisticsMetricsMXBean: "getCacheStatisticsFormatted" -> "getCacheGroupStatistics", "getCachePhysicalReadsStatistics" -> "getCacheGroupPhysicalReads", "getCacheLogicalReadsStatistics" -> "getCacheGroupLogicalReads"
8) Same as p.7 but for indexes.
9) "getStatHashIndexesNames", "getStatSortedIndexesNames", "getStatCachesNames" - I do not think we need this at all. It is better to simply list available constants in docs.
10) IoStatisticsMetricsMXBean.getCachePhysicalReadsStatistics, getIndexPhysicalReadsStatistics, getIndexLeafPhysicalReadsStatistics, getIndexInnerPhysicalReadsStatistics -> looks like these methods are never tested.
> Gathering local node IO statistics with cache and index dimensions
> ------------------------------------------------------------------
>
> Key: IGNITE-9777
> URL: https://issues.apache.org/jira/browse/IGNITE-9777
> Project: Ignite
> Issue Type: Task
> Components: persistence, sql
> Reporter: Yury Gerzhedovich
> Assignee: Yury Gerzhedovich
> Priority: Major
> Labels: iep-27
> Fix For: 2.8
>
>
> Should be gathered and exposed IO statistics on cache/index level.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)