[
https://issues.apache.org/jira/browse/HBASE-24967?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Zheng Wang updated HBASE-24967:
-------------------------------
Description:
The reason is that, in order to get the compaction state of table, we need to
iterate all regions and call getRegionInfo, the default timeout for rpc is
5000ms, so if there exists closed reigons, this line will be blocked for long
time.
{code:java}
CompactionState compactionState =
admin.getCompactionState(table.getName()).get();{code}
The solution in this patch is to add compaction state to RegionMetrics, so when
getting the compaction state, we could use the in memory state instead of doing
rpc.
Affect branch-2+, in branch-1 it will display exception stack info immediately.
was:
The reason is that, in order to get the compaction state of table, we need to
iterate all regions and call getRegionInfo, the default timeout for rpc is
5000ms, so if there exists closed reigons, this line will be blocked for long
time.
{code:java}
CompactionState compactionState =
admin.getCompactionState(table.getName()).get();{code}
My proposal is set timeout for it.
Affect branch-2+, in branch-1 it will display exception stack info immediately.
> The table.jsp cost long time to load if the table include closed regions
> ------------------------------------------------------------------------
>
> Key: HBASE-24967
> URL: https://issues.apache.org/jira/browse/HBASE-24967
> Project: HBase
> Issue Type: Improvement
> Components: UI
> Reporter: Zheng Wang
> Assignee: Zheng Wang
> Priority: Major
>
> The reason is that, in order to get the compaction state of table, we need to
> iterate all regions and call getRegionInfo, the default timeout for rpc is
> 5000ms, so if there exists closed reigons, this line will be blocked for long
> time.
> {code:java}
> CompactionState compactionState =
> admin.getCompactionState(table.getName()).get();{code}
> The solution in this patch is to add compaction state to RegionMetrics, so
> when getting the compaction state, we could use the in memory state instead
> of doing rpc.
> Affect branch-2+, in branch-1 it will display exception stack info
> immediately.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
