[
https://issues.apache.org/jira/browse/IOTDB-1373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17355624#comment-17355624
]
lan.liang commented on IOTDB-1373:
----------------------------------
So how can we work for "automatically" ? Do you have any design?
> Adjust the tsfile_size_threshold automatically
> ----------------------------------------------
>
> Key: IOTDB-1373
> URL: https://issues.apache.org/jira/browse/IOTDB-1373
> Project: Apache IoTDB
> Issue Type: Improvement
> Reporter: Jialin Qiao
> Priority: Major
>
> Currently, when we need to flush a memtable, we check if the size of target
> TsFile is larger than the tsfile_size_threshold. If the target TsFile is
> larger than this parameter, we close the TsFile after flushing.
>
> If a TsFile is not close, all ChunkMetadata will cached in memory, so we set
> tsfile_size_threshold to 1 to restrict that only two memtables are in TsFile.
> This is a simple way to avoid OOM.
>
> However, each file contains two memtable is not efficient.
> We need to enlarge tsfile_size_threshold and avoid OOM.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
