[
https://issues.apache.org/jira/browse/HBASE-26510?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yutong Xiao updated HBASE-26510:
--------------------------------
Description:
Currently, we use CACHE_DATA_ON_READ_KEY as BLOCKCACHE_ENABLED. This makes the
usage ambiguous.
In my opinion, we should use a flag to control the on/off of the whole
blockcache and one another flag to control if we cache data blocks on read
(cache on write stay the same).
I added a new BLOCKCACHE_ENABLED key to make the name and usage more clear.
However, there are some logic problems in the other functions in
TestCacheConfig. I have submit an MR for this. Hope to discuss about the logic.
was:Currently, we use CACHE_DATA_ON_READ_KEY as BLOCKCACHE_ENABLED. This
makes the usage ambiguous. I added a new BLOCKCACHE_ENABLED key to make the
name and usage more clear. However, there are some logic problems in the other
functions in TestCacheConfig. I have submit an MR for this. Hope to discuss
about the logic.
> Separate the blockcache enabled key and cache data on read key
> --------------------------------------------------------------
>
> Key: HBASE-26510
> URL: https://issues.apache.org/jira/browse/HBASE-26510
> Project: HBase
> Issue Type: Task
> Reporter: Yutong Xiao
> Assignee: Yutong Xiao
> Priority: Minor
>
> Currently, we use CACHE_DATA_ON_READ_KEY as BLOCKCACHE_ENABLED. This makes
> the usage ambiguous.
> In my opinion, we should use a flag to control the on/off of the whole
> blockcache and one another flag to control if we cache data blocks on read
> (cache on write stay the same).
> I added a new BLOCKCACHE_ENABLED key to make the name and usage more clear.
> However, there are some logic problems in the other functions in
> TestCacheConfig. I have submit an MR for this. Hope to discuss about the
> logic.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
