[
https://issues.apache.org/jira/browse/HBASE-10074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13838398#comment-13838398
]
Nick Dimiduk commented on HBASE-10074:
--------------------------------------
I don't read docbook, so I cannot comment about the markup. However, the
content is great! Here are some nits.
bq. <para>The master as is is allergic to tons of regions
Mind adding some JIRA references here?
bq. tons of regions on a few RS can cause the store file index to rise raising
heap usage and...
"store file index to rise rising" ? This sentence is confusing me.
bq. Keeping 5 regions per RS would be too low for a job, whereas 1000 will
generate too many maps.
How about "Hosting only 5 regions per RS will not be enough task splits for a
mapreduce job, while 1000 regions will generate far too many map tasks."
In section {{<section xml:id="ops.capacity.regions"><title>Determining region
count and size</title>}} you suggest "20-200 regions per RS" but previously you
said "20-100".
+1
> consolidate and improve capacity/sizing documentation
> -----------------------------------------------------
>
> Key: HBASE-10074
> URL: https://issues.apache.org/jira/browse/HBASE-10074
> Project: HBase
> Issue Type: Improvement
> Components: documentation
> Reporter: Sergey Shelukhin
> Assignee: Sergey Shelukhin
> Attachments: HBASE-10074.patch
>
>
> Region count description is in config section; region size description is in
> architecture sections; both of these have a lot of good technical details,
> but imho we could do better in terms of admin-centric advice.
> Currently, there's a nearly-empty capacity section; I'd like to rewrite it to
> consolidate capacity planning/sizing/region sizing information, and some
> basic configuration pertaining to it.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
