[
https://issues.apache.org/jira/browse/HBASE-13393?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14557468#comment-14557468
]
Sean Busbey commented on HBASE-13393:
-------------------------------------
{quote}
bq.We are still doing volatile boolean read on each update to memstore
Okie. We can change to non-volatile.
{quote}
Is it safe to change to a non-volatile? What are the concurrency expectations
here, when does it need to be read?
{quote}
bq. In 90+% of cases, tags are not used? What if a table or column family
config to say 'use tags'?
There is no such feature we have. Tags are not supported directly from clients.
Tags are now added to Cells only in the server side as part of ACL or
Visibilities. The client passes these information only as Attributes.
The reverse where a server sends a Tag to the client is also not there except
in the case of replication or cell replay where we use WalCodec (in case of WAL
replay).
{quote}
{code}
+ // In no tags case this NoTagsKeyValue.getTagsLength() is a cheap call.
+ // When we use ACL CP or Visibility CP which deals with Tags during
+ // mutation, the TagRewriteCell.getTagsLength() is a cheaper call. We do
not
+ // parse the byte[] to identify the tags length.
+ if(!tagsAvailable && e.getTagsLength() > 0) {
+ tagsAvailable = true;
+ }
{code}
just change this check to ignore the current state of tagsAvailable. a rare
spurious assignment to a volatile variable is better than a frequent read.
> Add facility to disable tags after parent goes in
> -------------------------------------------------
>
> Key: HBASE-13393
> URL: https://issues.apache.org/jira/browse/HBASE-13393
> Project: HBase
> Issue Type: Sub-task
> Reporter: stack
> Assignee: stack
> Attachments: HBASE-13393.patch, HBASE-13393_1.patch
>
>
> We used to specify 'no tags' by choosing to write files that were version 2.
> After parent goes in, make it so can ask a v3 file not to write tags.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
