[
https://issues.apache.org/jira/browse/HBASE-7320?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13901126#comment-13901126
]
ramkrishna.s.vasudevan commented on HBASE-7320:
-----------------------------------------------
bq.How you think we fix these Lars? Would the best approach now be to try and
implement a new Cell type altogether? That would shake out any reliance on KV?
Stack, you asked this question in one of the above comments.
I think your point is valid.
So my thinking was that may be if we could implement a different KeyValue
altogether. May be all the components of the kv are in individual byte arrays.
bq.The server just iterates the Codec and reconstitutes Cells as it thinks best.
So this what i mean by saying use a new codec but again tightly couple it with
current KeyValue. Hope you got what am trying to say here.
bq.But you know, I don't have this worked out end to end. The Codecs and their
API made sense doing the RPC. Might not be best fit for in and out of hfiles
(though I like the idea of the hfile block being able to come up into block
cache and our being able to keep blocks around and have Cell iterators over
them that rehydrate Cells only when necessary (and not if it can be avoided).
I think atleast in cases where we need to strip the tags from the KVs, we are
surely in need of a format which would help us do that easily without having a
needs to reconstruct the KV as how are bound to do now. I am just trying to
make a patch where in every place possible we will refer as Cell rather than
KV. This would mean that even if we change the format of the KV or use a new
type of Cell our internal code does not change.
> Remove KeyValue.getBuffer()
> ---------------------------
>
> Key: HBASE-7320
> URL: https://issues.apache.org/jira/browse/HBASE-7320
> Project: HBase
> Issue Type: Bug
> Reporter: Lars Hofhansl
> Assignee: stack
> Fix For: 0.99.0
>
> Attachments: 7320-simple.txt
>
>
> In many places this is simple task of just replacing the method name.
> There, however, quite a few places where we assume that:
> # the entire KV is backed by a single byte array
> # the KVs key portion is backed by a single byte array
> Some of those can easily be fixed, others will need their own jiras.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
