If you are bored this weekend, you can take a look at the table locks doc,
that I attached at HBASE-7305.

There has been a lot of confusion about the current status, design and
implementation, for which I take the blame.

Hope these notes help to answer the questions. If not, leave a note here,
and I'll update the doc to incorporate the answer.

https://issues.apache.org/jira/secure/attachment/12571721/HBaseTableLocks.pdf

Cheers,
Enis

Thanks for documenting that Enis. I was wondering what to do tonight ;)

JM

2013/3/1 Enis Söztutar <[EMAIL PROTECTED]>:
> If you are bored this weekend, you can take a look at the table locks doc,
> that I attached at HBASE-7305.
>
> There has been a lot of confusion about the current status, design and
> implementation, for which I take the blame.
>
> Hope these notes help to answer the questions. If not, leave a note here,
> and I'll update the doc to incorporate the answer.
>
> https://issues.apache.org/jira/secure/attachment/12571721/HBaseTableLocks.pdf
>
> Cheers,
> Enis

bq. master operations acquiring locks for every region, and region splits
acquiring the region lock for their region

Can you elaborate how the above conclusion was drawn ? Looks like region
split is not categorized as master operation.
To my knowledge, region split requires coordination on master side.

It would be nice if you can document how you tested this feature in the pdf.

Cheers

On Fri, Mar 1, 2013 at 6:47 PM, Enis Söztutar <[EMAIL PROTECTED]> wrote:

> If you are bored this weekend, you can take a look at the table locks doc,
> that I attached at HBASE-7305.
>
> There has been a lot of confusion about the current status, design and
> implementation, for which I take the blame.
>
> Hope these notes help to answer the questions. If not, leave a note here,
> and I'll update the doc to incorporate the answer.
>
>
> https://issues.apache.org/jira/secure/attachment/12571721/HBaseTableLocks.pdf
>
> Cheers,
> Enis
>

I read the zookeeper recipe page and then corelated your implementation
with that during this week. Anyway thanks for the doc too :)

Regards
Ram

On Sat, Mar 2, 2013 at 9:17 AM, Ted Yu <[EMAIL PROTECTED]> wrote:

> bq. master operations acquiring locks for every region, and region splits
> acquiring the region lock for their region
>
> Can you elaborate how the above conclusion was drawn ? Looks like region
> split is not categorized as master operation.
> To my knowledge, region split requires coordination on master side.
>
> It would be nice if you can document how you tested this feature in the
> pdf.
>
> Cheers
>
> On Fri, Mar 1, 2013 at 6:47 PM, Enis Söztutar <[EMAIL PROTECTED]> wrote:
>
> > If you are bored this weekend, you can take a look at the table locks
> doc,
> > that I attached at HBASE-7305.
> >
> > There has been a lot of confusion about the current status, design and
> > implementation, for which I take the blame.
> >
> > Hope these notes help to answer the questions. If not, leave a note here,
> > and I'll update the doc to incorporate the answer.
> >
> >
> >
> https://issues.apache.org/jira/secure/attachment/12571721/HBaseTableLocks.pdf
> >
> > Cheers,
> > Enis
> >
>