Hi Folks-

You probably already my email yesterday on this...
   https://issues.apache.org/jira/browse/HBASE-4089 (block cache report)

...and I just created this one...
   https://issues.apache.org/jira/browse/HBASE-4147 (StoreFile query
report)

What I'd like to run past the dev-list is this:  if Hbase had periodic
summary usage statistics, where should they go?  What I'd like to throw
out for discussion is that I'm suggesting that it should simply go to the
log files and users can slice and dice this on their own.  No UI (I.e.,
JSPs), no JMX, etc.
The summary out the output is this:
BlockCacheReport:  on configured interval, print out summary of blockcache
(at table/CF level) to log file. This one is point-in-time, not delta.

StoreFile read report:  on configured interval, print out summary of
StoreFile accesses and how much time was spent reading each StoreFile to
log file.  

Thoughts?

Doug

>

Can you cast some of these new attributes as metrics (This won't work
for the per-table/per-cf attributes).  Print to log could work.  Here
we json compound info like this when we stick it in log.  Maybe here
too?  Would be nice if an API call too... or add the data to
ClusterStatus fat report object.  Then it could be seen over in shell,
etc.

St.Ack

On Fri, Jul 29, 2011 at 4:15 AM, Doug Meil
<[EMAIL PROTECTED]> wrote:
>
> Hi Folks-
>
> You probably already my email yesterday on this...
>   https://issues.apache.org/jira/browse/HBASE-4089 (block cache report)
>
> ...and I just created this one...
>   https://issues.apache.org/jira/browse/HBASE-4147 (StoreFile query
> report)
>
> What I'd like to run past the dev-list is this:  if Hbase had periodic
> summary usage statistics, where should they go?  What I'd like to throw
> out for discussion is that I'm suggesting that it should simply go to the
> log files and users can slice and dice this on their own.  No UI (I.e.,
> JSPs), no JMX, etc.
>
>
> The summary out the output is this:
> BlockCacheReport:  on configured interval, print out summary of blockcache
> (at table/CF level) to log file. This one is point-in-time, not delta.
>
> StoreFile read report:  on configured interval, print out summary of
> StoreFile accesses and how much time was spent reading each StoreFile to
> log file.
>
> Thoughts?
>
> Doug
>
>>
>
>

What I'd prefer is something like:

interface BlockCacheReportSink {
  public void reportStats(BlockCacheReport report);
}

class LoggingBlockCacheReportSink {
  ... {
    log it with whatever formatting you want
  }
}

then a configuration which could default to the logging implementation, but
orgs could easily substitute their own implementation. For example, I could
see wanting to do an implementation where it keeps local RRD graphs of some
stats, or pushes them to a central management server.

The assumption is that BlockCacheReport is a fairly straightforward "struct"
with the non-formatted information available.

-Todd

On Fri, Jul 29, 2011 at 4:15 AM, Doug Meil <[EMAIL PROTECTED]>wrote:

>
> Hi Folks-
>
> You probably already my email yesterday on this...
>   https://issues.apache.org/jira/browse/HBASE-4089 (block cache report)
>
> ...and I just created this one...
>   https://issues.apache.org/jira/browse/HBASE-4147 (StoreFile query
> report)
>
> What I'd like to run past the dev-list is this:  if Hbase had periodic
> summary usage statistics, where should they go?  What I'd like to throw
> out for discussion is that I'm suggesting that it should simply go to the
> log files and users can slice and dice this on their own.  No UI (I.e.,
> JSPs), no JMX, etc.
>
>
> The summary out the output is this:
> BlockCacheReport:  on configured interval, print out summary of blockcache
> (at table/CF level) to log file. This one is point-in-time, not delta.
>
> StoreFile read report:  on configured interval, print out summary of
> StoreFile accesses and how much time was spent reading each StoreFile to
> log file.
>
> Thoughts?
>
> Doug
>
> >
>
>
--
Todd Lipcon
Software Engineer, Cloudera

There is also the matter of HServerLoad and how that is used by the shell and master UI to report on cluster status.

I'd like the shell to be able to let the user explore all of these different reports interactively.

At the very least, they should all be handled the same way.

And then there is Riley's work over at FB on a slow query log. How does that fit in? 
 
Best regards,
   - Andy

Problems worthy of attack prove their worth by hitting back. - Piet Hein (via Tom White)
>________________________________
>From: Todd Lipcon <[EMAIL PROTECTED]>
>To: [EMAIL PROTECTED]
>Sent: Friday, July 29, 2011 9:58 AM
>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>
>What I'd prefer is something like:
>
>interface BlockCacheReportSink {
>  public void reportStats(BlockCacheReport report);
>}
>
>class LoggingBlockCacheReportSink {
>  ... {
>    log it with whatever formatting you want
>  }
>}
>
>then a configuration which could default to the logging implementation, but
>orgs could easily substitute their own implementation. For example, I could
>see wanting to do an implementation where it keeps local RRD graphs of some
>stats, or pushes them to a central management server.
>
>The assumption is that BlockCacheReport is a fairly straightforward "struct"
>with the non-formatted information available.
>
>-Todd
>
>On Fri, Jul 29, 2011 at 4:15 AM, Doug Meil <[EMAIL PROTECTED]>wrote:
>
>>
>> Hi Folks-
>>
>> You probably already my email yesterday on this...
>>  https://issues.apache.org/jira/browse/HBASE-4089 (block cache report)
>>
>> ...and I just created this one...
>>  https://issues.apache.org/jira/browse/HBASE-4147 (StoreFile query
>> report)
>>
>> What I'd like to run past the dev-list is this:  if Hbase had periodic
>> summary usage statistics, where should they go?  What I'd like to throw
>> out for discussion is that I'm suggesting that it should simply go to the
>> log files and users can slice and dice this on their own.  No UI (I.e.,
>> JSPs), no JMX, etc.
>>
>>
>> The summary out the output is this:
>> BlockCacheReport:  on configured interval, print out summary of blockcache
>> (at table/CF level) to log file. This one is point-in-time, not delta.
>>
>> StoreFile read report:  on configured interval, print out summary of
>> StoreFile accesses and how much time was spent reading each StoreFile to
>> log file.
>>
>> Thoughts?
>>
>> Doug
>>
>> >
>>
>>
>
>
>--
>Todd Lipcon
>Software Engineer, Cloudera
>
>
>

On Fri, Jul 29, 2011 at 10:15 AM, Andrew Purtell <[EMAIL PROTECTED]>wrote:

> There is also the matter of HServerLoad and how that is used by the shell
> and master UI to report on cluster status.
>
> I'd like the shell to be able to let the user explore all of these
> different reports interactively.
>

+1. In some use cases the users of HBase don't have access to the log files.

With my above idea of adding an interface, we could have an implementation
which keeps the most recent report in memory and exposes it via a servlet as
well.
>
> At the very least, they should all be handled the same way.
>
> And then there is Riley's work over at FB on a slow query log. How does
> that fit in?
>
> Best regards,
>
>
>    - Andy
>
> Problems worthy of attack prove their worth by hitting back. - Piet Hein
> (via Tom White)
>
>
> >________________________________
> >From: Todd Lipcon <[EMAIL PROTECTED]>
> >To: [EMAIL PROTECTED]
> >Sent: Friday, July 29, 2011 9:58 AM
> >Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
> >
> >What I'd prefer is something like:
> >
> >interface BlockCacheReportSink {
> >  public void reportStats(BlockCacheReport report);
> >}
> >
> >class LoggingBlockCacheReportSink {
> >  ... {
> >    log it with whatever formatting you want
> >  }
> >}
> >
> >then a configuration which could default to the logging implementation,
> but
> >orgs could easily substitute their own implementation. For example, I
> could
> >see wanting to do an implementation where it keeps local RRD graphs of
> some
> >stats, or pushes them to a central management server.
> >
> >The assumption is that BlockCacheReport is a fairly straightforward
> "struct"
> >with the non-formatted information available.
> >
> >-Todd
> >
> >On Fri, Jul 29, 2011 at 4:15 AM, Doug Meil <[EMAIL PROTECTED]
> >wrote:
> >
> >>
> >> Hi Folks-
> >>
> >> You probably already my email yesterday on this...
> >>  https://issues.apache.org/jira/browse/HBASE-4089 (block cache report)
> >>
> >> ...and I just created this one...
> >>  https://issues.apache.org/jira/browse/HBASE-4147 (StoreFile query
> >> report)
> >>
> >> What I'd like to run past the dev-list is this:  if Hbase had periodic
> >> summary usage statistics, where should they go?  What I'd like to throw
> >> out for discussion is that I'm suggesting that it should simply go to
> the
> >> log files and users can slice and dice this on their own.  No UI (I.e.,
> >> JSPs), no JMX, etc.
> >>
> >>
> >> The summary out the output is this:
> >> BlockCacheReport:  on configured interval, print out summary of
> blockcache
> >> (at table/CF level) to log file. This one is point-in-time, not delta.
> >>
> >> StoreFile read report:  on configured interval, print out summary of
> >> StoreFile accesses and how much time was spent reading each StoreFile to
> >> log file.
> >>
> >> Thoughts?
> >>
> >> Doug
> >>
> >> >
> >>
> >>
> >
> >
> >--
> >Todd Lipcon
> >Software Engineer, Cloudera
> >
> >
> >
>

--
Todd Lipcon
Software Engineer, Cloudera

I'd rather see this output being able to be captured by something the sink
that Todd suggested, rather than focusing on shell access.  HServerLoad is
super-summary at the RS level, and both the items in 4089 and 4147 are
proposed to be "summarized" but still have reasonable detail (e.g., even
table/CF summary there could be dozens of entries given a reasonably
complex system).
On 7/29/11 1:15 PM, "Andrew Purtell" <[EMAIL PROTECTED]> wrote:

>There is also the matter of HServerLoad and how that is used by the shell
>and master UI to report on cluster status.
>
>I'd like the shell to be able to let the user explore all of these
>different reports interactively.
>
>At the very least, they should all be handled the same way.
>
>And then there is Riley's work over at FB on a slow query log. How does
>that fit in?
>
>Best regards,
>
>
>   - Andy
>
>Problems worthy of attack prove their worth by hitting back. - Piet Hein
>(via Tom White)
>
>
>>________________________________
>>From: Todd Lipcon <[EMAIL PROTECTED]>
>>To: [EMAIL PROTECTED]
>>Sent: Friday, July 29, 2011 9:58 AM
>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>
>>What I'd prefer is something like:
>>
>>interface BlockCacheReportSink {
>>  public void reportStats(BlockCacheReport report);
>>}
>>
>>class LoggingBlockCacheReportSink {
>>  ... {
>>    log it with whatever formatting you want
>>  }
>>}
>>
>>then a configuration which could default to the logging implementation,
>>but
>>orgs could easily substitute their own implementation. For example, I
>>could
>>see wanting to do an implementation where it keeps local RRD graphs of
>>some
>>stats, or pushes them to a central management server.
>>
>>The assumption is that BlockCacheReport is a fairly straightforward
>>"struct"
>>with the non-formatted information available.
>>
>>-Todd
>>
>>On Fri, Jul 29, 2011 at 4:15 AM, Doug Meil
>><[EMAIL PROTECTED]>wrote:
>>
>>>
>>> Hi Folks-
>>>
>>> You probably already my email yesterday on this...
>>>  https://issues.apache.org/jira/browse/HBASE-4089 (block cache report)
>>>
>>> ...and I just created this one...
>>>  https://issues.apache.org/jira/browse/HBASE-4147 (StoreFile query
>>> report)
>>>
>>> What I'd like to run past the dev-list is this:  if Hbase had periodic
>>> summary usage statistics, where should they go?  What I'd like to throw
>>> out for discussion is that I'm suggesting that it should simply go to
>>>the
>>> log files and users can slice and dice this on their own.  No UI (I.e.,
>>> JSPs), no JMX, etc.
>>>
>>>
>>> The summary out the output is this:
>>> BlockCacheReport:  on configured interval, print out summary of
>>>blockcache
>>> (at table/CF level) to log file. This one is point-in-time, not delta.
>>>
>>> StoreFile read report:  on configured interval, print out summary of
>>> StoreFile accesses and how much time was spent reading each StoreFile
>>>to
>>> log file.
>>>
>>> Thoughts?
>>>
>>> Doug
>>>
>>> >
>>>
>>>
>>
>>
>>--
>>Todd Lipcon
>>Software Engineer, Cloudera
>>
>>

> I'd rather see this output being able to be captured by something the sink that Todd suggested, rather than focusing on shell access.
I don't agree.
Look at what we have existing and proposed:

    - Java API access to server and region load information, that the shell uses

    - A proposal to dump some stats into log files, that then has to be scraped

    - A proposal (by the FB guys) to export some JSON via a HTTP servlet

This is not good design, this is a bunch of random shit stuck together. 

Note that what Todd proposed does not preclude adding Java client API support for retrieving it.

At a minimum all of this information must be accessible via the Java client API, to enable programmatic monitoring and analysis use cases. I'll add the shell support if nobody else cares about it, that is a relatively small detail, but one I think is important. 

Best regards,
    - Andy
Problems worthy of attack prove their worth by hitting back. - Piet Hein (via Tom White)
>________________________________
>From: Doug Meil <[EMAIL PROTECTED]>
>To: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>; "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
>Sent: Friday, July 29, 2011 11:39 AM
>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>
>
>I'd rather see this output being able to be captured by something the sink
>that Todd suggested, rather than focusing on shell access.  HServerLoad is
>super-summary at the RS level, and both the items in 4089 and 4147 are
>proposed to be "summarized" but still have reasonable detail (e.g., even
>table/CF summary there could be dozens of entries given a reasonably
>complex system).
>
>
>
>
>On 7/29/11 1:15 PM, "Andrew Purtell" <[EMAIL PROTECTED]> wrote:
>
>>There is also the matter of HServerLoad and how that is used by the shell
>>and master UI to report on cluster status.
>>
>>I'd like the shell to be able to let the user explore all of these
>>different reports interactively.
>>
>>At the very least, they should all be handled the same way.
>>
>>And then there is Riley's work over at FB on a slow query log. How does
>>that fit in?
>>
>>Best regards,
>>
>>
>>   - Andy
>>
>>Problems worthy of attack prove their worth by hitting back. - Piet Hein
>>(via Tom White)
>>
>>
>>>________________________________
>>>From: Todd Lipcon <[EMAIL PROTECTED]>
>>>To: [EMAIL PROTECTED]
>>>Sent: Friday, July 29, 2011 9:58 AM
>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>
>>>What I'd prefer is something like:
>>>
>>>interface BlockCacheReportSink {
>>>  public void reportStats(BlockCacheReport report);
>>>}
>>>
>>>class LoggingBlockCacheReportSink {
>>>  ... {
>>>    log it with whatever formatting you want
>>>  }
>>>}
>>>
>>>then a configuration which could default to the logging implementation,
>>>but
>>>orgs could easily substitute their own implementation. For example, I
>>>could
>>>see wanting to do an implementation where it keeps local RRD graphs of
>>>some
>>>stats, or pushes them to a central management server.
>>>
>>>The assumption is that BlockCacheReport is a fairly straightforward
>>>"struct"
>>>with the non-formatted information available.
>>>
>>>-Todd
>>>
>>>On Fri, Jul 29, 2011 at 4:15 AM, Doug Meil
>>><[EMAIL PROTECTED]>wrote:
>>>
>>>>
>>>> Hi Folks-
>>>>
>>>> You probably already my email yesterday on this...
>>>>  https://issues.apache.org/jira/browse/HBASE-4089 (block cache report)
>>>>
>>>> ...and I just created this one...
>>>>  https://issues.apache.org/jira/browse/HBASE-4147 (StoreFile query
>>>> report)
>>>>
>>>> What I'd like to run past the dev-list is this:  if Hbase had periodic
>>>> summary usage statistics, where should they go?  What I'd like to throw
>>>> out for discussion is that I'm suggesting that it should simply go to
>>>>the
>>>> log files and users can slice and dice this on their own.  No UI (I.e.,
>>>> JSPs), no JMX, etc.
>>>>
>>>>
>>>> The summary out the output is this:
>>>> BlockCacheReport:  on configured interval, print out summary of

On Fri, Jul 29, 2011 at 1:34 PM, Andrew Purtell <[EMAIL PROTECTED]> wrote:
> This is not good design, this is a bunch of random shit stuck together.
>

Right.  Lets not do random shit (Strike the dumping to log files if I
was one who suggested it).
St.Ack

I'll claim the log-file suggestion, but right now we have very little
operational insight and I'll take info in log-files over nothing.
On 7/29/11 4:41 PM, "Stack" <[EMAIL PROTECTED]> wrote:

>On Fri, Jul 29, 2011 at 1:34 PM, Andrew Purtell <[EMAIL PROTECTED]>
>wrote:
>> This is not good design, this is a bunch of random shit stuck together.
>>
>
>Right.  Lets not do random shit (Strike the dumping to log files if I
>was one who suggested it).
>St.Ack

On Fri, Jul 29, 2011 at 2:07 PM, Doug Meil <[EMAIL PROTECTED]>wrote:

>
> I'll claim the log-file suggestion, but right now we have very little
> operational insight and I'll take info in log-files over nothing.
>
>
Agree log files are better than nothing. But let's design in an interface so
that we can build other tooling on top.
>
>
>
> On 7/29/11 4:41 PM, "Stack" <[EMAIL PROTECTED]> wrote:
>
> >On Fri, Jul 29, 2011 at 1:34 PM, Andrew Purtell <[EMAIL PROTECTED]>
> >wrote:
> >> This is not good design, this is a bunch of random shit stuck together.
> >>
> >
> >Right.  Lets not do random shit (Strike the dumping to log files if I
> >was one who suggested it).
> >St.Ack
>
>
--
Todd Lipcon
Software Engineer, Cloudera

Thanks everybody.  I'll put something together on paper (PDF,
specifically) and attach it to the JIRA for and comments will be most
welcome.
On 7/29/11 5:34 PM, "Todd Lipcon" <[EMAIL PROTECTED]> wrote:

>On Fri, Jul 29, 2011 at 2:07 PM, Doug Meil
><[EMAIL PROTECTED]>wrote:
>
>>
>> I'll claim the log-file suggestion, but right now we have very little
>> operational insight and I'll take info in log-files over nothing.
>>
>>
>Agree log files are better than nothing. But let's design in an interface
>so
>that we can build other tooling on top.
>
>
>>
>>
>>
>> On 7/29/11 4:41 PM, "Stack" <[EMAIL PROTECTED]> wrote:
>>
>> >On Fri, Jul 29, 2011 at 1:34 PM, Andrew Purtell <[EMAIL PROTECTED]>
>> >wrote:
>> >> This is not good design, this is a bunch of random shit stuck
>>together.
>> >>
>> >
>> >Right.  Lets not do random shit (Strike the dumping to log files if I
>> >was one who suggested it).
>> >St.Ack
>>
>>
>
>
>--
>Todd Lipcon
>Software Engineer, Cloudera

Hi,

I'm for publishing all performance metrics in JMX (in addition to exposing it wherever else you guys decide).  That's because JMX is probably the easiest for our SPM for HBase [1] to get to HBase performance metrics and I suspect we are not alone.

Otis
[1] http://sematext.com/spm/hbase-performance-monitoring/index.html
----
Sematext :: http://sematext.com/ :: Solr - Lucene - Hadoop - HBase
Hadoop ecosystem search :: http://search-hadoop.com/

>________________________________
>From: Andrew Purtell <[EMAIL PROTECTED]>
>To: Doug Meil <[EMAIL PROTECTED]>; "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
>Sent: Friday, July 29, 2011 4:34 PM
>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>
>> I'd rather see this output being able to be captured by something the sink that Todd suggested, rather than focusing on shell access.
>
>
>I don't agree.
>
>
>Look at what we have existing and proposed:
>
>    - Java API access to server and region load information, that the shell uses
>
>    - A proposal to dump some stats into log files, that then has to be scraped
>
>    - A proposal (by the FB guys) to export some JSON via a HTTP servlet
>
>This is not good design, this is a bunch of random shit stuck together. 
>
>Note that what Todd proposed does not preclude adding Java client API support for retrieving it.
>
>At a minimum all of this information must be accessible via the Java client API, to enable programmatic monitoring and analysis use cases. I'll add the shell support if nobody else cares about it, that is a relatively small detail, but one I think is important. 
>
>Best regards,
>
>
>    - Andy
>
>
>Problems worthy of attack prove their worth by hitting back. - Piet Hein (via Tom White)
>
>
>>________________________________
>>From: Doug Meil <[EMAIL PROTECTED]>
>>To: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>; "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
>>Sent: Friday, July 29, 2011 11:39 AM
>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>
>>
>>I'd rather see this output being able to be captured by something the sink
>>that Todd suggested, rather than focusing on shell access.  HServerLoad is
>>super-summary at the RS level, and both the items in 4089 and 4147 are
>>proposed to be "summarized" but still have reasonable detail (e.g., even
>>table/CF summary there could be dozens of entries given a reasonably
>>complex system).
>>
>>
>>
>>
>>On 7/29/11 1:15 PM, "Andrew Purtell" <[EMAIL PROTECTED]> wrote:
>>
>>>There is also the matter of HServerLoad and how that is used by the shell
>>>and master UI to report on cluster status.
>>>
>>>I'd like the shell to be able to let the user explore all of these
>>>different reports interactively.
>>>
>>>At the very least, they should all be handled the same way.
>>>
>>>And then there is Riley's work over at FB on a slow query log. How does
>>>that fit in?
>>>
>>>Best regards,
>>>
>>>
>>>   - Andy
>>>
>>>Problems worthy of attack prove their worth by hitting back. - Piet Hein
>>>(via Tom White)
>>>
>>>
>>>>________________________________
>>>>From: Todd Lipcon <[EMAIL PROTECTED]>
>>>>To: [EMAIL PROTECTED]
>>>>Sent: Friday, July 29, 2011 9:58 AM
>>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>>
>>>>What I'd prefer is something like:
>>>>
>>>>interface BlockCacheReportSink {
>>>>  public void reportStats(BlockCacheReport report);
>>>>}
>>>>
>>>>class LoggingBlockCacheReportSink {
>>>>  ... {
>>>>    log it with whatever formatting you want
>>>>  }
>>>>}
>>>>
>>>>then a configuration which could default to the logging implementation,
>>>>but
>>>>orgs could easily substitute their own implementation. For example, I
>>>>could
>>>>see wanting to do an implementation where it keeps local RRD graphs of
>>>>some
>>>>stats, or pushes them to a central management server.
>>>>
>>>>The assumption is that BlockCacheReport is a fairly straightforward
>>>>"struct"
>>>>with the non-formatted information available.

What is it all about? HBase sucks. Too many problems to newcomers,
few-weeks-warm-up to begin with!!!!!!!!!!!!!!!! Is it really-really
supported by Microsoft employees?!


And, SEO of course:
==================

--
Fuad Efendi
416-993-2060
Tokenizer Inc., Canada
Data Mining, Search Engines
http://www.tokenizer.ca
On 11-07-29 7:49 PM, "Otis Gospodnetic" <[EMAIL PROTECTED]> wrote:

>Hi,
>
>I'm for publishing all performance metrics in JMX (in addition to
>exposing it wherever else you guys decide).  That's because JMX is
>probably the easiest for our SPM for HBase [1] to get to HBase
>performance metrics and I suspect we are not alone.
>
>Otis
>[1] http://sematext.com/spm/hbase-performance-monitoring/index.html
>----
>Sematext :: http://sematext.com/ :: Solr - Lucene - Hadoop - HBase
>Hadoop ecosystem search :: http://search-hadoop.com/
>
>
>
>>________________________________
>>From: Andrew Purtell <[EMAIL PROTECTED]>
>>To: Doug Meil <[EMAIL PROTECTED]>; "[EMAIL PROTECTED]"
>><[EMAIL PROTECTED]>
>>Sent: Friday, July 29, 2011 4:34 PM
>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>
>>> I'd rather see this output being able to be captured by something the
>>>sink that Todd suggested, rather than focusing on shell access.
>>
>>
>>I don't agree.
>>
>>
>>Look at what we have existing and proposed:
>>
>>    - Java API access to server and region load information, that the
>>shell uses
>>
>>    - A proposal to dump some stats into log files, that then has to be
>>scraped
>>
>>    - A proposal (by the FB guys) to export some JSON via a HTTP servlet
>>
>>This is not good design, this is a bunch of random shit stuck together.
>>
>>Note that what Todd proposed does not preclude adding Java client API
>>support for retrieving it.
>>
>>At a minimum all of this information must be accessible via the Java
>>client API, to enable programmatic monitoring and analysis use cases.
>>I'll add the shell support if nobody else cares about it, that is a
>>relatively small detail, but one I think is important.
>>
>>Best regards,
>>
>>
>>    - Andy
>>
>>
>>Problems worthy of attack prove their worth by hitting back. - Piet Hein
>>(via Tom White)
>>
>>
>>>________________________________
>>>From: Doug Meil <[EMAIL PROTECTED]>
>>>To: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>;
>>>"[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
>>>Sent: Friday, July 29, 2011 11:39 AM
>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>
>>>
>>>I'd rather see this output being able to be captured by something the
>>>sink
>>>that Todd suggested, rather than focusing on shell access.  HServerLoad
>>>is
>>>super-summary at the RS level, and both the items in 4089 and 4147 are
>>>proposed to be "summarized" but still have reasonable detail (e.g., even
>>>table/CF summary there could be dozens of entries given a reasonably
>>>complex system).
>>>
>>>
>>>
>>>
>>>On 7/29/11 1:15 PM, "Andrew Purtell" <[EMAIL PROTECTED]> wrote:
>>>
>>>>There is also the matter of HServerLoad and how that is used by the
>>>>shell
>>>>and master UI to report on cluster status.
>>>>
>>>>I'd like the shell to be able to let the user explore all of these
>>>>different reports interactively.
>>>>
>>>>At the very least, they should all be handled the same way.
>>>>
>>>>And then there is Riley's work over at FB on a slow query log. How does
>>>>that fit in?
>>>>
>>>>Best regards,
>>>>
>>>>
>>>>   - Andy
>>>>
>>>>Problems worthy of attack prove their worth by hitting back. - Piet
>>>>Hein
>>>>(via Tom White)
>>>>
>>>>
>>>>>________________________________
>>>>>From: Todd Lipcon <[EMAIL PROTECTED]>
>>>>>To: [EMAIL PROTECTED]
>>>>>Sent: Friday, July 29, 2011 9:58 AM
>>>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>>>
>>>>>What I'd prefer is something like:
>>>>>
>>>>>interface BlockCacheReportSink {
>>>>>  public void reportStats(BlockCacheReport report);
>>>>>}
>>>>>
>>>>>class LoggingBlockCacheReportSink {

You should ask for your money back!!

On Sun, Jul 31, 2011 at 3:10 PM, Fuad Efendi <[EMAIL PROTECTED]> wrote:
> What is it all about? HBase sucks. Too many problems to newcomers,
> few-weeks-warm-up to begin with!!!!!!!!!!!!!!!! Is it really-really
> supported by Microsoft employees?!
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> And, SEO of course:
> ==================>
>
> --
> Fuad Efendi
> 416-993-2060
> Tokenizer Inc., Canada
> Data Mining, Search Engines
> http://www.tokenizer.ca
>
>
>
>
>
>
>
>
> On 11-07-29 7:49 PM, "Otis Gospodnetic" <[EMAIL PROTECTED]> wrote:
>
>>Hi,
>>
>>I'm for publishing all performance metrics in JMX (in addition to
>>exposing it wherever else you guys decide).  That's because JMX is
>>probably the easiest for our SPM for HBase [1] to get to HBase
>>performance metrics and I suspect we are not alone.
>>
>>Otis
>>[1] http://sematext.com/spm/hbase-performance-monitoring/index.html
>>----
>>Sematext :: http://sematext.com/ :: Solr - Lucene - Hadoop - HBase
>>Hadoop ecosystem search :: http://search-hadoop.com/
>>
>>
>>
>>>________________________________
>>>From: Andrew Purtell <[EMAIL PROTECTED]>
>>>To: Doug Meil <[EMAIL PROTECTED]>; "[EMAIL PROTECTED]"
>>><[EMAIL PROTECTED]>
>>>Sent: Friday, July 29, 2011 4:34 PM
>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>
>>>> I'd rather see this output being able to be captured by something the
>>>>sink that Todd suggested, rather than focusing on shell access.
>>>
>>>
>>>I don't agree.
>>>
>>>
>>>Look at what we have existing and proposed:
>>>
>>>    - Java API access to server and region load information, that the
>>>shell uses
>>>
>>>    - A proposal to dump some stats into log files, that then has to be
>>>scraped
>>>
>>>    - A proposal (by the FB guys) to export some JSON via a HTTP servlet
>>>
>>>This is not good design, this is a bunch of random shit stuck together.
>>>
>>>Note that what Todd proposed does not preclude adding Java client API
>>>support for retrieving it.
>>>
>>>At a minimum all of this information must be accessible via the Java
>>>client API, to enable programmatic monitoring and analysis use cases.
>>>I'll add the shell support if nobody else cares about it, that is a
>>>relatively small detail, but one I think is important.
>>>
>>>Best regards,
>>>
>>>
>>>    - Andy
>>>
>>>
>>>Problems worthy of attack prove their worth by hitting back. - Piet Hein
>>>(via Tom White)
>>>
>>>
>>>>________________________________
>>>>From: Doug Meil <[EMAIL PROTECTED]>
>>>>To: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>;
>>>>"[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
>>>>Sent: Friday, July 29, 2011 11:39 AM
>>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>>
>>>>
>>>>I'd rather see this output being able to be captured by something the
>>>>sink
>>>>that Todd suggested, rather than focusing on shell access.  HServerLoad
>>>>is
>>>>super-summary at the RS level, and both the items in 4089 and 4147 are
>>>>proposed to be "summarized" but still have reasonable detail (e.g., even
>>>>table/CF summary there could be dozens of entries given a reasonably
>>>>complex system).
>>>>
>>>>
>>>>
>>>>
>>>>On 7/29/11 1:15 PM, "Andrew Purtell" <[EMAIL PROTECTED]> wrote:
>>>>
>>>>>There is also the matter of HServerLoad and how that is used by the
>>>>>shell
>>>>>and master UI to report on cluster status.
>>>>>
>>>>>I'd like the shell to be able to let the user explore all of these
>>>>>different reports interactively.
>>>>>
>>>>>At the very least, they should all be handled the same way.
>>>>>
>>>>>And then there is Riley's work over at FB on a slow query log. How does
>>>>>that fit in?
>>>>>
>>>>>Best regards,
>>>>>
>>>>>
>>>>>   - Andy
>>>>>
>>>>>Problems worthy of attack prove their worth by hitting back. - Piet
>>>>>Hein
>>>>>(via Tom White)
>>>>>
>>>>>
>>>>>>________________________________
>>>>>>From: Todd Lipcon <[EMAIL PROTECTED]>
>>>>>>To: [EMAIL PROTECTED]
>>>>>>Sent: Friday, July 29, 2011 9:58 AM

Great,

My VP is former MS-98 :)))

(BTW, many thanks to http://www.cloudera.com/ employees; please make Whirr
really "rrrrrŠŠŠ" "open source"? I don't see any meaningŠ reinventing a
bike is cheaper nowadays!!!)
And, SEO of course: (remember: Google is SMARTER!!! You are just "clone".)

=========================================================================

--
Fuad Efendi
416-993-2060
Tokenizer Inc., Canada
Data Mining, Search Engines
http://www.tokenizer.ca
On 11-07-31 6:21 PM, "Ryan Rawson" <[EMAIL PROTECTED]> wrote:

>You should ask for your money back!!
>
>On Sun, Jul 31, 2011 at 3:10 PM, Fuad Efendi <[EMAIL PROTECTED]> wrote:
>> What is it all about? HBase sucks. Too many problems to newcomers,
>> few-weeks-warm-up to begin with!!!!!!!!!!!!!!!! Is it really-really
>> supported by Microsoft employees?!
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> And, SEO of course:
>> ==================>>
>>
>> --
>> Fuad Efendi
>> 416-993-2060
>> Tokenizer Inc., Canada
>> Data Mining, Search Engines
>> http://www.tokenizer.ca
>>
>>
>>
>>
>>
>>
>>
>>
>> On 11-07-29 7:49 PM, "Otis Gospodnetic" <[EMAIL PROTECTED]>
>>wrote:
>>
>>>Hi,
>>>
>>>I'm for publishing all performance metrics in JMX (in addition to
>>>exposing it wherever else you guys decide).  That's because JMX is
>>>probably the easiest for our SPM for HBase [1] to get to HBase
>>>performance metrics and I suspect we are not alone.
>>>
>>>Otis
>>>[1] http://sematext.com/spm/hbase-performance-monitoring/index.html
>>>----
>>>Sematext :: http://sematext.com/ :: Solr - Lucene - Hadoop - HBase
>>>Hadoop ecosystem search :: http://search-hadoop.com/
>>>
>>>
>>>
>>>>________________________________
>>>>From: Andrew Purtell <[EMAIL PROTECTED]>
>>>>To: Doug Meil <[EMAIL PROTECTED]>; "[EMAIL PROTECTED]"
>>>><[EMAIL PROTECTED]>
>>>>Sent: Friday, July 29, 2011 4:34 PM
>>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>>
>>>>> I'd rather see this output being able to be captured by something the
>>>>>sink that Todd suggested, rather than focusing on shell access.
>>>>
>>>>
>>>>I don't agree.
>>>>
>>>>
>>>>Look at what we have existing and proposed:
>>>>
>>>>    - Java API access to server and region load information, that the
>>>>shell uses
>>>>
>>>>    - A proposal to dump some stats into log files, that then has to be
>>>>scraped
>>>>
>>>>    - A proposal (by the FB guys) to export some JSON via a HTTP
>>>>servlet
>>>>
>>>>This is not good design, this is a bunch of random shit stuck together.
>>>>
>>>>Note that what Todd proposed does not preclude adding Java client API
>>>>support for retrieving it.
>>>>
>>>>At a minimum all of this information must be accessible via the Java
>>>>client API, to enable programmatic monitoring and analysis use cases.
>>>>I'll add the shell support if nobody else cares about it, that is a
>>>>relatively small detail, but one I think is important.
>>>>
>>>>Best regards,
>>>>
>>>>
>>>>    - Andy
>>>>
>>>>
>>>>Problems worthy of attack prove their worth by hitting back. - Piet
>>>>Hein
>>>>(via Tom White)
>>>>
>>>>
>>>>>________________________________
>>>>>From: Doug Meil <[EMAIL PROTECTED]>
>>>>>To: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>;
>>>>>"[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
>>>>>Sent: Friday, July 29, 2011 11:39 AM
>>>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>>>
>>>>>
>>>>>I'd rather see this output being able to be captured by something the
>>>>>sink
>>>>>that Todd suggested, rather than focusing on shell access.
>>>>>HServerLoad
>>>>>is
>>>>>super-summary at the RS level, and both the items in 4089 and 4147 are
>>>>>proposed to be "summarized" but still have reasonable detail (e.g.,
>>>>>even
>>>>>table/CF summary there could be dozens of entries given a reasonably
>>>>>complex system).
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>On 7/29/11 1:15 PM, "Andrew Purtell" <[EMAIL PROTECTED]> wrote:
>>>>>
>>>>>>There is also the matter of HServerLoad and how that is used by the

re:  "Whirr"

I'm sure Cloudera would love to hear your feedback on Whirr, but please
address it to them directly and not on the Hbase dist-list.

re:  "Is it really-really supported by Microsoft employees?!"

It is really, really not.
As you pointed out, and as is cited in the Apache Hbase book, Hbase was
inspired from BigTable.  And as for the refund that Ryan suggested I'm
sure Google would be happy to make good on it.  Watch out for the exchange
rate at the border.
On 7/31/11 6:38 PM, "Fuad Efendi" <[EMAIL PROTECTED]> wrote:

>Great,
>
>
>
>My VP is former MS-98 :)))
>
>
>
>(BTW, many thanks to http://www.cloudera.com/ employees; please make Whirr
>really "rrrrrŠŠŠ" "open source"? I don't see any meaningŠ reinventing a
>bike is cheaper nowadays!!!)
>
>
>
>
>
>
>And, SEO of course: (remember: Google is SMARTER!!! You are just "clone".)
>
>=========================================================================>
>
>--
>Fuad Efendi
>416-993-2060
>Tokenizer Inc., Canada
>Data Mining, Search Engines
>http://www.tokenizer.ca
>
>
>
>
>
>
>
>
>On 11-07-31 6:21 PM, "Ryan Rawson" <[EMAIL PROTECTED]> wrote:
>
>>You should ask for your money back!!
>>
>>On Sun, Jul 31, 2011 at 3:10 PM, Fuad Efendi <[EMAIL PROTECTED]> wrote:
>>> What is it all about? HBase sucks. Too many problems to newcomers,
>>> few-weeks-warm-up to begin with!!!!!!!!!!!!!!!! Is it really-really
>>> supported by Microsoft employees?!
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> And, SEO of course:
>>> ==================>>>
>>>
>>> --
>>> Fuad Efendi
>>> 416-993-2060
>>> Tokenizer Inc., Canada
>>> Data Mining, Search Engines
>>> http://www.tokenizer.ca
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> On 11-07-29 7:49 PM, "Otis Gospodnetic" <[EMAIL PROTECTED]>
>>>wrote:
>>>
>>>>Hi,
>>>>
>>>>I'm for publishing all performance metrics in JMX (in addition to
>>>>exposing it wherever else you guys decide).  That's because JMX is
>>>>probably the easiest for our SPM for HBase [1] to get to HBase
>>>>performance metrics and I suspect we are not alone.
>>>>
>>>>Otis
>>>>[1] http://sematext.com/spm/hbase-performance-monitoring/index.html
>>>>----
>>>>Sematext :: http://sematext.com/ :: Solr - Lucene - Hadoop - HBase
>>>>Hadoop ecosystem search :: http://search-hadoop.com/
>>>>
>>>>
>>>>
>>>>>________________________________
>>>>>From: Andrew Purtell <[EMAIL PROTECTED]>
>>>>>To: Doug Meil <[EMAIL PROTECTED]>; "[EMAIL PROTECTED]"
>>>>><[EMAIL PROTECTED]>
>>>>>Sent: Friday, July 29, 2011 4:34 PM
>>>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>>>
>>>>>> I'd rather see this output being able to be captured by something
>>>>>>the
>>>>>>sink that Todd suggested, rather than focusing on shell access.
>>>>>
>>>>>
>>>>>I don't agree.
>>>>>
>>>>>
>>>>>Look at what we have existing and proposed:
>>>>>
>>>>>    - Java API access to server and region load information, that the
>>>>>shell uses
>>>>>
>>>>>    - A proposal to dump some stats into log files, that then has to
>>>>>be
>>>>>scraped
>>>>>
>>>>>    - A proposal (by the FB guys) to export some JSON via a HTTP
>>>>>servlet
>>>>>
>>>>>This is not good design, this is a bunch of random shit stuck
>>>>>together.
>>>>>
>>>>>Note that what Todd proposed does not preclude adding Java client API
>>>>>support for retrieving it.
>>>>>
>>>>>At a minimum all of this information must be accessible via the Java
>>>>>client API, to enable programmatic monitoring and analysis use cases.
>>>>>I'll add the shell support if nobody else cares about it, that is a
>>>>>relatively small detail, but one I think is important.
>>>>>
>>>>>Best regards,
>>>>>
>>>>>
>>>>>    - Andy
>>>>>
>>>>>
>>>>>Problems worthy of attack prove their worth by hitting back. - Piet
>>>>>Hein
>>>>>(via Tom White)
>>>>>
>>>>>
>>>>>>________________________________
>>>>>>From: Doug Meil <[EMAIL PROTECTED]>
>>>>>>To: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>;
>>>>>>"[EMAIL PROTECTED]" <[EMAIL PROTECTED]>

This is the first time I've ever seen a thread here go off into what the fuckville. Bookmarked.
 
    - Andy

Problems worthy of attack prove their worth by hitting back. - Piet Hein (via Tom White)
>________________________________
>From: Doug Meil <[EMAIL PROTECTED]>
>To: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
>Sent: Sunday, July 31, 2011 4:17 PM
>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>
>
>re:  "Whirr"
>
>I'm sure Cloudera would love to hear your feedback on Whirr, but please
>address it to them directly and not on the Hbase dist-list.
>
>re:  "Is it really-really supported by Microsoft employees?!"
>
>It is really, really not.
>
>
>As you pointed out, and as is cited in the Apache Hbase book, Hbase was
>inspired from BigTable.  And as for the refund that Ryan suggested I'm
>sure Google would be happy to make good on it.  Watch out for the exchange
>rate at the border.
>
>
>
>
>On 7/31/11 6:38 PM, "Fuad Efendi" <[EMAIL PROTECTED]> wrote:
>
>>Great,
>>
>>
>>
>>My VP is former MS-98 :)))
>>
>>
>>
>>(BTW, many thanks to http://www.cloudera.com/ employees; please make Whirr
>>really "rrrrrŠŠŠ" "open source"? I don't see any meaningŠ reinventing a
>>bike is cheaper nowadays!!!)
>>
>>
>>
>>
>>
>>
>>And, SEO of course: (remember: Google is SMARTER!!! You are just "clone".)
>>
>>=========================================================================>>
>>
>>--
>>Fuad Efendi
>>416-993-2060
>>Tokenizer Inc., Canada
>>Data Mining, Search Engines
>>http://www.tokenizer.ca
>>
>>
>>
>>
>>
>>
>>
>>
>>On 11-07-31 6:21 PM, "Ryan Rawson" <[EMAIL PROTECTED]> wrote:
>>
>>>You should ask for your money back!!
>>>
>>>On Sun, Jul 31, 2011 at 3:10 PM, Fuad Efendi <[EMAIL PROTECTED]> wrote:
>>>> What is it all about? HBase sucks. Too many problems to newcomers,
>>>> few-weeks-warm-up to begin with!!!!!!!!!!!!!!!! Is it really-really
>>>> supported by Microsoft employees?!
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> And, SEO of course:
>>>> ==================>>>>
>>>>
>>>> --
>>>> Fuad Efendi
>>>> 416-993-2060
>>>> Tokenizer Inc., Canada
>>>> Data Mining, Search Engines
>>>> http://www.tokenizer.ca
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On 11-07-29 7:49 PM, "Otis Gospodnetic" <[EMAIL PROTECTED]>
>>>>wrote:
>>>>
>>>>>Hi,
>>>>>
>>>>>I'm for publishing all performance metrics in JMX (in addition to
>>>>>exposing it wherever else you guys decide).  That's because JMX is
>>>>>probably the easiest for our SPM for HBase [1] to get to HBase
>>>>>performance metrics and I suspect we are not alone.
>>>>>
>>>>>Otis
>>>>>[1] http://sematext.com/spm/hbase-performance-monitoring/index.html
>>>>>----
>>>>>Sematext :: http://sematext.com/ :: Solr - Lucene - Hadoop - HBase
>>>>>Hadoop ecosystem search :: http://search-hadoop.com/
>>>>>
>>>>>
>>>>>
>>>>>>________________________________
>>>>>>From: Andrew Purtell <[EMAIL PROTECTED]>
>>>>>>To: Doug Meil <[EMAIL PROTECTED]>; "[EMAIL PROTECTED]"
>>>>>><[EMAIL PROTECTED]>
>>>>>>Sent: Friday, July 29, 2011 4:34 PM
>>>>>>Subject: Re: HBASE-4089 & HBASE-4147 - on the topic of ops output
>>>>>>
>>>>>>> I'd rather see this output being able to be captured by something
>>>>>>>the
>>>>>>>sink that Todd suggested, rather than focusing on shell access.
>>>>>>
>>>>>>
>>>>>>I don't agree.
>>>>>>
>>>>>>
>>>>>>Look at what we have existing and proposed:
>>>>>>
>>>>>>    - Java API access to server and region load information, that the
>>>>>>shell uses
>>>>>>
>>>>>>    - A proposal to dump some stats into log files, that then has to
>>>>>>be
>>>>>>scraped
>>>>>>
>>>>>>    - A proposal (by the FB guys) to export some JSON via a HTTP
>>>>>>servlet
>>>>>>
>>>>>>This is not good design, this is a bunch of random shit stuck
>>>>>>together.
>>>>>>
>>>>>>Note that what Todd proposed does not preclude adding Java client API
>>>>>>support for retrieving it.
>>>>>>
>>>>>>At a minimum all of this information must be accessible via the Java

>
>Doug, sorry for off-topic, it was extremely hot day; and I finally found
>that it is cheaper to get single dedicated cc1.4xlarge (EC2) for
>$1.60/hour than to rely on advocated (Whirr and Friends) c1.xlarge
>($1.68/hour); unbelievable (hot!) 10 minutes for garbage collection of 1
>Gb heap only, 1000 (small!!!) transactions PER MINUTE only. It can be
>explained by "virtual" sharing of the same cc1.4xlarge with at least two
>of other users (3 x $1.68 = $2.04) . Especially "swap", "cache" is used
>instead of file, and (of course) it is slower.
No any problem with Hadoop thoughŠ

BTW, I need to compare with Cassandra, another Big Table clone.

(P.S. I used dedicated servers before, and never had problems with Hbase;
except EC2)

-Fuad

Cassandra is not a big table clone.

Bigtable specifies 2 things:
- a data model with column families, etc
- a distribution model, a distributed architecture, etc

Only HBase provides both these things.  Cassandra provides elements of
(a) and uses the dynamo model to do distribution.

I believe in HBase because I think it's distribution model is simpler,
more robust and reliable than the dynamo model.

On Mon, Aug 1, 2011 at 7:40 PM, Fuad Efendi <[EMAIL PROTECTED]> wrote:
>>
>>Doug, sorry for off-topic, it was extremely hot day; and I finally found
>>that it is cheaper to get single dedicated cc1.4xlarge (EC2) for
>>$1.60/hour than to rely on advocated (Whirr and Friends) c1.xlarge
>>($1.68/hour); unbelievable (hot!) 10 minutes for garbage collection of 1
>>Gb heap only, 1000 (small!!!) transactions PER MINUTE only. It can be
>>explained by "virtual" sharing of the same cc1.4xlarge with at least two
>>of other users (3 x $1.68 = $2.04) . Especially "swap", "cache" is used
>>instead of file, and (of course) it is slower.
> No any problem with Hadoop thoughŠ
>
> BTW, I need to compare with Cassandra, another Big Table clone.
>
> (P.S. I used dedicated servers before, and never had problems with Hbase;
> except EC2)
>
> -Fuad
>
>
>

>
>
>
>re:  "Is it really-really supported by Microsoft employees?!"
>
>It is really, really not.
>
I believe Hbase was contributed to Apache by a Powerset, currently owned
by Microsoft; and (same contributors) were full-time supporting Hbase and
having salaries from Microsoft for at least a year; it was first
(implicit) contribution from Microsoft to Apache.

No one at powerset is currently contributing to HBase.  "Was" is the
key here - in the past, etc.

I guess MSFT never got to integrating the HBase API with the C# LINQ
system and Visual Studio.  Maybe it's that azure table services?

On Mon, Aug 1, 2011 at 7:50 PM, Fuad Efendi <[EMAIL PROTECTED]> wrote:
>>
>>
>>
>>re:  "Is it really-really supported by Microsoft employees?!"
>>
>>It is really, really not.
>>
>
>
> I believe Hbase was contributed to Apache by a Powerset, currently owned
> by Microsoft; and (same contributors) were full-time supporting Hbase and
> having salaries from Microsoft for at least a year; it was first
> (implicit) contribution from Microsoft to Apache.
>
>
>

Yes.  Hbase was started by powerset and two of us contributed first as powerset employees and then later after our acquisition by microsoft and a pause of a few months while some interesting legal issues were worked through, i resumed contributing to apache hbase as an employee of Microsoft. A bunch of backscenes work by mgmnt at powerset and by the ms integration team made what I thought the impossible happen (at the time I was highly skeptical and was out talking to all and sundry looking for someone else to assume hbase sponsorship).   I do not think our work was the first open source ms contrib -- I seem to remember fixes for some sqlserver open source driver happening before we were let resume -- but I am pretty sure it was the most sustained (I was near full time working on hbase).   I remained a ms employee free to contrib to hbase, hadoop itself had not been vetted by ms legal so was off limits, for over a year until the day we all realized hbase no longer fit the direction in which the powerset division was headed.  Then I moved on.

Microsoft  sponsored early hbase dev.

Stack

On Aug 1, 2011, at 19:50, Fuad Efendi <[EMAIL PROTECTED]> wrote:

>>
>>
>>
>> re:  "Is it really-really supported by Microsoft employees?!"
>>
>> It is really, really not.
>>
>
>
> I believe Hbase was contributed to Apache by a Powerset, currently owned
> by Microsoft; and (same contributors) were full-time supporting Hbase and
> having salaries from Microsoft for at least a year; it was first
> (implicit) contribution from Microsoft to Apache.
>
>