Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Accumulo >> mail # dev >> multi-table isolated batch scanner


Copy link to this message
-
Re: multi-table isolated batch scanner
On Mon, Apr 15, 2013 at 6:19 PM, Adam Fuchs <[EMAIL PROTECTED]> wrote:
> Keith,
>
> In this case we're filling the buffer before we can amortize the search
> cost. We're using a document-partitioned table design and we have to do a
> local sort before we can get the first result.
>

I am not sure exactly what you are doing.  To me it sounds like you
may be doing the following, but not sure.

 * Seeking N iterators
 * Doing what the intersecting iterator does, joining docids or
seeking iterators
 * Collecting a set of key values (docids or documents?) and sorting
them?  How much is collected before sort?  Why sort?  Is filtering
done after the sort?

Or did you meant something else by 'local sort' ? like a sort on the
client side?  But this does not seem the entire story as you mentioned
something about multiple passes.  Are you hash partitioning the data?
Are you building something in memoery besides the buffered output
data?
> I have found that increasing the buffer size also increases the latency for
> getting the first results. This application is both latency and throughput
> sensitive. In addition, increasing the batch size too much puts significant
> memory requirements on the process running the batch scanner.
>
> Adam
>
>
>
> On Mon, Apr 15, 2013 at 5:33 PM, Keith Turner <[EMAIL PROTECTED]> wrote:
>
>> On Mon, Apr 15, 2013 at 5:06 PM, Adam Fuchs <[EMAIL PROTECTED]> wrote:
>> > Chris,
>> >
>> > The desire for isolation stems from the desire to amortize some
>> computation
>> > over a number of results. Say it takes 5 seconds to compute an
>> intersection
>>
>> Would increasing the size of the key/value buffer help in your case?
>> The iterator stack is not torn down until that buffer fills up or the
>> end of tablet is reached.  Are you concerned about the cost of
>> reconstructing the iterator stack across tablets?
>>
>> > of a couple of sets within the iterators, and then streaming back the
>> > results takes a minute or so. If I have to redo the 5 second computation
>> > many times, as in to support the reconstruction of the iterator tree,
>> then
>> > that computation may start to dominate my query performance. Primarily,
>> > this means I need to be able to continue a scan without having to rebuild
>> > the iterators. Isolation in the scanner has that side effect. Proper
>> > isolation would be a "nice-to-have", but I can deal with not having it.
>> >
>> > Adam
>> >
>> >
>> >
>> > On Mon, Apr 15, 2013 at 4:13 PM, Christopher <[EMAIL PROTECTED]>
>> wrote:
>> >
>> >> Adam-
>> >>
>> >> It seems like you're talking about two features at once:
>> >> 1) Multi-table batch scanner.
>> >> 2) Scan Isolation on batch scanners like we have on regular scanners.
>> >> Is that correct?
>> >>
>> >> I can see the utility of a multi-table batch scanner, but I haven't
>> >> seen a compelling need for implementing isolation on the
>> >> batch-scanners. Do you have a use case in mind for that?
>> >>
>> >> Also, it seems that your use case for isolation is not so much the
>> >> isolated reads, but the statefulness of the iterator stack on the
>> >> server side. Is this correct? If so, I'm even more curious about your
>> >> use case for this, since that statefulness is only guaranteed per-row.
>> >>
>> >>
>> >> --
>> >> Christopher L Tubbs II
>> >> http://gravatar.com/ctubbsii
>> >>
>> >>
>> >> On Mon, Apr 15, 2013 at 3:10 PM, Adam Fuchs <[EMAIL PROTECTED]> wrote:
>> >> > Thanks Bill,
>> >> >
>> >> > I care about latency and throughput. First available result ordering
>> is
>> >> > fine, though.
>> >> >
>> >> > Does Guava just chain through a collection of iterators, completing
>> one
>> >> > then moving to the next?
>> >> >
>> >> > Adam
>> >> >
>> >> >
>> >> >
>> >> > On Mon, Apr 15, 2013 at 3:06 PM, William Slacum <
>> >> > [EMAIL PROTECTED]> wrote:
>> >> >
>> >> >> How are you expecting to get results back? Guava's Iterables could
>> >> concat a
>> >> >> bunch of a Scanners together, if you didn't care about the throughput