accumulo-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Keith Turner <ke...@deenlo.com>
Subject Re: Feedback about techniques for tuning batch scanning for my problem
Date Mon, 23 May 2016 16:11:14 GMT
Until ACCUMULO-4164[1] is released you may see a performance difference
between 1 and 2 rfile level indexes.  However this depends on your #of
seeks to scan ratio.  If you seek to a spot and read a good bit of data
from there, it probably will not matter.

Accumulo caches a certain number of open rfiles[2].  For these cached open
rfiles the root level node is kept in memory. However, levels of the index
below the root are always read from index cache.  This is where the change
in ACCUMULO-4164 to avoid a copy from index cache helps.  So you may want
to consider ensuring the caching of open rfile is large enough and that the
index depth is one (for now).

[1]: https://issues.apache.org/jira/browse/ACCUMULO-4164
[2]:
http://accumulo.apache.org/1.7/accumulo_user_manual#_tserver_scan_files_open_max

On Mon, May 23, 2016 at 11:48 AM, Mario Pastorelli <
mario.pastorelli@teralytics.ch> wrote:

> This is interesting and I think that's the main parameters that I need to
> tune. We have a lot of memory to spare for Accumulo so the index should fix
> in memory, but I'll check when I change those parameters. With small
> records, I think it makes sense to have smaller blocks for faster lookups.
> Thank you!
>
> On Mon, May 23, 2016 at 5:16 PM, Keith Turner <keith@deenlo.com> wrote:
>
>> Mario
>>
>> You could experiment with adjusting table.file.compress.blocksize[1] and
>> table.file.compress.blocksize.index[2].  Making the data block size smaller
>> will increase the files index size and may lower random seek times
>> depending on how much of the index fits in memory.  Adjusting the index
>> block size will control the depth of the index tree.   You can adjust these
>> settings, compact, and run "accumulo rfile-info" on a file to see
>> information about the index size, depth, and number of data blocks.  You
>> can also adjust the index[6] and data[5] cache size settings and turn
>> data[3] and index[4] caching on or off for your table.
>>
>> [1]:
>> http://accumulo.apache.org/1.7/accumulo_user_manual#_table_file_compress_blocksize
>> [2]:
>> http://accumulo.apache.org/1.7/accumulo_user_manual#_table_file_compress_blocksize_index
>> [3]:
>> http://accumulo.apache.org/1.7/accumulo_user_manual#_table_cache_block_enable
>> [4]:
>> http://accumulo.apache.org/1.7/accumulo_user_manual#_table_cache_index_enable
>> [5]:
>> http://accumulo.apache.org/1.7/accumulo_user_manual#_tserver_cache_data_size
>> [6]:
>> http://accumulo.apache.org/1.7/accumulo_user_manual#_tserver_cache_index_size
>>
>> Keith
>>
>> On Thu, May 19, 2016 at 11:08 AM, Mario Pastorelli <
>> mario.pastorelli@teralytics.ch> wrote:
>>
>>> Hey people,
>>> I'm trying to tune a bit the query performance to see how fast it can go
>>> and I thought it would be great to have comments from the community. The
>>> problem that I'm trying to solve in Accumulo is the following: we want to
>>> store the entities that have been in a certain location in a certain day.
>>> The location is a Long and the entity id is a Long. I want to be able to
>>> scan ~1M of rows in few seconds, possibly less than one. Right now, I'm
>>> doing the following things:
>>>
>>>    1. I'm using a sharding byte at the start of the rowId to keep the
>>>    data in the same range distributed in the cluster
>>>    2. all the records are encoded, one single record is composed by
>>>       1. rowId: 1 shard byte + 3 bytes for the day
>>>       2. column family: 8 byte for the long corresponding to the hash
>>>       of the location
>>>       3. column qualifier: 8 byte corresponding to the identifier of
>>>       the entity
>>>       4. value: 2 bytes for some additional information
>>>    3. I use a batch scanner because I don't need sorting and it's faster
>>>
>>> As expected, it takes few seconds to scan 1M rows but now I'm wondering
>>> if I can improve it. My ideas are the following:
>>>
>>>    1. set table.compaction.major.ration to 1 because I don't care about
>>>    the ingestion performance and this should improve the query performance
>>>    2. pre-split tables to match the number of servers and then use a
>>>    byte of shard as first byte of the rowId. This should improve both writing
>>>    and reading the data because both should work in parallel for what I
>>>    understood
>>>    3. enable bloom filter on the table
>>>
>>> Do you think those ideas make sense? Furthermore, I have two questions:
>>>
>>>    1. considering that a single entry is only 22 bytes but I'm going to
>>>    scan ~1M records per query, do you think I should change the BatchScanner
>>>    buffers somehow?
>>>    2. anything else to improve the scan speed? Again, I don't care
>>>    about the ingestion time
>>>
>>> Thanks for the help!
>>>
>>> --
>>> Mario Pastorelli | TERALYTICS
>>>
>>> *software engineer*
>>>
>>> Teralytics AG | Zollstrasse 62 | 8005 Zurich | Switzerland
>>> phone: +41794381682
>>> email: mario.pastorelli@teralytics.ch
>>> www.teralytics.net
>>>
>>> Company registration number: CH-020.3.037.709-7 | Trade register Canton
>>> Zurich
>>> Board of directors: Georg Polzer, Luciano Franceschina, Mark Schmitz,
>>> Yann de Vries
>>>
>>> This e-mail message contains confidential information which is for the
>>> sole attention and use of the intended recipient. Please notify us at once
>>> if you think that it may not be intended for you and delete it immediately.
>>>
>>
>>
>
>
> --
> Mario Pastorelli | TERALYTICS
>
> *software engineer*
>
> Teralytics AG | Zollstrasse 62 | 8005 Zurich | Switzerland
> phone: +41794381682
> email: mario.pastorelli@teralytics.ch
> www.teralytics.net
>
> Company registration number: CH-020.3.037.709-7 | Trade register Canton
> Zurich
> Board of directors: Georg Polzer, Luciano Franceschina, Mark Schmitz, Yann
> de Vries
>
> This e-mail message contains confidential information which is for the
> sole attention and use of the intended recipient. Please notify us at once
> if you think that it may not be intended for you and delete it immediately.
>

Mime
View raw message