Mailing-List: contact commits-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cassandra.apache.org
Date: Mon, 28 Apr 2014 21:25:15 +0000 (UTC)
From: "Martin Bligh (JIRA)" <jira@apache.org>
To: commits@cassandra.apache.org
Message-ID: <JIRA.12711022.1398719936047.194435.1398720315522@arcas>
In-Reply-To: <JIRA.12711022.1398719936047@arcas>
References: <JIRA.12711022.1398719936047@arcas>
Subject: [jira] [Commented] (CASSANDRA-7103) Very poor performance with
 simple setup
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/CASSANDRA-7103?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13983567#comment-13983567 ] 

Martin Bligh commented on CASSANDRA-7103:
-----------------------------------------

BTW, I'm aware that this is
(a) a single node
(b) half my time_order entries are zero (which I don't think matters as it's a single node anyway), so my partition key doesn't have much variance
(c) disk is not performant (but we're not even trying to write to it from iostat, so I don't think this matters).
(d) I'm writing to one table
(e) I'm using a single writer.

So I'm creating a "hotspot" of some form. But really, 

1. I think it should be able to handle more than 700 writes a second to one table.
2. It shouldn't degrade to about 10 writes per second.

Sure, I could throw masses of hardware at it, and make it scale a bit better, but ... unless it can perform better than this on a single table, single node, I don't see how it'd perform in any reasonable fashion on a larger cluster.


> Very poor performance with simple setup
> ---------------------------------------
>
>                 Key: CASSANDRA-7103
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7103
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>         Environment: Fedora 19 (also happens on Ubuntu), Cassandra 2.0.7. dsc standard install
>            Reporter: Martin Bligh
>
> Single node (this is just development, 32GB 20 core server), single disk array.
> Create the following table:
> CREATE TABLE reut (
>   time_order bigint,
>   time_start bigint,
>   ack_us map<int, int>,
>   gc_strategy map<text, int>,
>   gc_strategy_symbol map<text, int>,
>   gc_symbol map<text, int>,
>   ge_strategy map<text, int>,
>   ge_strategy_symbol map<text, int>,
>   ge_symbol map<text, int>,
>   go_strategy map<text, int>,
>   go_strategy_symbol map<text, int>,
>   go_symbol map<text, int>,
>   message_type map<text, int>,
>   PRIMARY KEY (time_order, time_start)
> ) WITH
>   bloom_filter_fp_chance=0.010000 AND
>   caching='KEYS_ONLY' AND
>   comment='' AND
>   dclocal_read_repair_chance=0.000000 AND
>   gc_grace_seconds=864000 AND
>   index_interval=128 AND
>   read_repair_chance=0.100000 AND
>   replicate_on_write='true' AND
>   populate_io_cache_on_flush='false' AND
>   default_time_to_live=0 AND
>   speculative_retry='99.0PERCENTILE' AND
>   memtable_flush_period_in_ms=0 AND
>   compaction={'class': 'SizeTieredCompactionStrategy'} AND
>   compression={};
> Now I just insert data into it (using python driver, async insert, prepared insert statement). Each row only fills out one of the gc_*, go_*, or ge_* columns, and there's something like 20-100 entries per map column, occasionally 1000, but it's nothing huge. 
> First run 685 inserts in 1.004860 seconds (681.687053 Operations/s).
> OK, not great, but that's fine.
> Now throw 50,000 rows at it.
> Now run the first run again, and it takes 53s to do the same insert of 685 rows - I'm getting about 10 rows per second. 
> It's not IO bound - "iostat 1" shows quiescent for 9 seconds, then ~640KB write, then sleeps again - seems like the fflush sync.
> Run "nodetool flush" and performance goes back to as before!!!!
> Not sure why this gets so slow - I think it just builds huge commit logs and memtables, but never writes out to the data/ directory with sstables because I only have one table? That doesn't seem like a good situation. 
> Worse ... if you let the python driver just throw stuff at it async (I think this allows up to 128 request if I understand the underlying protocol, then it gets so slow that a single write takes over 10s and times out). Seems to be some sort of synchronization problem in Java ... if I limit the concurrent async requests to the left column below, I get the number of seconds elapsed on the right:
> 1: 103 seconds
> 2: 63 seconds
> 8: 53 seconds
> 16: 53 seconds
> 32: 66 seconds
> 64: so slow it explodes in timeouts on write (over 10s each).
> I guess there's some thundering herd type locking issue in whatever Java primitive you are using to lock concurrent access to a single table. I know some of the Java concurrent.* stuff has this issue. So for the other tests above, I was limiting async writes to 16 pending.


--
This message was sent by Atlassian JIRA
(v6.2#6252)