Can't get performance for scanning the column that is not row key.

[Sungju Hong <eq...@gmail.com>]

Hello,

I have cassandra 1.2 and about fifty nodes.

I sometime meet requirements for scanning the column that is not row key
fast.
I know that secondary index doesn't help.

someone advised to read sstables directly. is it helpful?

I will appreciate for any idea handling this topic.

Thanks.

Re: Can't get performance for scanning the column that is not row key.

[Alain RODRIGUEZ <ar...@gmail.com>]

Wow.

I have cassandra 1.2 and about fifty nodes.


This Cassandra version is no longer supported, 2.0 isn't either and 2.1 is
reaching end of life...

You could use materialised views, SASI indexes or this kind of new
features. Also CQL performances are better in recent versions etc.

You don't expose data about the table you want to read and how you want to
read it, it is hard to tell.

But I am almost sure you could take profit of some new features, bug fixes
and optimisation. Plus I would bet that no one is going to open the
Cassandra 1.2 code anymore to help you. So if you want free or paid
support, go to 2.1 at least, and keep in mind this one is EOL too...

someone advised to read sstables directly. is it helpful?


It would probably be harder than upgrading... And that's what Cassandra
does. I doubt you will be able to do it better than Cassandra does,
excepted, maybe on some rare corner cases...

Good luck,
C*heers,
-----------------------
Alain Rodriguez - alain@thelastpickle.com
France

The Last Pickle - Apache Cassandra Consulting
http://www.thelastpickle.com

2016-07-07 2:39 GMT+02:00 Sungju Hong <eq...@gmail.com>:

> Hello,
>
> I have cassandra 1.2 and about fifty nodes.
>
> I sometime meet requirements for scanning the column that is not row key
> fast.
> I know that secondary index doesn't help.
>
> someone advised to read sstables directly. is it helpful?
>
> I will appreciate for any idea handling this topic.
>
> Thanks.
>
>
>