[prev in list] [next in list] [prev in thread] [next in thread] 

List:       cassandra-dev
Subject:    Re: lazy boy example using column family
From:       Michael Greene <michael.greene () gmail ! com>
Date:       2009-07-27 16:31:25
Message-ID: f4d6a21a0907270931r3185807boae11bde5aa2c680d () mail ! gmail ! com
[Download RAW message or body]

http://issues.apache.org/jira/browse/CASSANDRA-44 is the tracking
issue for making column families easier to modify in an existing
cluster.

Michael

On Mon, Jul 27, 2009 at 11:13 AM, Evan Weaver<eweaver@gmail.com> wrote:
> Transferring this over to dev.
>
> I think it's unacceptable to have to restart the entire cluster under
> any circumstances aside from code upgrades, even if the operation in
> question is "rare". Note that in column-oriented use, adding or
> removing a CF is not rare at all.
>
> Let's not turn a bug into a feature, here. More usability is never bad.
>
> Evan
>
> On Mon, Jul 27, 2009 at 12:38 AM, Mark Robson<markxr@gmail.com> wrote:
>>
>>
>> 2009/7/27 <mobiledreamers@gmail.com>
>>>
>>> i m trying to use cassandra in a mode where everytime i create a new
>>> columnfamily i do not want to restart all the nodes
>>
>> In my opinion you should not be doing that anyway.
>>
>> Because families can have as many columns as you like anyway, it should not
>> normally be necessary to create new families. And when you do, it will
>> probably be part of a major release that will have some impact on your
>> infrastructure.
>>
>> You should define an operational procedure for restarting all the nodes for
>> occasional use by ops engineers; they could choose to automate it.
>>
>> Restarting all nodes isn't that onerous provided it's not done too often.
>> It's not going to take > a few hours, even if you have quite a few.
>>
>> Mark
>>
>
>
>
> --
> Evan Weaver
>


[prev in list] [next in list] [prev in thread] [next in thread]