You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by A J <s5...@gmail.com> on 2011/11/15 17:20:33 UTC
Continuous export of data out of database
Hello
VoltDB has an export feature to stream the data out of the database.
http://voltdb.com/company/blog/voltdb-export-connecting-voltdb-other-systems
This is different from Cassandra's export feature
(http://wiki.apache.org/cassandra/Operations#Import_.2BAC8_export)
which is more of a different way of snapshotting.
My question is : is streaming data out on a continuous basis (as in
VoltDB) possible in some fashion in Cassandra ?
Thanks
Bala
Re: Continuous export of data out of database
Posted by Robert Jackson <ro...@promedicalinc.com>.
It would work if you could use LOCAL_QUORUM in each independent data center, but with EACH_QUORUM the whole read/write operation would block until a quorum was reached in each data center. This would be particularly bad if using a single replica in the "backup" data center and it was unavailable (the read/write would fail).
Unfortunately , I can't think of another solution with Cassandra itself. You could use some sort of client side queueing and keep the two clusters separated completely. This would probably add quite a bit of complexity to your app...
Robert Jackson
----- Original Message -----
| From: "A J" <s5...@gmail.com>
| To: user@cassandra.apache.org
| Sent: Tuesday, November 15, 2011 12:20:52 PM
| Subject: Re: Continuous export of data out of database
| The issue with that is that I wish to have EACH_QUORUM in our other 2
| datacenters but not in the third DC.
| Could not figure a way to accomplish that so exploring have a
| near-realtime backup copy in the third DC via some streaming process.
| On Tue, Nov 15, 2011 at 12:12 PM, Robert Jackson
| <ro...@promedicalinc.com> wrote:
| > The thing that I thought if initially would be setting up your
| > cluster in a
| > multi-datacenter config[1]. In that scenario you could add an
| > additional
| > machine in a second datacenter with RF=1. We are using a variant of
| > this
| > setup to separate long running calculations from our interactive
| > systems.
| > [1] -
| > http://www.datastax.com/dev/blog/deploying-cassandra-across-multiple-data-centers
| >
| > Robert Jackson
| > ________________________________
Re: Continuous export of data out of database
Posted by A J <s5...@gmail.com>.
The issue with that is that I wish to have EACH_QUORUM in our other 2
datacenters but not in the third DC.
Could not figure a way to accomplish that so exploring have a
near-realtime backup copy in the third DC via some streaming process.
On Tue, Nov 15, 2011 at 12:12 PM, Robert Jackson
<ro...@promedicalinc.com> wrote:
> The thing that I thought if initially would be setting up your cluster in a
> multi-datacenter config[1]. In that scenario you could add an additional
> machine in a second datacenter with RF=1. We are using a variant of this
> setup to separate long running calculations from our interactive systems.
> [1] -
> http://www.datastax.com/dev/blog/deploying-cassandra-across-multiple-data-centers
>
> Robert Jackson
> ________________________________
Re: Continuous export of data out of database
Posted by Robert Jackson <ro...@promedicalinc.com>.
The thing that I thought if initially would be setting up your cluster in a multi-datacenter config[1]. In that scenario you could add an additional machine in a second datacenter with RF=1. We are using a variant of this setup to separate long running calculations from our interactive systems.
[1] - http://www.datastax.com/dev/blog/deploying-cassandra-across-multiple-data-centers
Robert Jackson
----- Original Message -----
| From: "A J" <s5...@gmail.com>
| To: user@cassandra.apache.org
| Sent: Tuesday, November 15, 2011 11:20:33 AM
| Subject: Continuous export of data out of database
| Hello
| VoltDB has an export feature to stream the data out of the database.
| http://voltdb.com/company/blog/voltdb-export-connecting-voltdb-other-systems
| This is different from Cassandra's export feature
| (http://wiki.apache.org/cassandra/Operations#Import_.2BAC8_export)
| which is more of a different way of snapshotting.
| My question is : is streaming data out on a continuous basis (as in
| VoltDB) possible in some fashion in Cassandra ?
| Thanks
| Bala