Re: [ZODB-Dev] ZODB replication and high-availability across data centers

2010-10-21 Thread Alan Runyan
 we are currently investigating options for a large-scale solution
 where the machines (ZEO clients and ZEO server) should be separated
 across two data centers. Incoming traffic should be balanced between
 both data centers. Automatic failover and online-replication are
 desirable - high-performance is a must.

 The well-known options are:

 ZRS
 (ZEO)
 Relstorage on top of some RDBMS-related clustering/replication mechanism
 NEO
 DRDB
 ZEORaid

We have not explicitly set something like this up; although our hosting
partner has SAN replication between data centers.  This is another
solution; while low level - may be valid to explore.  They use
it with Oracle and MSSQLServer to do cross datacenter replication.

cheers
alan
___
For more information about ZODB, see the ZODB Wiki:
http://www.zope.org/Wikis/ZODB/

ZODB-Dev mailing list  -  ZODB-Dev@zope.org
https://mail.zope.org/mailman/listinfo/zodb-dev


Re: [ZODB-Dev] ZODB replication and high-availability across data centers

2010-10-21 Thread Roché Compaan
On Thu, 2010-10-21 at 18:07 +0100, Chris Withers wrote:
 On 21/10/2010 16:44, Andreas Jung wrote:
  The well-known options are:
 
  Relstorage on top of some RDBMS-related clustering/replication mechanism
 
 This is where I'd look, since the RDBMS guys likely have patterns that 
 match the one you're seeking...

True, but they also have SQL which can be replicated with much less
overhead. Replicating at such a low level requires replicating whole
objects even if a single attribute changes whether you use Relstorage or
not. This is why I would consider replication at application level,
especially in a limited bandwith environment.

-- 
Roché Compaan
Upfront Systems   http://www.upfrontsystems.co.za

___
For more information about ZODB, see the ZODB Wiki:
http://www.zope.org/Wikis/ZODB/

ZODB-Dev mailing list  -  ZODB-Dev@zope.org
https://mail.zope.org/mailman/listinfo/zodb-dev


Re: [ZODB-Dev] ZODB replication and high-availability across data centers

2010-10-21 Thread Shane Hathaway
On 10/21/2010 09:44 AM, Andreas Jung wrote:
 -BEGIN PGP SIGNED MESSAGE-
 Hash: SHA1

 Hi there,

 we are currently investigating options for a large-scale solution
 where the machines (ZEO clients and ZEO server) should be separated
 across two data centers. Incoming traffic should be balanced between
 both data centers. Automatic failover and online-replication are
 desirable - high-performance is a must.

 The well-known options are:

 ZRS
 (ZEO)
 Relstorage on top of some RDBMS-related clustering/replication mechanism

Six Feet Up is running RelStorage + MySQL + async replication for one of 
their customers.  AFAIK, it works well.  Also, I expect the new 
replication features in Postgres 9 to work well with RelStorage.

Depending on what you're doing, balancing between data centers could be 
very complicated.  You might start simple by directing traffic to only 
one data center at a time.

Shane
___
For more information about ZODB, see the ZODB Wiki:
http://www.zope.org/Wikis/ZODB/

ZODB-Dev mailing list  -  ZODB-Dev@zope.org
https://mail.zope.org/mailman/listinfo/zodb-dev