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

List:       zodb-dev
Subject:    Re: [ZODB-Dev] ZODB blob support.
From:       Chris McDonough <chrism () plope ! com>
Date:       2005-03-21 19:22:13
Message-ID: 1111432933.16360.48.camel () athlon
[Download RAW message or body]

On Mon, 2005-03-21 at 13:39, Jeremy Hylton wrote:
> On Mon, 21 Mar 2005 18:08:35 +0000, Toby Dickenson
> <tdickenson@devmail.geminidataloggers.co.uk> wrote:
> > On Monday 21 March 2005 17:50, Chris McDonough wrote:
> > 
> > > - People are are unwilling to accept any weakening of data integrity
> > >   guarantee and who want the ZEO server itself to be responsible
> > >   for writing data (the ZEO clients must send the server the data
> > >   over RPC).  OTOH, if configured properly, the data may be read
> > >   from a shared filesystem.
> > 
> > I suspect some weakening is inevitable, unless you severely restrict the
> > choice of shared filesystem. Every network filesystem has its own ideas about
> > when client caches should be flushed.
> 
> I wouldn't want to use a network filesystem in production.

I suspect that the typical setup that Jim has proposed would ensure that
only reads are directed over network filesystems.  A local disk storage
would be used on the ZEO server which would allow for more atomicity at
commit time.  That filesystem would be shared to clients over the
network for read only; clients would never write directly to it.

- C


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

ZODB-Dev mailing list  -  ZODB-Dev@zope.org
http://mail.zope.org/mailman/listinfo/zodb-dev
[prev in list] [next in list] [prev in thread] [next in thread]