[OpenAFS] Practicality of near-continuous replication on busy volumes

Dominic Hargreaves dominic.hargreaves@it.ox.ac.uk
Thu, 3 Jul 2014 17:49:05 +0100

Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable


We are looking to improve the resilience of our VLE infrastructure, which
uses OpenAFS to house some data; in particular to provide multi-server
and multi-site data replication. The database part of this will be
provided by MySQL streaming replication, and we are hoping that the AFS
part can be provided by giving the VLE a multi-site AFS volume,=20
configuring the application to always talk to the read/write volume, and
then scheduling a 'vos release' every minute or so via cron or some other
method. Then the manual failover process would involve promoting the
secondary site manually.

Has anyone got any experience of doing this kind of thing with a busy
and large-ish (100s of GB) volume? Is it workable in principle or have
people found that there is a practical limit to how quickly/often
vos releases can be done - or any other flaws in the scheme?


Dominic Hargreaves, Systems Development and Support Section
IT Services, University of Oxford, 13 Banbury Road, Oxford, OX2 6NN

Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature

Version: GnuPG v1.4.12 (GNU/Linux)