File Catalog Distribution/Replication (phase one: Grid - work around)
Updated on Fri, 2005-11-04 11:35. Originally created by deph on 2005-11-04 10:26.
This blog is to present, record discussions and ideas regarding the distribution of file catalog database to remote grid enabled sites.
MySQL replication will be used, initially to provide a complete file catalog at each of the sites. The challenge is that each sit will be able to update this catalog, so this information needs to be centrally aggregated and then redistributed to each site as a complete/comprehensive database.
To this end, we will dedicate a single machine db09 (dubbed mega-master). This machine will hold BNL's master, be a slave to each sites master database, and combine all these databases into one mega-master that is replicated to a "whole" slave at each site. This final slave will be what is queried by grid jobs.
Specific configurations and technical details will be presented as comments to this blog.
MySQL replication will be used, initially to provide a complete file catalog at each of the sites. The challenge is that each sit will be able to update this catalog, so this information needs to be centrally aggregated and then redistributed to each site as a complete/comprehensive database.
To this end, we will dedicate a single machine db09 (dubbed mega-master). This machine will hold BNL's master, be a slave to each sites master database, and combine all these databases into one mega-master that is replicated to a "whole" slave at each site. This final slave will be what is queried by grid jobs.
Specific configurations and technical details will be presented as comments to this blog.
»
- deph's blog
- Login or register to post comments