Re: Effect of SAN-level replication on Informix

On Tuesday, July 31, 2012 1:54:40 PM UTC-6, red_valsen wrote: > Does anyone have experience with SAN-level replication and its effect on Informix, especially with respect to disaster recovery/restore? We're examining different redundancy schemes of our environment across a WAN using Redhat Linux 4/5/6, Hitachi SANs, VMware and Informix (and other DBMSes). Some staff have expressed reservations that SAN-level replication can adequately substitute for database-style transactions in event of sudden failure, that data granularity may be too fine, perhaps resulting in not just failed transactions but referential integrity violations, further data corruption or simple failure when a failover or restore (at O/S or instance-level) is required. > > > > Has anyone dealt with these issues? Disk based replication can be used with Informix. However, you need to be aware that a disk based replication solution requires far more network bandwidth than with HDR or ER. With disk based replication you are transmiting the logical log changes, the physical log file writes, sort space IO, temporary tables, etc. With HDR you only transmit the logical log. Also disk based replication does create a lag on all writes - so it will impact performance.