On Sat, Aug 9, 2014 at 3:59 PM, Charles F Sullivan <[email protected]> wrote: > Okay thanks, that's interesting to hear. > > The site where we do recovery of the most critical services, and which must > be done quickly, is even a little closer than yours. > > I just started using vSphere Replication to sync four servers (with another > four to be added) to that site. It works okay, but the sync is slow in that > it's not using the available bandwidth, which apparently is expected for the > way the product is designed. It's mostly a problem on the initial sync, but > I also find that the RPO of 15 minutes is often not being met. How do you > find the replication speed to be using SRM?
I don't use vSphere replication, we have the 2 Recover Points provide the replication and synchronization. The Recover Points are configured so that SRM actually manages it, tho. The replication speed seems pretty good to me, altho I don't have numbers right at hand. We did have an issue where we had to interrupt a testing plan, and that locked up SRM, and apparently corrupted the RP tables. We had to have EMC connect in remotely to fix it, by removing SRM as the RP manager, breaking the connection link, and re-establishing. And then uninstalling and re-installing SRM at both sides. (and recreating all settings - that was a nightmare couple of days). Anyway, once they re-established the link between the 2 RPs, they did re-synchronize overnight. Mind you, they weren't that far out of sync, but there were like 23 LUNs that had to synchronize.
