12-16-2010 02:48 AM
Not sure if this is a problem or not? A customer has recently implemented EMC Recoverpoint on his existing network which initially consisted of a various mix of hosts with path failover going through 2 x 4100B (one in each path) switches to SAN storage (DS4000). As part of the Recoverpoint install 2 x 7600 switches were installed these were then confgured with ISL's to the existing 4110B's then MSFR (Frame redirect) VI's and VT's were configured to allow a SAN tap off to the Recoverpoint RPA's. Now as the DPC in the 7600 is pretty much controlling the SAN traffic my question is, yes there is a question!, If I loose the 7600 switches would I loose all connectivity between the Hosts and SAN Storage? i.e. there is no fall-back option when configured with the 7600?
Being a newbie to this not sure how much sense this makes?
Any answers greatly appreciated.
12-16-2010 03:21 PM
may not have the best answer but hopefully can help with some insight.
My understanding is that with Frame Redirect, only I/O for those volumes that are replicated is actually redirected to the splitter.
We had a failed AP7600B recently and our experience confirmed our understanding: all the hosts with Frame Redirect lost their paths to storage on the fabric where the spitter failed - this storage was set up for replication. However, we did have 1 of those hosts with Frame Redirect that had a volume that was not being replicated and that volume was accessible to the host durinig the AP7600 outage while its replicated volumes were inaccessible to the host down that fabric.
All other hosts with no Frame Redirect and no replication had all paths available.
Hope this helps.