Hi, hope someone can help with an issue i have come across in testing.
I have 2 vSphere 5.5 clusters running SRM 5.5 and attached HUS-VM storage.
Each site is protecting the other, and each site has 4 datastores replicated cross site to s-vols.
Horcm file has been loaded on eash SRM server, and arrays paired succesfully.
Testing SRM has worked succesfully from both sites in a "planned failover", where failover/reprotect and failback works perfectly, with P-Vols and S-Vols being mounted and unmounted as required. all hosts and VCS/SRM VM's were running during testing.
I simulated a "site failure" by ungraceful shutdown of hosts, all protected VM's, VCS and SRM vm's at protected site offline.
At recovery site, recovery plan for protected site ran succesfully, s-vols presented to read/write and protected VM's powered on.
My issue arises when..
1. Protected site becomes available - hosts power on, HA will remember state of VM's and power these on (even though running at recovery site) .. bad..!!!
2. protected datastores on origional protected site are still read/write !!
3. PVOL and SVOL's are not paired at the storage level
4. By re-running SRM recovery plan at recovery site again( as recommended), this will power down the VM's at protected site, change P-Vols to read only / unmount and re-pair the PLVOLs -SVOLS, then by running reprotect, and fail back return to origional state.
I need to be able to either:
At Protected site, prevent the VM's from powering on, but as a HA cluster the VM state is resumed, the option:Virtual machine startup/shutdown with host will not work,
Prior to protected site becoming available, issue command at recovery site SRM server like pairresync ?? etc to change origional PVOLS to read only, enable pair and start sync:.. with the protected site SRM server offline.
If anyone could point me in the right direction for this, would be verrrrry grateful