Site Recovery Manager 5: The BIG Feature List

 Administration, Disaster Recovery, Software Releases, Tips and Tricks, VMware News
Aug 162011
 

As you probably know by now SRM5 is just over the horizon.  You have probably heard me mention numerous times how SRM has always been my favorite non-vSphere product from VMware.  Some great news is that they have made some great improvements in SRM5 and added the most-requested functionality.  Here we go:

  • vSphere Replication – The biggest feature add.  An additional replication option which allows you to replicate your VMs without having the storage perform the replication.  Even allows you to replicate to/from local storage on the ESXi hosts.  There are some important limits to vSphere Replication.  It’s not for everything/everyone but it does do quite a bit for the first release.
    • Requires vSphere 5
    • Managed from the vSphere client directly
    • ISOs and Floppys are not replicated
    • Powered off/Suspended VMs are not replicated
    • Non-critical files are not replicated (swap files, dumps, logs, etc.)
    • VMs can have snapshots on the protected side but they are automatically collapsed on the recovery side
    • Physical RDMs not supported (but virtual RDMs are)
    • Fault Tolerant VMs, Linked Clones and VM Templates are not supported
    • Automated Failback of vSphere Replicated VMs is not supported in SRM 5.0
    • Requires VM Hardware version 7 or 8 (required for Change Block Tracking)
    • Supports up to 500 VMs
    • Asynchronous only
    • Minimum replication frequency is every 15 minutes, max is every 24 hours
    • Initial copy can be seeded by sneaker net (taking the initial on a portable HD and importing at the destination, i.e. does not need to seed the initial copy over the wire)
    • File-level consistency (except for planned migration – see below) quiesces OS file system before sending changed blocks to the DR site (does not quiesce applications)
    • Included in both Standard and Enterprise Editions of SRM
    • vSphere Replication is not available outside SRM5
  • Scalability Improvements
    • 1000 Total Protected VMs (Same as SRM4.1)
    • 500 Protected VMs in a single protection group (same as SRM v4.1)
    • 250 Protection Groups (Up from 150 in v4.1)
    • 30 Simultaneous running recovery plans (Up from 3 in v4.1 – this is the biggest improvement in scalability)
  • Planned Migration – This is a big feature add.  This is another option when you are going failover.  In 4.1 the only option was to start up the VMs from the last good replication and go.  This option now allow you to migrate when there is an impending disaster and the protected side is still up.  Planned migration will shut down the VMs on the protected side then initiate a replication of the storage frames (or vSphere Replication) to get the last drop of changed data to the recovery side before powering on the VMs and bringing them up.  One extremely important advantage to this method: the VMs are always in a application-consistent state when they come up in DR.  (Absolutely love this feature)
  • Failback – the single most-requested feature in SRM4.  Once a failover occurs, the admin clicks the “Reprotect” link to reset the recovery plan for failback and reverse replication.  Once completed, the recovery plan can be tested or run in the reverse direction and recovery the VMs to the origional protected site.  (This is outstanding for enterprises that are required to do a true failover for DR testing.
  • User Interface improvements – Slightly different look and feel.
    • both sides are visible without vCenter linked mode
    • IP changes for VMs during recovery can now be entered in the GUI (thank you VMware!)
    • Placeholder VMs at the DR side now have a unique icon (with a thunderbolt thru it) to identify them easily in the DR vCenter.
    • Reports now include the user ID that initiated the Failover or DR test.
    • Reports now include more information about the storage steps (including the device friendly names)
  • IPv6 Support – Ipv6 is now supported for all links.
  • IP Customization performance increase – big performance improvement in the actual IP conversion in the VM
  • In guest callouts – now you can run a script inside the VM, run a script on the SRM server or insert a breakpoint to post a message (these also now have maximum timeouts as an option) during the recovery plans
  • New APIs on both the Protected and Recovery Sides – new commands for 3rd party integration (note these are SOAP based and not PowerShell or PowerCLI)
  • Dependency Improvements – There are now 5 priority groups for each recovery plan.  Each priority group has to finish completely before the recovery plan will start with the next group.  Within a single priority group, you can also set dependencies (similar to how Windows Services set dependencies) so that a particular VM will not recover before it’s dependencies have recovered (note-this is within a single priority group and cannot span priority groups.)
  • Licensing – There are now two editions of SRM, Standard and Enterprise.  Both are feature identical.  Standard is for sites up to 75 VMs and Enterprise is for sites up to 1000 VMs (the technical limit).  All existing customers who maintain support will get SRM Enterprise when they go to SRM5.  SRM Standard is a new offering for SMBs and Remote Offices.  When customers need to grow beyond 75 VMs at a site, they can upgrade their existing VMs to SRM Enterprise and then continue buying SRM5 Enterprise VM-Packs.  Licensing still sold in packs of 25 VMs.  Only need to purchase for the VMs that you are going to protect.