As we wrap this up, lets talk about VERITAS's truly special, and more "advanced" features. Please don't for an instant think this is the complete and definitive list. Rather, this is a small list of features that I really think separate VxVM from the other volume managers available. All of these topics will be covered in detail in the "Advanced VERITAS Theory" course later on.
Dynamic Multi-Pathing: Also know as DMP, this feature allows for arrays which can be attached to multiple controllers (such as the Sun A5000 family of arrays) to leverage them for both load balancing and redundancy. So if your disks are attached to both c1 and c2, and your using a 100Mb/s fiber connection, with DMP you'll access the disks at 200Mb/s. And if c1 dies, for any reason, you still have access to the disk through c2, which means zero downtime for your volume. DMP can really save your butt if there is a failure.
Hot Spares: You can allocate VM Disks as spares, so that if a disk in one of your plexes dies for some reason the spare will immediately take it's place and again, zero downtime. You can have as many hot spares as you like, and VM Disks allocated as spares can are independent of any particular disk group, plex or volume, so you don't need to have different spares for different volumes. Spares are managed by the vxrelocd as mentioned earlier.
Dirty Region Logging: Also known as DRL, this feature can be thought of as a volumes diary. Before writes are made to a DRL'ed volume VERITAS makes a note of the write in the DRL, marking that region "dirty". After the note is made in the log, the write occurs. The idea here is that if the system crashes, when the system comes back online it won't know what writes were in progress. So, if a DRL log exists for the volume, it checks the log and then sync's dirty regions from the other mirror(s). In this way, we don't have to worry about corrupted data. DRL only applies to mirrored volumes, even though DRL can be added to unmirrored volumes (it just won't do anything useful).
Snapshotting: Users of NetApp Filers will be familiar with this. Snapshotting is used for online backups. When you "snapshot" a volume, you are saying "I want to create a static picture of what the volume looks like right now". Then once the snapshot is created it appears as a new temporary volume, which can be mounted and backed up. Once the backup is done you "release" the snapshot. The kool thing is that while the snapshot exists the volume is still read and writable to everyone else. Zero downtime backups!
OK, I'm going to leave the list at that for now. Don't get frustrated if your thinking "Hey, that's kool! But your not telling me anything useful!", because all of that will come later in the "Advanced VERITAS Theory" course, where I will talk about these topics and more in detail. For the meantime, just know that they exists and can be used.