BonzaiDuck
Lifer
OK. I've posted another thread at storage on the hardware issue.
I had an absolutely stellar IDE hot-swap bay and caddies with 500GB IDE drives which were truly swappable with a 32-bit OS and the bundled "Swap Manager" software. That system is being "parted out" and replaced.
I've contemplated the problem of using the bay, caddies and drives on my WHS-2011 (64-bit) system. If I connect them via IDE, I can't hot-swap the drives; I must shut down the server and then swap before bootup. "Swap Manager" doesn't work, and the drive doesn't appear in the "Safe Removal" list as long as it's connected to the mobo IDE controller.
I had partial success by using a USB2 to IDE cable/adapter internally. Totally hot-swappable!! No danger to the not-much-used IDE drives.
But.
Too slow. I started backing up the "Client Computer Backups" folder of my drive-pool yesterday at about 1PM. Nightly client backups occur automatically between midnight and 6AM to the server drive pool, and the hot-swap backup was only 60% complete with the USB2 speed. So I'm going to replace the hardware with a USB3 solution -- I await delivery of the two essential parts.
Two of four client backups failed last night because of this overlap (can't think of another reason.) I'm using RichCopy 4.0 (downloadable from MS/MSDN/TechNet -- whatever). Since I can't "image" my pooled drives' "Virtual" disk with Acronis or WHS (native backup), this is the best solution. I can create profiles for each set of source folders and target backup disk; load them; press "start" and walk away. Instead of scheduling these backups automatically, this solution only has the overhead of the one-time profile setup, and I can take those backups at a frequency of my own choosing.
I notice that the file dates shown in the target disk's folders reflect that previous night's successful backups for all four clients.
How is RichCopy handling this? Is the "successful" handling of the previous successful backups pre-empting the most recent nightly backups (client to server drive-pool) from proceeding? I suspect that the other two machines whose backups proceeded successfully last night (this morning) had already completed copying to the USB2 drive.
Also, now that I've embraced a USB3 PCI-E x1 controller ($20) for the server and a USB3-to-IDE cable/adapter to work with it, how much extra work (stress, wear etc.) does the target drive undergo for working continuously over 24 hours as opposed to probably 1/5th or less of that time with the anticipated USB3 solution?
I had an absolutely stellar IDE hot-swap bay and caddies with 500GB IDE drives which were truly swappable with a 32-bit OS and the bundled "Swap Manager" software. That system is being "parted out" and replaced.
I've contemplated the problem of using the bay, caddies and drives on my WHS-2011 (64-bit) system. If I connect them via IDE, I can't hot-swap the drives; I must shut down the server and then swap before bootup. "Swap Manager" doesn't work, and the drive doesn't appear in the "Safe Removal" list as long as it's connected to the mobo IDE controller.
I had partial success by using a USB2 to IDE cable/adapter internally. Totally hot-swappable!! No danger to the not-much-used IDE drives.
But.
Too slow. I started backing up the "Client Computer Backups" folder of my drive-pool yesterday at about 1PM. Nightly client backups occur automatically between midnight and 6AM to the server drive pool, and the hot-swap backup was only 60% complete with the USB2 speed. So I'm going to replace the hardware with a USB3 solution -- I await delivery of the two essential parts.
Two of four client backups failed last night because of this overlap (can't think of another reason.) I'm using RichCopy 4.0 (downloadable from MS/MSDN/TechNet -- whatever). Since I can't "image" my pooled drives' "Virtual" disk with Acronis or WHS (native backup), this is the best solution. I can create profiles for each set of source folders and target backup disk; load them; press "start" and walk away. Instead of scheduling these backups automatically, this solution only has the overhead of the one-time profile setup, and I can take those backups at a frequency of my own choosing.
I notice that the file dates shown in the target disk's folders reflect that previous night's successful backups for all four clients.
How is RichCopy handling this? Is the "successful" handling of the previous successful backups pre-empting the most recent nightly backups (client to server drive-pool) from proceeding? I suspect that the other two machines whose backups proceeded successfully last night (this morning) had already completed copying to the USB2 drive.
Also, now that I've embraced a USB3 PCI-E x1 controller ($20) for the server and a USB3-to-IDE cable/adapter to work with it, how much extra work (stress, wear etc.) does the target drive undergo for working continuously over 24 hours as opposed to probably 1/5th or less of that time with the anticipated USB3 solution?
