Specs said you can get an "expander" for that card so it supports up to 256 drives (?)
SAS expanders are not typically something that one usually buys separately, they are typically built into larger disk cages or enclosures. In some cases, large disk enclosures will have multiple levels of expanders (if you want to pack 100 disks into an enclosure, a single expander chip won't do).
Do you know anything about the "WL" brand of hard drives?
They're said to be made as white label drives then rebranded and sold by many OEM's, etc.
There are only 2.5 manufacturers of disk drives: Seagate (which includes Samsung), Western Digital (sometimes known as WD, and some disks still known as Hitachi), and Toshiba (which I count as 0.5 because they are small). So where do off-brand drives come from? Typically rejected drives from one of the big vendors. They could take two possible paths: either the manufacturer sold them to a big user (90% of all enterprise disks are sold to a dozen big companies, like Amazon/Apple/Baidu/Facebook/Google/Microsoft/Tencent or Dell/HP/Amazon/Oracle), failed QA testing there or showed errors early on, and were sold to unscrupulous resellers that hide their true history. Or they are drives that failed QA testing at the manufacturer, but I don't think the big manufacturers would be willing to sell those.
In all this, one has to remember that the disk manufacturers have very strict QA systems. And they grade the quality of disk drives: the best ones go to preferred customers, who also get access to internal QA information on a per-drive basis and use the drives in-house, and who pay a premium (typically the cloud superscalers); the decent ones go to price-conscious customers (such as Dell) who resell the drives as part of systems, and the not-so-good ones go into the retail channel. The joke used to be that the worst drives are sold at Fry's (a chain of electronics supermarkets on the west coast of the US, in particular in Silicon Valley, infamous for selling junk and having impossible to navigate return systems), but that's no longer true since Fry's has gone out of business.
If I had nothing useful to do, I would waste $50 on one of those drives, find out exactly what kind it really is (the firmware will give it away), and then perhaps bring it to some friend who works for one of the drive makers and we take it apart together. But I have too many useful things to do.
Summary: STAY AWAY.
I've also heard that its a good idea to mix in different hard drive brands.
Opinions on that differ. At the consumer level, where you have no information about the reliability of individual disks, it might be a good idea, just to guard against the unfortunate coincidence that you buy all drives from one manufacturer/model made roughly at the same time and place, and that kind happens to be unreliable. But modern enterprise-grade drives are so good, with a decent RAID layer on top, they will be reliable enough. At the large-scale user level (the customers who buy a million disks at a time), there are large groups of people who study, measure and forecast drive reliability, and who consciously adjust data placement to maximize reliability and minimize cost. This is one of the reasons that small users simply can't compete with cloud providers: You can't afford the group of 5 PhDs and 10 software engineers that perform such optimizations, but you can rent disk space in the cloud from companies that do.
(about draining a disk that is about to remove)
This one sounds really interesting. Do you have the steps listed somewhere that you're willing to share, so I could experiment?
Look at the "zpool remove"command.
Not needed. Just buy quality drives.
Look at backblaze stats.
THIS. For the consumer who buys small quantities of drives from the retail environment, looking at Backblaze is the best idea, because that's where their disks also come from. The problem with this approach is: by the time Backblaze has good high-statistics data (like having used 10,000 disks for 4 years), the disk model is probably obsolete, and can no longer be found in the retail channel, except for used or rejected drives. So the idea here is to look for patterns, like all the disks from manufacturer "Elephant" with model names that start with "Dumbo" are very good, and then follow that pattern. I'll give you a hint: My spinning drives are (WD) Hitachi HGST enterprise-grade drives, with model numbers starting with H.
How many Hot-swap slots would I need (of the 8 available) to pull this off, *if* its even possible with my small rig?
Could I use an external hard drive or two, in an external drive enclosure, attached w/usb cable?
One spare (empty) slot is enough. Sure, you could do it with external enclosures, but that's a hassle: You have to put the new/spare/old disk in the external enclosure, USB is probably slower, then remove it again and put it into its real location. Easier to have a spare slot.
I'm starting to think along these lines for my final setup:
Rack server case has maximum of 3 hot swap cages @ 4 drive bays/cage = 12 hot swap drive bays:
Given that you are planning to use 6 drives (and your assignment looks reasonable), I think two bays, meaning two spare slots, seems adequate. And cheaper.
