Todd,
I was wondering why it hasn't been updated since Jan., and I found your message saying its now EOL. I gotta say, that really sucks. iSCSI Enterprise still has a lot of issues.
A while back I mentioned how the Target IP Access hardly works right. Maybe not a big deal, but certainly clunky for a final version, don't you think?
Another issue which caused me a lot of grief is that there are typos in the console menu under "Tuning -> iSCSI Daemon Options". When selecting an item, it tells you what the default value is supposed to be. Both "Max Burst Length" and "First Burst Length" are incorrect. "Max" is supposed to be 262144. But it says 262214. "First" is also wrong. It may not seem like a big deal, but I was unable to connect for a while one day until I realized the stated default values were incorrect.
When I first bought this product, it claimed "Mac OSX Initiator Support" on the product page. But it didn't. ATTO XtendSAN 1.0 (the only Mac initiator that existed at the time) never worked. Now 2 years later, neither of the two Mac Initiators work properly. XtendSAN 2.0 works, but miserably with this target. Write speeds are less than 20% of what they should be. The SNS globalSAN initiator works, and it has incredible performance with this target. Near wire-speed. But it occasionally hangs and times out during large data transfers. Kind of a problem.
At one point, one of the past updates said it "now supports volumes >2TB". From what I can tell, that never really happened. At least not for me. No matter how large, my OS only sees 2TB. And one of the initiators I use (globalSAN) absolutely supports >2TB in Mac OSX, at least when using their target products (which happen to use Linux-based Wasabi DOMs, similar to Open-E).
The 3ware driver Open-E iSCSI uses has been screwed up from day one. If you use all 16 SATA ports, any data you write to that disk will get corrupted. It doesn't matter if you create a giant 16-disk RAID 5, or split it into 2 or more RAID volumes. Anything written will get corrupted. Regardless of what initiator I've used. And you'll never know about it until its too late. This isn't some freak thing that happened to me once or twice. I've tested this countless times and repeated it consistently over the past 2 years using numerous versions of Open-E. Before and after the 3ware drivers were updated. I've tried multiple cards of the same type (9550SX-ML16). I've contacted 3ware support about it. I removed the Open-E DOM and tried both Windows and Ubuntu OSes using the same hardware RAIDs w/ 16-disks - and the problem didn't occur. I even tried the StarWind demo under Windows and connected to it as an iSCSI disk, and it worked fine. The corruption only happens when it is an Open-E target, and the 3ware support tech absolutely believed the Open-E implementation to be the problem. And despite all of that, if there is still any doubt, I also have a Wasabi iSCSI target using the same 3ware card (the reason I have two of them) and it handles 16 disks just fine (despite its other shortcomings and lack of an upgrade path).
Lowering the disk count to 14 solved the problem, and I've found workarounds to most of the other problems. Which is why I haven't really complained too much until now. But this EOL thing has rubbed me the wrong way? Why does booting off a USB disk vs. an IDE disk constitute an entirely new product while discontinuing updates for the original one?
Sorry for the rant, but as you can see this product hasn't worked out too well for me. I appreciate the low-cost and broad hardware support. I didn't have $30k to spend on a SAN, so this solution saved me a lot. But seeing 1.72 is considered the final version of the product - I think that's very unfortunate.
