Hi folks,
I started using Areca RAID cards after they were recommended to me on this forum as one of the best performing RAID controllers. That was around the ARC-1231ML days. That was a fantastic card and I loved it. Since then, I've also used the ARC-1680, ARC-1880 and ARC-1882 (all "ix", 24-port cards), upgrading my array through the generations and creating new ones. I've used them on multiple machines (workstations and servers). Overall they haven't been bad, but there are a few nagging problems which have persisted through all generations of this controller and are driving me crazy, e.g:
- Doesn't play well with consumer drives as they age; tends to drop them too aggressively after timeouts (yeah, guessing it's a TLER thing - although I've been careful to only pick drives recommended for RAID arrays, and have played with timeout settings on the card)
- Card sometimes becomes unresponsive in the web GUI, and sometimes even in BIOS management. Typically occurs after something unusual happens. e.g. Right now I've got an array that was somehow mucked up (am not clear how; there were no drive failures, although I was messing with boot records on unrelated disks). If I boot the card with all the disks plugged in, then TAB to the BIOS, it doesn't get past the "Select controller" screen (i.e. the list of controllers goes away, but the background just stays blue instead of it loading all the menu options). If I boot with some drives unplugged, the remaining ones show "Free" even though they should show as part of an incomplete RAID set. As I plug new ones in, those show correctly. If I boot with ALL of them unplugged and plug 'em in one by one, they show up correctly, until the last one is plugged in at which point the controller starts to rebuild the array and freezes up at 0%. It's this kind of strange edge case behavior that makes me nervous. I'm likely going to have to destroy & rebuild the array. I've also seen the web GUI go unresponsive during operations that utilize a lot of CPU (e.g. RAID6 rebuild). Occurs both through the proxy service and via the dedicated ethernet port. In general, it seems to me there are some edge case bugs in the firmware, and the management processes aren't "out of band" enough. I need a card that's impeccably reliable when things start to go wrong, not one that exhibits fragile behavior and breaks down when weird things happen.
- I have only once successfully recovered an array (even after minor events where I would have expected a decent chance of recoverability). I suspect this is more a lack of knowledge on my part (i.e. no documentation / clear process in recovery procedure) or due to taking the wrong steps due to controller exhibiting flakey behavior in edge cases.
- The later models have very long delays while loading firmware during boot (for those times you need to do a lot of rebooting on the PC without caring about RAID, would be nice if they had a hotkey to skip the load and disable the card for this session).
I've had other minor annoyances with the card over time, again, mostly firmware related. There are a lot of things I love (e.g. the fact that they've maintained compatibility with older RAID arrays through the ages; the clean, non-cluttered, keyboard-friendly DOS-style BIOS; the out-of-band ethernet port; generous RAM; etc). But over the next few years I anticipate moving my large "platter-based" storage arrays to dedicated ZFS machines, and using NVMe SSD's where workstation performance is required.
In the meantime, I'm wondering if there are alternative high-performance solutions, controllers or other hardware/software folks here might recommend (either in existence today or coming soon). I hear LSI has caught up in terms of raw performance (although I hate their GUI's).
Thanks!
Reply With Quote
Bookmarks