Best card/hdd combo for 5x750gb raid5?

A

arnemetis

2[H]4U
Joined
Aug 2, 2004
Messages
4,090
Hi everyone,
Been lurking around this subforum for a while reading up on raid and popular cards. For my new pc I intend on having 5x750gb drives in raid5 to have plenty of space which will also be protected in the event one of the disks fails, so I don't lose everything. I originally had planned on using my upcoming motherboard's onboard sata since it supports raid5, however it is my understanding that when it comes time to upgrade, I won't be able to take the array with me, and the performance isn't very good, and lastly it would bog down my cpu with the parity calculations. The most this setup will see is perhaps downloading from newsgroups at 8mbps while I try to play a game (which will be installed on another hard drive anyway), so far from an intense scenario. I am going to be running a heavily overclocked quad and sli, all watercooled, so I can't have the cpu doing all the work since that's reserved for games or folding. :D

Since I anticipate this array will either travel to the pc after this one, or at least be retired to a file server, it looks like I'll need a hardware card to keep it independant of the motherboard. This Areca 1220 8 port card looks to be best in my situation, and Areca seems to hold good ground here on the forums. Any comments or suggestions on this would be greatly appreciated.

Secondly, the hard drives. I've settled on the seagate 7200.11 series 750gb drives (not the enterprise drives), unless there's good reason to go elsewhere. Five year warranty trumps any theoretical speed advantages. Welcoming suggestions here as well of course.

Lastly, if by some miracle I've managed to put together the right combo for my needs, please give me the thumbs up rather than just passing by, figuring its just fine! Thanks.
 
I don't think those Seagate drives have TLER, and AFAIK there is no way to enable TLER on them either. (most?) Western Digital drives can have TLER activated by use of a DOS utility. See this thread for more information.

I don't actually know how often a drive will be dropped because of the lack of TLER, but in a RAID5 array it seems prudent to have it, seeing as data security is the whole point of the array.

When it comes to the controler card I have no real knowledge, so don't pay too much attention to this paragraph, but I believe the Adaptec 5805 is much quicker then the areca 1220 (which is a relatively old card?). Since you're not looking for performance (and the array isn't too large anyway) this isn't all that important, but a faster processor should also speed ut the build/repair processes.
 
Thanks for the input everyone. Regarding that Adaptec 5808 card, even if it is faster, It costs about $100 more and doesn't have standard sata port connections, its uses multi lane sas connectors. Will probably stick with the Areca unless a less expensive alternative is proposed.

That TLER is a valid concern now...I thought it was only affecting WD drives. I guess now I need some suggestions regarding this issue as well. I see three options now:
1. Get the WD drives and use the WDTLER utility as described in a thread in this forum
2. Get the Seagate ES drives that run about $60 a pop more than the regular ones, but which are supposedly 'designed' for this kind of application
3. Stick with the regular seagates and hope this problem isn't as widewpread as it is with the WD's

Would appreciate some further suggestions on this, thanks!
 
arnemetis said:
Thanks for the input everyone. Regarding that Adaptec 5808 card, even if it is faster, It costs about $100 more and doesn't have standard sata port connections, its uses multi lane sas connectors. Will probably stick with the Areca unless a less expensive alternative is proposed.

That TLER is a valid concern now...I thought it was only affecting WD drives. I guess now I need some suggestions regarding this issue as well. I see three options now:
1. Get the WD drives and use the WDTLER utility as described in a thread in this forum
2. Get the Seagate ES drives that run about $60 a pop more than the regular ones, but which are supposedly 'designed' for this kind of application
3. Stick with the regular seagates and hope this problem isn't as widewpread as it is with the WD's

Would appreciate some further suggestions on this, thanks!
Click to expand...
I've never heard of TLER affecting anything but WD drives. I've had 5 x 750 GB 7200.10 Seagate drives in RAID 5 on a Highpoint 2320 for over a year with 0 issues.

edit: I've heard Hitachi drives have the least issues with Areca controllers. There was a thread about it at 2cpu forums, but I can't find it. You might look around their storage forum.
 
All drives should be equally affected by the lack of TLER-like features. Any drive that spends too much time trying to fix an error will (from what I've understood) be dropped by the RAID controller.

Be aware that TLER is just Western Digital's name for this feature. Seagate calls it ERC (Error Recovery Control), while Samsung and Hitachi call it CCTL (Command Completion Time Limit).

Again I have no idea how often this will actually be a problem, but it's pretty obvious that not having TLER/ERC/CCTL will increase the probability of loosing the array. The question is by how much.
 
I can't believe this. Do most people buy the expensive ones? It's my understanding however that it's just once in a while a drive will drop out, and I'll have to rebuilt it. I shouldn't lose the data (or brick the dropped hard drive), right? That's what I'm most concerned with. If I have to sit through a few hours of rebuild a couple times a year to save $400, It's worth it. This computer won't be on 24x7, nor will it be seeing heavy load, and almost never seeing multi-user networked loads. Is this all I have to worry about?
 
Well it seems reasonable that the error recovery process is more likely to take too long when the drive is under heavy load, so under a light usage scenario it might not be very likely that there will be a problem (at least that's what I've read). Ofcourse that means that the likelyhood of another drive dropping out while the array is busy rebuilding is much higher, which is a bit scary.

Samsung has an article about it here.

"What Leads to Long Error recovery situations?

Enterprise workloads with lots of I/O including random I/O and Streaming I/O. Applications which create high levels of I/O, lots of read/write requests..."

"At that point, the RAID controller will rebuild the raid volume, while the RAID volume is performing parity rebuild, basically the data on the failed drive is recreated for the new drives using parity recalculation. Parity recalculation involves high I/O"

Interestingly they don't come out an say that another drive failure is more likely during rebuild, but it seems implicit do me, unless the array is always under heavy load anyway.
 
EnderW said:
I've heard Hitachi drives have the least issues with Areca controllers. There was a thread about it at 2cpu forums, but I can't find it. You might look around their storage forum.
Click to expand...
found it, this is a post from a guy who sells Areca controllers
http://forums.2cpu.com/showpost.php?p=735370&postcount=17

although if you haven't bought the card yet, I'd take a hard look at the Highpoint 3520 and Adaptec 5805
 
You must log in or register to reply here.
Back
Top