![]() ![]() I had 4 of the 6 of them die on me during the lifespan of that RAID. My previous Promise R6 was populated with a batch of those horrible Seagate 3TB mechanisms released just after the flood. The fan noise on the OWC is acceptable to me – and can always be improved by swapping out the fan with a premium fan if you're super picky. And you can always load up a bare OWC enclosure with whatever quiet drives you like (or whatever Promise is using). But, a Pegasus RAID using the same drives will be just as loud. Luckily, I can keep the RAID in an adjacent room. My mechanical OWC uses HGST/Toshiba MD04 series drives, which admittedly are a bit too chattery/gronky when reading and writing for my liking. That is more a function of the type of mechanisms used in the RAID. Performance is on par.Īs for noise, I assume people in this thread are talking about disk access noise. I have two OWC RAIDs and had a Pegasus R6 previously. The backup could be done via Carbon Copy at night or continuously using Time Machine, or various other ways. This would give almost the same read performance as the R8 in RAID6 and 4x the write performance - PLUS provide a backup. All RAID systems must be fully backed up, so the cost of the backup method must be included.įor significantly less money than the Pegasus R8 you could get TWO 32TB OWC Thunderbay 4s, run them in RAID-0 with one as a backup. It gives 6x read speed and no write speed improvement over a single drive. However you'd normally run that in RAID-6 (double parity) which reduces capacity to 36TB. The Promise Pegasus R8 supports up to 8 drives, so 8 x 6TB drives = 48TB. The SoftRAID utility is easier to use and gives better feedback on drive status than the equivalent Promise Utility. ![]() There is no significant host CPU overhead from running SoftRAID, and in fact the entire system seems faster. Promise makes a very good product, I just like the additional flexibility of the OWC arrays. It's possible Promise has improved their rebuild or sync performance on newer models. I didn't save the similar data from my rebuild tests on the OWC Thunderbay 4 in RAID-5 using SoftRAID, but I recall it was much faster. A few years ago I did many tests on multiple R4 chassis and documented the data here: /photos/i-pdz6p2j/0/632a78ee/O/i-pdz6p2j.jpg If you lost a drive in RAID-5 it might take several days to rebuild that, during which performance was drastically degraded. The Promise Pegasys R4 had extremely slow rebuild times, (aka as RAID sync times). OWC and SoftRAID don't have those restrictions. It's likely due to the low-level interaction between their microcontroller and the HDDs. This even includes what firmware version is on the drive. Promise has a rigid, limited approval list for what drives are supported in a given raid chassis. By contrast if I someday don't like OWC, I can pull those drives and use them without losing data in an Akito, Lacie or any other software RAID box. The Promise hardware RAID format is proprietary. If I need to pull the drives from my Promise array, they can't be used in any other chassis without reformatting. The OWC arrays aren't as quiet but they aren't loud.Īnother big advantage of SoftRAID is avoiding lock-in to a proprietary hardware vendor. However my tests did not show any performance benefit of their hardware RAID vs OWC's software RAID. The Promise arrays are excellent and fairly quiet. This gives me the option of switching any to a higher RAID format if I want, plus I trust SoftRAID's data integrity and performance more than macOS's AppleRAID. All the OWC arrays use SoftRAID, even though macOS supports RAID-0 natively. I also have an 8TB OWC SSD array in RAID-0. One of those was previously a 16TB array in RAID-5 I upgraded it to 32TB and all three are now in RAID-0. I have an 8TB Promise Pegasus R4 in RAID-5 and three 32TB OWC Thunderbolt 4 arrays, all using SoftRAID. Is there a compelling reasons to spend the extra money for a Promise over an OWC Thunderbay?.Does having a software-controlled RAID impact FCPX in any NOTICEABLE negative way, since it's sharing CPU resources with the RAID? My question is whether there is a noticeable speed difference. The RAIDs that use hardware controllers (Promise) are considerably more expensive than ones that use software controllers (OWC Thunderbay 4). ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |