Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Pegasus2 R4 drives
- Thread starter agtoau
- Start date
- Sort by reaction score
I ordered a Diskless last night, but it's saying a 1-2 month wait time. While I'm waiting, I guess I'll start researching drives. Is there any particular reason everyone here likes the specific drive they've listed? I'm not up to speed on the Reds/Blacks, from WD. Another other drives that guys recommend for me to research?
Getting drives for the Promise2 R4 is a bit of a gong show.
Good link if you've not seen it. http://www.extremetech.com/extreme/175089-who-makes-the-most-reliable-hard-drives
I have an empty diskless p2 r4 - waiting for HDD's.
The Promise compatibility chart for the P2 is horrible.
I spent many hours yesterday trying to find the ancient and few enterprise class drives on the list. Let alone a vender that would ship to Canada (additional headache for me).
So I have gone back an forth many times about sticking to the compatibility list or not.
I located some HUA722020ALA330's (the best drive on the list i think - SATA 3, 32 MB
) but I think they only carry a 1 yr warranty at this point (possibly because only 3rd party vendors have them). Not the 5 yrs they originally did when they were first released. I'm not happy about that at all, plus it will probably be impossible to get a replacement if one fails.. so I thought about ordering a 5th.. but gnaw. Screw the compatibility list.
Kind of a crappy situation to be in all because of Promise.
At this point I am going to go with either the ;
HUS724020ALE640 (1st) Ultrastar 7K4000
WD2000FYYZ (2nd) WD RE
or MG03ACA200 (3rd) Toshiba - which I know I'll get 5 yrs warranty on from a vender I have used many times over the years.
Good link if you've not seen it. http://www.extremetech.com/extreme/175089-who-makes-the-most-reliable-hard-drives
I have an empty diskless p2 r4 - waiting for HDD's.
The Promise compatibility chart for the P2 is horrible.
I spent many hours yesterday trying to find the ancient and few enterprise class drives on the list. Let alone a vender that would ship to Canada (additional headache for me).
So I have gone back an forth many times about sticking to the compatibility list or not.
I located some HUA722020ALA330's (the best drive on the list i think - SATA 3, 32 MB
) but I think they only carry a 1 yr warranty at this point (possibly because only 3rd party vendors have them). Not the 5 yrs they originally did when they were first released. I'm not happy about that at all, plus it will probably be impossible to get a replacement if one fails.. so I thought about ordering a 5th.. but gnaw. Screw the compatibility list.Kind of a crappy situation to be in all because of Promise.
At this point I am going to go with either the ;
HUS724020ALE640 (1st) Ultrastar 7K4000
WD2000FYYZ (2nd) WD RE
or MG03ACA200 (3rd) Toshiba - which I know I'll get 5 yrs warranty on from a vender I have used many times over the years.
I finally received my screws. The ones that worked were M3 x 5mm with a flat head and cones toward the screw. If the head is too curved upward, it will protrude and collide with the drive below it in the trays. But it worked perfectly with the ones I got.
The P2 R4 definitely reads them and identifies them correctly as SSDs along with their respective specs. I set both SSDs as two independent RAID0 (which basically is a JBOD setup). They definitely transfer data pretty fast. I will test their reaction speed to sample streaming once I have finished transferring all of the libraries from my backup drives.
On the BlackMagic app, I am getting anywhere between 250 MB/s - 450 MB/s per SSD, depending on the load.
That suits me fine for most of my purposes. One drive of the two is a little faster for some reason (both in write and read by about 100 MB/s), even though they have identical settings and are identical drives. Perhaps how full the drive is affects it or perhaps the average size of the individual files. But it doesn't matter. I now have 2 independent streaming sources. And if I hookup my older Lacie Rugged 256GB SSD to the 2nd TB port on the R4, I'll have 3 separate sources for streaming so I'll have even less traffic clutter for spontaneous reads across various libraries. And if it gets cluttered through TB, I'll simply hook it up to 2 separate TB2 ports for truly independent transfers.
*EDIT*
Never mind the slower/faster SSD comment. When I reset my computer, the drives both gave practically identical read/write speeds of roughly 450MBps/300MBps, respectively. And a note to those who try to replicate (in whole or in part) my setup with the P2 R4, if you want to maximize read speed at the expense of some write speed (for streaming samples), the best cache policy, which I have tested, is ReadAhead and WriteThru. WriteBack increases write speed at the expense of read speed.
Last edited:
That's a great news that P2 R4 recognizes SSDs. Just a curious question, does P2 R4's power links with computer's? Which means P2 R4 turns off as computer's off; turns on when the computer is on. If that's the case, do drives mounted instantly?
Yes, I was very happy when I saw that it actually identified it as an SSD and not another HDD.
To answer your question, yes. The R4 has synchronicity with the computer. When the computer is turned off or put to sleep, the R4 also shuts down. Likewise, if you are sending many GB that will take several hours to complete, when your screen turns off to save power, the R4 keeps doing it's job. It "knows" when it can turn off. It also turns off if it is idle for more than 30 min or so. And to reach full power again, it takes about 15 sec and stays hot and ready until idle for extended periods of time.
I also got to test the streaming capabilities of the R4. It works very, very well with the SSDs. I now realize how demanding the Hollywood series is though. The Hollywood series from EastWest has such detail that sometimes transients in the samples cause the stream speed to spike to 2GB/s (that's right... 2000 MB/s), which is WAY above the streaming capability of any single SSD today. But still, it is much better than the limitations of a HHD, since SSDs react instantly to random access. I will simply have to allocate my resources appropriately, but at least it is easier now with my 2 SSDs. I might have to load some patches directly into my RAM. Crazy.
If you have EWQLSO or Choirs, SSDs will work just fine with no issues whatsoever. But the Hollywood series is another monster altogether.
Can the P2 R4 take two SSD's in raid 0 and then have a 3rd HDD for time machine back ups with an empty 4th bay?
Two drive RAID
Yes. I have a two drive RAID 0 and two other drives in bays 3 and 4 as regular drives without any problem.
Yes. I have a two drive RAID 0 and two other drives in bays 3 and 4 as regular drives without any problem.
I concur with the above post. P2 R4 is pretty flexible. You can do any combination you like, so long as the parameters you seek fall under the minimum requirements of the array type you want.
What you want to do is perfectly feasible.
I think they are dishonestly backtracking. A lot of people including myself contacted Blackmagic and were told the chips are SATA II. I've also seen benchmarks that top out at 350 MB/s read speeds. At this point Blackmagic should make a press release to put the questions to rest.
Finally received my nMP.
I tried my original JBOD and RAID0 w/ 2 SSDs on my TB1 port and I got 450MBps on both cases. However, on my TB2 port I got 850MBps for the RAID0 and the usual 450MBps for the JBOD.
It would appear that TB2 definitely makes a difference but I don't know if that is due purely to the TB connection type or the firmware running it in the P2 R4.
Thought I would share.
I'm sure with 4 SSDs, it make begin to congest bandwidth though. Probably capping at around 1400MBps or so.
I tried my original JBOD and RAID0 w/ 2 SSDs on my TB1 port and I got 450MBps on both cases. However, on my TB2 port I got 850MBps for the RAID0 and the usual 450MBps for the JBOD.
It would appear that TB2 definitely makes a difference but I don't know if that is due purely to the TB connection type or the firmware running it in the P2 R4.
Thought I would share.
I'm sure with 4 SSDs, it make begin to congest bandwidth though. Probably capping at around 1400MBps or so.
How much of a performance hit is the Write Thru.. setting ? vs. Write back.
I'm curious if its like 150 mb/s penalty?
I just setup my Pegasus2 R4 lastnight night in RAID5 .. 4 x 2TB - Hitachi 7K4000 - HGST HUS724020ALE640
Stripe Size: 256KB
Sector Size: 512 bytes
Read Cache Mode: Read Ahead
Write Cache Mode: Write Thru
Updated the firmware to 5.04.000.18
Plugged into a nMP 6 core over TB2. Only one monitor hooked up atm, on TB bus 1, P2 r4 is on TB bus 2 AFAIK.
Tested with Blackmagic and AJA..
I'm getting about 195 MB/s write, and 440 MB/s read.
I'm curious if its like 150 mb/s penalty?
I just setup my Pegasus2 R4 lastnight night in RAID5 .. 4 x 2TB - Hitachi 7K4000 - HGST HUS724020ALE640
Stripe Size: 256KB
Sector Size: 512 bytes
Read Cache Mode: Read Ahead
Write Cache Mode: Write Thru
Updated the firmware to 5.04.000.18
Plugged into a nMP 6 core over TB2. Only one monitor hooked up atm, on TB bus 1, P2 r4 is on TB bus 2 AFAIK.
Tested with Blackmagic and AJA..
I'm getting about 195 MB/s write, and 440 MB/s read.
It's good to know that they are still using Hitachi's ( HGST ) drives as I plan on going ahead with the replacement, taking me up to 8TB of space. I've put my 4X 1TB drives up for sale ( still under warranty) and have been shopping around for the 2TB's..best price for 4 delivered so far is £116.00 each...I may be able to do better and can't do the upgrade until the others have sold anyway. In RAID 0 my R4 benches at the speeds shown.
Attachments
Actually No.. I bought these.. enterprise 7K4000's .. this is in a "diskless" P2 R4.
I hated what I saw on the promise compatibility list, i wasn't gonna buy a Sata2 drive with 32 mb cache.., with 1 yr, 3 if your lucky yr warranty for a raid.
That's certainly changed a lot since I bought mine...5 Years on the drives, 2 on the controller. They charge enough for them...Makes you wonder if they want to sell them at all.
I wouldn't worry about that list....As long as you buy enterprise you should be fine...Heck, if you wanted to cheap out on speed and performance I reckon you could probably put greens in there, although it's a little pointless.
I'm sourcing Hitachi because they've proved to be reliable, not because they are on the list....Best price so far on 4X2TB's is £116 each delivered or $194, but I'm still looking...Just put my 4X1TB up for sale...they should go as they still have warranty on them...I have all the media backed up on a 4TB USB 3 drive so it won't be hard after re-building the array.
Sounds right. I got almost the exact same thing when I had my 4 HDD RAID 5 setup. Using the same drives as you actually.
But you can freely switch between writeback and writethru without resynchronizing. The penalty is kind of hefty at about 200 MBps for write but you gain about 150 MBps of read in my experience.
OK cool to know. I will do a test tonight. I haven't found a lot of info on the writeback vs. wrote thru feature. I just know I choose the "safer" option w/penalty. I am guessing safter means less chance for data corruption?
I did a test with the Write Back setting selected. Pretty interesting.
I just setup my Pegasus2 R4 is in RAID5 .. 4 x 2TB - Hitachi 7K4000 - HGST HUS724020ALE640
Stripe Size: 256KB
Sector Size: 512 bytes
Read Cache Mode: Read Ahead
Write Cache Mode: Write Back
Updated the firmware to 5.04.000.18
Plugged into a nMP 6 core over TB2. Only one monitor hooked up atm, on TB bus 1, P2 r4 is on TB bus 2 AFAIK.
Tested with Blackmagic and AJA.. I used the first Blackmagic numbers, because I think the cache comes into play in the 2nd round which spikes the numbers into the 600 even 700 range. So my numbers below are avg.'s after 1 BM, and 2 AJA tests.
I'm getting about 465 MB/s write (+270), and 340 MB/s read (-100) compared to with the "write thru" setting enabled.
Does the P2 raid controller have a battery backup, to negate the need for write thru?
I need to think about my avg. use case. If I would rather have and extra 100 mb/s read, or a plus 270 mb/s write?
I think this raid will mostly be media and bulk storage (read).. and scratch disks will be on the internal 2x-4x faster 1TB SSD.
I just setup my Pegasus2 R4 is in RAID5 .. 4 x 2TB - Hitachi 7K4000 - HGST HUS724020ALE640
Stripe Size: 256KB
Sector Size: 512 bytes
Read Cache Mode: Read Ahead
Write Cache Mode: Write Back
Updated the firmware to 5.04.000.18
Plugged into a nMP 6 core over TB2. Only one monitor hooked up atm, on TB bus 1, P2 r4 is on TB bus 2 AFAIK.
Tested with Blackmagic and AJA.. I used the first Blackmagic numbers, because I think the cache comes into play in the 2nd round which spikes the numbers into the 600 even 700 range. So my numbers below are avg.'s after 1 BM, and 2 AJA tests.
I'm getting about 465 MB/s write (+270), and 340 MB/s read (-100) compared to with the "write thru" setting enabled.
Does the P2 raid controller have a battery backup, to negate the need for write thru?
I need to think about my avg. use case. If I would rather have and extra 100 mb/s read, or a plus 270 mb/s write?
I think this raid will mostly be media and bulk storage (read).. and scratch disks will be on the internal 2x-4x faster 1TB SSD.
Last edited:
Just to explain the difference between Write Through and Write Back…
Write Through: When writing data to the disk, the cache is not used. Instead, OS X will tell the Pegasus to write the data to the drive, and the Pegasus waits for the data to be completely written to the drives before returning the successful completion.
Write Back: This uses the cache when writing data to the disk. In this case, OS X tells the Pegasus to write a given block of data to the disk. The Pegasus saves this block quickly in the cache and tells OS X the write was successful. The data is not actually written to the disks until some time later (not too much later, just as soon as the disks can seek to the right location and perform the write operation).
Setting it to Write Back is "less safe" than Write Through. The safety issue comes into play during a power outage. If the power goes out between the time that the Pegasus told OS X the data was written, and the time when the data is actually on the disks themselves, data corruption could take place.
However, I think the chance of corruption is pretty small because as far as I know (can't find the source at the moment), the Pegasus2 has a battery-backed cache. [Edit: Page 116 of the manual claims there is a battery for the cache.] If a power outage happens as described above, the battery should power the cache until the power goes back on and the Pegasus can finish writing the cache to disk.
As for the read speed decreasing when Write Back is enabled, the only reason I can see for this to be the case is that there is less cache available for reading (because some is being used for writes). The difference shouldn't be anywhere near 100MB/sec though. Not sure why you are seeing that.