Post Reply 
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Questions about Drobo 5D performance
08-14-2013, 04:47 PM
Post: #11
RE: Questions about Drobo 5D performance
Drobo (in general) is aimed more at the "black box that just does its thing" market.
Having built my own RAID boxes, the plus is definitely low maintenance and ease-of-use. I could easily walk my sister (my wife is now too geeky to use as an example) through replacing out a drive.
The minuses is not being able to get to the "nitty gritty" stuff like individual drive SMART data, performance tuning, etc.

Speed in dual parity (Dual Disk Redundancy aka DDR) doesn't seem to change much, as far as I can tell.

Enterprise SSDs are not supported - probably because they're simply drawing the line between consumer and enterprise products.

Rebuild time - I can't say for the 5D's faster CPU, but on my Drobo S units, it's approx 24 per protected TB used, or less, barring any further drive issues.

Near-instantaneous rebuild? If you reinsert a drive before the unit goes into rebuild, it's instantly re-added.
New drives are added pretty much instantly, but if rebuild is in progress, rebuild will have to go to completion.
In my experience it can take about a minute before Drobo "notices" the drive went away, but that's very anecdotal.

--Brandon | WHS2011+Drive Bender/2x Drobo v2/Drobo S G1/ Drobo S G2/Transporter
Drobo provides fault-tolerance, it's NOT a substitute for regular backups.
Drobo Best Practices - Official and Community-sourced.
Find all posts by this user
Quote this message in a reply
08-14-2013, 05:11 PM
Post: #12
RE: Questions about Drobo 5D performance
(08-14-2013 04:47 PM)bhiga Wrote:  Speed in dual parity (Dual Disk Redundancy aka DDR) doesn't seem to change much, as far as I can tell.

Based on what tests? Any numbers? Dual redundancy is costly for write I/Os and calculations, usually bringing down write speeds significantly, especially random I/Os. 5D can't be exempt from it, especially that its brain isn't a quad core i7. Read speeds - depends.

(08-14-2013 04:47 PM)bhiga Wrote:  Enterprise SSDs are not supported - probably because they're simply drawing the line between consumer and enterprise products.

They support enterprise class SATA drives. Don't think that's the reason.

(08-14-2013 04:47 PM)bhiga Wrote:  Rebuild time - I can't say for the 5D's faster CPU, but on my Drobo S units, it's approx 24 per protected TB used, or less, barring any further drive issues.

24? Can't be minutes, gotta be hours? Ouch. That's roughly 10 times slower than some RAID6 units I've tested.

(08-14-2013 04:47 PM)bhiga Wrote:  Near-instantaneous rebuild? If you reinsert a drive before the unit goes into rebuild, it's instantly re-added.

Sorry if I wasn't very clear. Here are the steps:

Take a half-full (or nearly full - doesn't matter) Drobo 5D in single or dual parity mode:
1. Take out a drive (doesn't matter, "dirty" or gracefully)
2. Write 1GB to the Drobo while it's in degraded mode
3. Re-insert that same drive
4. Measure time it takes to get back to "fully healthy"

How long will it take to rebuild? Roughly - is it measured in minutes or hours? Or days? The question here is, does it behave like MSS and Linux RAID, only updating what's missing within seconds or minutes, or like most vanilla RAIDs, initiating a full rebuild that takes hours and days?
Visit this user's website Find all posts by this user
Quote this message in a reply
08-14-2013, 06:20 PM
Post: #13
RE: Questions about Drobo 5D performance
Hi, DroboS owner recently upgraded two disks from 2TB -> 4 TB disks. My total volume is 8+TB of data.
Rebuild when upgrading these disks was approximately 20 hours for the entire 8+TB volume.
In SDR FTR.
I only saw 24 hours per TB in my 4bay older Drobov2.
Find all posts by this user
Quote this message in a reply
08-14-2013, 06:24 PM
Post: #14
RE: Questions about Drobo 5D performance
Thanks Yoyoma. Wonder if 5D rebuilds even faster.
Visit this user's website Find all posts by this user
Quote this message in a reply
08-15-2013, 04:12 AM
Post: #15
RE: Questions about Drobo 5D performance
On some of the older drobos they said that dual parity knocks 10-20% off speeds

Drobo DOES support a "quick check" where you can pop a drive out, re insert it very shortly later, and you are back to healthy in about 1-2 minutes.... But I have no idea what criteria it uses to decide if it can just quick check or needs to do a full rebuild, so don't bank on it

My Avatar is my Panther Chameleon named 'Leon.

I have a owned a Drobo v1, v2, DroboPro, Drobo 5D, DroboFS, Drobo 5N and DroboPro FS

I currently have a Windows 2016 Server with 8 x 8TB Seagate Archive Drives, 14 x 8TB WD Red and 3 x 8TB WD Gold
Find all posts by this user
Quote this message in a reply
08-15-2013, 08:54 AM
Post: #16
RE: Questions about Drobo 5D performance
(08-14-2013 05:11 PM)beans Wrote:  Based on what tests? Any numbers? Dual redundancy is costly for write I/Os and calculations, usually bringing down write speeds significantly, especially random I/Os. 5D can't be exempt from it, especially that its brain isn't a quad core i7. Read speeds - depends.
My observation is based on my Drobo S, and well, it's not the fastest, so it may simply be 10-20% of slow is a tiny difference. Smile

(08-14-2013 05:11 PM)beans Wrote:  They (Enterprise SSDs) support enterprise class SATA drives. Don't think that's the reason.
The B1200i handles SSDs a bit differently from the 5D. On the B1200i, they're just part of the drive pool (they replace drives) and data is intelligently (magically?) tiered. On the 5D the SSD has a dedicated slot and is a dedicated cache.

(08-14-2013 05:11 PM)beans Wrote:  24? Can't be minutes, gotta be hours? Ouch. That's roughly 10 times slower than some RAID6 units I've tested.
Sorry, was rushing out the door - yes hours, not seconds or minutes.
Yes, Drobo rebuild times are slower than vanilla RAID. One thing to note, however, is that Drobo rebuild is based on used storage, not the raw array size. Thus, rebuild time for a given configuration is not a constant as it is with vanilla RAID.

The 24 hours/TB used is a maximum, and as yoyoma says, the newer units are much faster. Depending on the drive that was removed and how much data was actually being stored on it, it can be much, much faster.

The only real way to get hard numbers for this is to specify a fixed configuration of drives and stored data, though.

(08-14-2013 05:11 PM)beans Wrote:  Sorry if I wasn't very clear. Here are the steps:
1. Take out a drive (doesn't matter, "dirty" or gracefully)
2. Write 1GB to the Drobo while it's in degraded mode
3. Re-insert that same drive
4. Measure time it takes to get back to "fully healthy"
I haven't tried this myself, so I could be wrong here, but to my understanding so far, #2 (the write) will force a rebuild/downsize based on the remaining storage, if it's not already happening.
#3 When the drive is re-added, it will be treated as if it was a new drive being added to the pool.
Therefore, the answer to #4 is that it's definitely not MSS/Linux instant rebuild, because of the rebuild that already started between #1 and #2.

Simply adding a drive is pretty much instant, the migration of data onto the new drive is handled silently in the background.
But once a rebuild starts, there's really no (happy) way to get out of it.


As you can see already, there are a lot of unknowns in the Drobo world. Fact of the matter is, it's not designed for RAID-knowledgeable people. Most of the "smarties" here that are RAID-knowledgeable (rambo, rdo, Docchris, diamondsw, etc) landed here for reasons other than maximum speed.

--Brandon | WHS2011+Drive Bender/2x Drobo v2/Drobo S G1/ Drobo S G2/Transporter
Drobo provides fault-tolerance, it's NOT a substitute for regular backups.
Drobo Best Practices - Official and Community-sourced.
Find all posts by this user
Quote this message in a reply
08-15-2013, 01:44 PM (This post was last modified: 08-15-2013 01:54 PM by beans.)
Post: #17
RE: Questions about Drobo 5D performance
(08-15-2013 04:12 AM)Docchris Wrote:  Drobo DOES support a "quick check" where you can pop a drive out, re insert it very shortly later, and you are back to healthy in about 1-2 minutes...

Thanks Doc, that helps.
(08-15-2013 08:54 AM)bhiga Wrote:  As you can see already, there are a lot of unknowns in the Drobo world. Fact of the matter is, it's not designed for RAID-knowledgeable people. Most of the "smarties" here that are RAID-knowledgeable (rambo, rdo, Docchris, diamondsw, etc) landed here for reasons other than maximum speed.

Appreciate the answers - thanks Brandon!

In video, it's consistency and predictability of performance that matters the most, besides pure speeds. Hopefully Drobos will catch up in that department, as well as in value, now that there are some interesting low cost alternatives.
Visit this user's website Find all posts by this user
Quote this message in a reply
08-15-2013, 02:53 PM
Post: #18
RE: Questions about Drobo 5D performance
(08-15-2013 01:44 PM)beans Wrote:  In video, it's consistency and predictability of performance that matters the most, besides pure speeds. Hopefully Drobos will catch up in that department, as well as in value, now that there are some interesting low cost alternatives.
Most definitely. I think for specific uses, Drobo can meet the needs. We have a Drobo Pro that we use at work, connected on iSCSI, but we use it more for archive and "little" edits. Anything major we work on local workstation RAID.

For general purpose video editing, especially for contract-type houses where the formats and editing style can vary a lot from one project to the next, the scalability and "hands off" management of Drobo is countered.

A properly configured and built video editing station's local RAID storage will definitely outperform a consumer-level Drobo, and as long as the editor is busy editing and not breaking their system by messing with things, the scalability and ease-of-management advantages of Drobo don't apply as strongly.

However, Drobo does well as a fault-tolerant, easy-to-expand backup/archive device, and for more-general file storage.

--Brandon | WHS2011+Drive Bender/2x Drobo v2/Drobo S G1/ Drobo S G2/Transporter
Drobo provides fault-tolerance, it's NOT a substitute for regular backups.
Drobo Best Practices - Official and Community-sourced.
Find all posts by this user
Quote this message in a reply
08-16-2013, 05:49 AM
Post: #19
RE: Questions about Drobo 5D performance
hi beans,

just fyi, i've done some editing on the older Drobo gen1 and Drobo gen2 (4-slots) without problems as mentioned here:
http://www.drobospace.com/forums/showthr...http://www.drobospace.com/forums/showthread.php?tid=2046&highlig

i've also done some semi-pro editing on a 2nd gen Drobo-S (5-slots) in DDR mode (3 drives at the time), and didnt have any problems.

Usually with the Drobo-S, i start with the source material on the main computer's hard drive, and output the result to the same drive (or the drobo-S) i think i've also used source material from the Drobo-S to output the other way when i didnt have enough free space on the computer.

sure, the rendering/processing still takes "time", what with a couple of video tracks, audio, and other enhancements/efffects and PinP visuals, but i never felt that it was slow Smile

usually you can speed things up by turning off certain programs or services (or windows indexing/virus scanning) while you do it, as those seem to add lots of slowdown on their own)


if you mean USB Attached SCSI support, im not sure, but if you are planning on setting up video storage systems for others (like a service), then the best way is to actually get some 1st hand experience for yourself....but it never hurts to do some "free" research before splashing out on a purchase) Smile


Is your Main Objective to provide a hardware solution which has a large amount of storage space, with built-in drive failure features? if so, i think a 5d is good.

is your MO, to provide semi-pro setups for such stored video, as above, but also to provide a way to pull in multiple video's for the purpose of 1 person editing, then it should be ok, provided that the rest of the system is up to spec... eg the drobo is just one piece of the puzzle... and editing is CPU intensive, and the software used needs to be up to spec too.

if your MO is to provide a large safer storage system, which can allow multiple people (workers) to to simultaneous editing to and from the drobo, then i think you'll experience considerable slowdown (based on what i've read)

can you be a bit more specific with your MO? Smile

(btw i have XP home SP2, a Drobo v1 with 2x 1TB/2x 1.5TB WD greens, & a bkp Drobo v2 with the same + a DroboShare: unused)
& a DroboS v2 with 3xWD15EADS &2x1TB in DDR mode on win7, & a drobo5D (all usb)
  • btw i did a sustained (write) operation for about 6 hours, and got 13.2MB / sec ...objection? "sustained" :)
    (16.7MB/s on a v2 & 47-96MB/s drobo-s)
Find all posts by this user
Quote this message in a reply
08-16-2013, 07:10 PM
Post: #20
RE: Questions about Drobo 5D performance
(08-16-2013 05:49 AM)Paul Wrote:  can you be a bit more specific with your MO? Smile

Hi Paul. It sounds like you'd like to figure out if 5D will work for me based on my "MO". That won't be necessary, thank you. I just want answers to the questions I asked.

Thank you for your thoughts and suggestions though. While 5D (much like single 5400rpm drives) can be used for editing, it's rarely if ever used by professional editors, for reasons I mentioned already, and which other posters helped me figure out.

My "MO" is consulting, systems and storage solutions for professional editing and content creation workflows. ("Professional" here isn't about content quality, only the "MO": systems that are used for editing a minimum of 8 hours a day, often more.)

HTH.
Visit this user's website Find all posts by this user
Quote this message in a reply
Post Reply 


Forum Jump: