Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

DIT/Loader Cart

Nick,

GET RID OF FIREWIRE!!!!!!!!!! It is your main problem! Get an eSata card and bump all your drives and Red SSD to eSata. SSD drives connected via firefire are still slow as firewire and a waste. Cheap laptop drives are faster than firewire, so don't waste money on SSD drive until you get rid of firewire.

I would avoid xServe and xSan on set. Your macPro with Atto 6G SAS/eSata card will make your life a ton better. Atto makes one card you can connect 8 eSata devices for about $500 and another that you can connect 16 eSata devices for about $1000.

Thunderbolt will be useful when things a re shipping, but a good eSata card and lots of eSata devices will be faster than Thunderbolt.

The little 1u four drive raid that OWC sells is a nice solution. They also have a square version if you don't rackmount your stuff.
http://eshop.macsales.com/shop/hard-drives/RAID/Rack_Mount/

You can config that as Raid 0,1,3,5 and as Clone to make multiple exact copies. In clone mode your computer only sees one disk in the enclosure, but the enclosure writes the same data to all the drive in it. Nice way to make multiple copies with one write. Lots of ways you can use those to improve your situation.

Hopefully that atleast gets you a start.


Dusty

The reason I haven't pushed for eSata is that our G-Drives only gain a slight (about 30%) benefit from it. I had an eSata card in my workstation briefly, and wasn't receiving much of a real world bump from it. However, it's not really the same scenario. I'm also still wrapping my head around bus speeds and bandwidth. Am I right in thinking that all 4 Firewire connections on a Mac Pro are linked and share bandwidth? So, since I've got 3 devices running on the same Mac Pro, I'm "robbing Peter to pay Paul"? Any specs specifically I should be looking at when I shop around for an eSata card?

If that is the case, the bandwidth from good eSata card would really amp things up. Now that you say it, I'm not sure why I wasn't really considering it . . .

And thanks for the OWC rack suggestion. I'm also looking into something from Synologic, now that Graeme mentioned that he uses them for his home/business stuff. OWC is a name that is popping up more and more when I look into storage; people seem to like them.
 
The little 1u four drive raid that OWC sells is a nice solution. They also have a square version if you don't rackmount your stuff.
http://eshop.macsales.com/shop/hard-drives/RAID/Rack_Mount/

You can config that as Raid 0,1,3,5 and as Clone to make multiple exact copies. In clone mode your computer only sees one disk in the enclosure, but the enclosure writes the same data to all the drive in it. Nice way to make multiple copies with one write. Lots of ways you can use those to improve your situation.

Hopefully that atleast gets you a start.


Dusty

But how does that speed things up when everything still has to go to the client's G-Drives?
 
But how does that speed things up when everything still has to go to the client's G-Drives?

Well, we ARE the client. If we had to swap drives from a rack-mountable OWC device like that on set, to one in our post-production facility, it's not out of the question. My personal workflow would be altered a bit, but I doubt most other people in the office would notice. It may not be an ideal solution, and I'd like to avoid the idea of a "bare" drive floating from shoot location to post-production facility, but still, it's possible.

And also NOT something I would consider if I was a freelance DIT building out my cart. Not without a more polished way to present the drives to a client. Handing over a hard-drive without an enclosure, in my mind, sends the wrong message to the client.
 
And also NOT something I would consider if I was a freelance DIT building out my cart. Not without a more polished way to present the drives to a client. Handing over a hard-drive without an enclosure, in my mind, sends the wrong message to the client.

I was on a music vid last month where they shot 2.5 TB's from 2 Epics in 14 hours, and wanted DNxHD 220 transcodes, so I ended up writing 6 TB's total. I had to stop transcoding for the last 4-5 hours of the shoot because transcoding and downloading the cards at the same time was choking the bandwidth of the drives, and they needed the SSD cards back ASAP. Even if I had a RAID with fast enough bandwidth to do both w/o slowing down, so what? Everything still has to be transferred from that RAID to the client's drives.

What I'm really looking forward to is Thunderbolt G-Raids as client drives for Epic shoots...this month, supposedly...
 
Thunderbolt is garbage.

Firewire 800 is useless.

Laptops are NEARLY useless unless in a ultra-small travel situation.

But that's just, like, my opinion.
Hope you don't mind me refining a bit.

Thunderbolt is evolving.

Firewire 800 is a mediocre interface.

Laptops are powerful when used appropriately and in multiples.

Sorry Dan, couldn't resist.
 
my only gripe with thunderbolt drives is that there are no other connections on it. meaning there is no other way to mount the drive.

A. connectors die after awhile, i'm sure ppl have had their fw800 or even esata ports kick the bucket on the drive itself. and the usb saved their ass in transferring the files off to a fresher drive. TB eliminates that possibly and you'll probably have to rip open the drive and take out the internals and mount that to get to your footage. but most likely it's gonna be a raid so it's gotta go back to the company to get fixed.

B. it also eliminates it as a client drive. face it, clients aren't the most tech savvy, a lot of post houses are using a legacy systems. I've been to post houses where they don't even use esata at all. they need the fw800. so you get all these great offload speeds, but as of now it really wont' be a client drive. they won't know what to do with it more than half the time. not to mention PC implementation is just starting, only imacs and laptops have TB, we don't even have the tower that can use it yet, and even so, post houses, and other freelance post production people don't upgrade their computers as often as your average DIT.

TB has a long way to go, i'm more interested in it as a PCI E interface more than anything else, and i think 6g esata for client drives are gonna the new trend in this next year or two. 6g esata, with all your legacy ports on it for maximum compatibility. i think in a proper 4 by raid 0 with 6g esata, moving somewhere close to 400 mb/sec will be really respectable for offloads on set with how immense file sizes are getting.
 
Joe,

When I am on a show that is shooting tons of footage, I have a 20TB raid 6 that can sustain almost 1000MB/s. I copy everything to that and from that to client drives. I can download multiple SSDs and transcode without maxing out the raid. No Thunderbolt drive can match my big raid.

Dusty
 
Nick,

Yes, there are more firewire ports on a MacPro than independent buses, so robbing petter.... I think most MacPros have two firewire buses. One front and one rear, but I could be wrong on that. A single G-drive might only see 30% speed increase on esata over firewire, but multiple g-drives and multiple SSD readers on firefire will see drastic speed increase when moved to esata, because each esata port has 300 or 600MB/s vs about 100MB/s per bus for firewire 800.

For esata card I would get the Atto H680 (8 sata devices) or ATTO H6F0 (16 sata devies). Both of those cards are SAS cards and you can use a SAS to esata fan cable to connect to esata devices.

G-drives a great. I just mentioned the OWC raid to get you to think about other options.

And for protecting bare drives there are all kinds of protective cases for that. But if the people you are handing the drive to are not competent to not break it, they have bigger problems. The reason to use or not use bare drives should have more to do with the clients ability to use them. I have worked with people who want bare drives so they have less to carry back to the office and others who would know what to do with a bare drive if I gave it to them.



Dusty
 
Good points.


my only gripe with thunderbolt drives is that there are no other connections on it. meaning there is no other way to mount the drive.

A. connectors die after awhile, i'm sure ppl have had their fw800 or even esata ports kick the bucket on the drive itself. and the usb saved their ass in transferring the files off to a fresher drive. TB eliminates that possibly and you'll probably have to rip open the drive and take out the internals and mount that to get to your footage. but most likely it's gonna be a raid so it's gotta go back to the company to get fixed.

B. it also eliminates it as a client drive. face it, clients aren't the most tech savvy, a lot of post houses are using a legacy systems. I've been to post houses where they don't even use esata at all. they need the fw800. so you get all these great offload speeds, but as of now it really wont' be a client drive. they won't know what to do with it more than half the time. not to mention PC implementation is just starting, only imacs and laptops have TB, we don't even have the tower that can use it yet, and even so, post houses, and other freelance post production people don't upgrade their computers as often as your average DIT.

TB has a long way to go, i'm more interested in it as a PCI E interface more than anything else, and i think 6g esata for client drives are gonna the new trend in this next year or two. 6g esata, with all your legacy ports on it for maximum compatibility. i think in a proper 4 by raid 0 with 6g esata, moving somewhere close to 400 mb/sec will be really respectable for offloads on set with how immense file sizes are getting.
 
What you need

What you need

What are the smallest amount of components you would need to be able to have instant dailys and then put them on an iPad by the end of the shoot
 
Joe,

When I am on a show that is shooting tons of footage, I have a 20TB raid 6 that can sustain almost 1000MB/s. I copy everything to that and from that to client drives. I can download multiple SSDs and transcode without maxing out the raid. No Thunderbolt drive can match my big raid.Dusty

It'll be interesting to see what a 12TB Pegasus R6, daisy chained to (2) Thunderbolt G-RAIDS, or any future, similar config from different brands, will do. 10 gigabit bi-directional bandwidth would beat your current set-up, and you could supposedly do it on a laptop, unless those specs are vaporware...
 
Joe,

Just because Thunderbolt does 10Gb/s doesn't mean the drives can max it out. My raid is capable of 48Gb/s on the 10 esata ports it uses, but the 10 drives can only push just under 10Gb/s.

I am interested in seeing how Thunderbolt drives perform. The thunderbolt versions should be about the same speed as the esata versions, because the drives are the bottleneck and not the bus.


Dusty
 
keep in mind these big independent thunderbolt raids are hardware raided. I haven't seen anybody really fill up, let's say one of these pegasus drives yet. I'm hearing different things, but one of them including a SEVERE performance drop once you hit 80% capacity. a ATTO sas card and the software continually tunes a raid and keeps an eye on everything for you with SMART checks, and helps alleviate some of these issues. typically from what I've seen, on a custom raid setup you see about a 30% drop in performance or depending on your drives when you reach capacity, while I read one review on the pegasus that the drive slowed down over 60-70% when it was hitting capacity. there is more than one way to RAID, and not every RAID is equal to one another. hopefully they'll be able to tune a lot of these issues out in the future.

I'm really not impressed with what g tech is doing with TB, it's 2 drives daisy chained together effectively making a 4 bay raid 0. it's freakin impractical. I hope they smarten up and just go for making it a 4 bay chassis, and raiding it like that. it would make a lot more sense that way. since it's RAID 0, it's not like you can plug both drives in different computers and see your data. they have to be daisy chained at all times... it's kind of wack.
 
'm really not impressed with what g tech is doing with TB, it's 2 drives daisy chained together effectively making a 4 bay raid 0. it's freakin impractical.

Whoops you're right, didn't notice they raided 2 of them together for that article...oh well

Thanks for the info gents, good to hear
 
For esata card I would get the Atto H680 (8 sata devices) or ATTO H6F0 (16 sata devies). Both of those cards are SAS cards and you can use a SAS to esata fan cable to connect to esata devices.

Our process is fairly consistent at this point; I don't foresee us ever needing more than 4 devices attached to eSata. 99% of the time we'd be attaching a Red Station, and then two drives. So, any particular thoughts on what model card to get for a Mac Pro with 4 ports on it? Otherwise, I'll just go scan NewEgg, B&H, and Tiger Direct for an hour or so.
 
Joe,

When I am on a show that is shooting tons of footage, I have a 20TB raid 6 that can sustain almost 1000MB/s. I copy everything to that and from that to client drives. I can download multiple SSDs and transcode without maxing out the raid. No Thunderbolt drive can match my big raid.

Dusty

Thought about this for a minute, and I'm just getting more and more curious. I'm trying to figure out how this saves time, while also being safe and secure (in terms of data integrity). At what stage are you running R3D Data Manager, Shotput, or another form of checksum verification? Since you're copying the data twice, how is this saving you time? I see how it would cycle the Red Drives (be they CF Cards, Hard Drives, or SSDs) back into the production pipeline faster. But I only see it taking up more of your time in every other sense. Am I wrong?

Believe me, I'm OK being wrong about this. In fact, I think I'd rather be wrong about it.
 
Back
Top