OT: Saving & Opening CS6 Files across a network

@bentoms The problem with SMB is it lacks the auto-reconnect feature of AFP 3.x, so there is a 120 second window for any network anomaly to resolve itself, so the session ID is preserved and open documents can be saved without fear of corruption.

We've been supporting Xinet, Helios, etc., environments for years, and integrators like IO Integration, NAPC, etc., design and build infrastructure allowing users to work off the server. I think some solutions even allow locking the file on the server so only one person can work on it at once, although SMB is still a problem (from what we've seen).

Personally, I'd go with what Adobe supports...they're your support, so if you don't follow their lead and the $#!+ hits the fan, you might find yourself stranded.

Don

I guess we're risk takers. We have a really reliable and fast network (most places 1Gb to the desktop) and really reliable W2K8R2 servers (now on Nexus hardware with 10Gb uplinks) with fast disk (Direct Attached Storage, Dell MD1200 I believe) on the shares that need it (although we're not coming close to peak IO). Now, on the Windows workstation side of the house we're even ballsy enough to capture and edit HD video projects over the network on the Windows share. We've rarely had issues. On the Mac side of the house there isn't as much video editing in Premiere Pro, more iMovie so you HAVE to work locally really, but a lot of use of Adobe Photoshop, Dreamweaver, and InDesign and they often work over the network via SMB, to the same servers. Unless I'm not being told things, no real complaints. You didn't really say what products you were looking to use. Are Adobe's recommendations the way to go if you have support issues, absolutely. I haven't had issues that have required support and I fear for my life and time if I actually needed their support. Working locally will always eliminate the "what if" the network or server goes down. I would argue the same about "what if" the internal hard drive failed on the workstation, unless you have RAID? At least in my server environment there is RAID to help protect against that. Also, the business of copying up files, especially large raw video files, adds additional time going back and forth. For us the risk has been worth the reward in time saved, access to files from any workstation, ease for the user, and data integrity. If you lack confidence in your servers and network reliability (which I have all the confidence in the world in ours) then may want to contemplate that.

Thanks both.

FWIW we've seen issues like: DreamWeaver keeping files open after the document is closed, closing DreamWeaver on the mac that last opened the file resolves.

Our users mainly use Illustrator, Photoshop, DreamWeaver & Flash (those whom edit video always work local).

I posted this to the mac enterprise list & was given some versioning solutions to look at, so will have a nose @ that too.

We're also migrating these clients from 10.6.8 & cs5 to 10.8 & CS6.. this should help as one of their biggest complaints is share browsing speed. Maybe the speed increase will allow them to give the work local then upload worlkflow a go?

Oh, & the links to Adobe's notes:

http://helpx.adobe.com/premiere-pro/kb/networks-removable-media-dva.html
http://helpx.adobe.com/photoshop/kb/networks-removable-media-photoshop-cs5.html
http://helpx.adobe.com/illustrator/kb/illustrator-support-networks-removable-media.html

Adobe's stance for many, many years (and yes, all of us know this stance) has always been "don't work off of network shares, work local". It's the same song and dance Quark used back in the day. In the 15 plus years I've been working in the pre-press/printing/advertising world, I've had very, very few issues with working off a network share (from simple AFP share to Xinet share to ExtremeZ-IP share). The one issue that I see fairly consistently here is a Photoshop saving issue. It is described in this fairly long Adobe forum post:

http://forums.adobe.com/thread/732422

Our environment here is a 9 TB file share on a Promise RAID (RAID 50) connected via FC to an XServe running 10.6.8 hosting the share via AFP. The server has two 1gig connections (bonded) to the network, and all of my users are running on 1gig unless on wi-fi. And it's a variety of client OS from 10.6.8 up to 10.8.2 all running CS 5.5 DP.

My users are used to this Photoshop issue and know how to work around it, but it is annoying.

Although I don't work in the creative field anymore, I had once done so managing Macs in such an environment for over a decade, and I have to echo what stevewood is stating. Yeah, it might be considered "bad practice" to work directly off a server according to Adobe, but for years we did exactly that and had very few issues. The one application that was the most susceptible to files becoming corrupted if there was any issue with connectivity was Photoshop. We occasionally would get a file that just wouldn't open (or if it did it was all weird looking) if it was left open and a server went offline or someone's Mac crashed just when saving the document. But that's what you have a good backup plan for right? Right??
Other than that, the 'work off the server' workflow was generally fine. Besides, getting creative users to comply with a copy locally, work, save, then copy back process, though possible, was akin to herding cats.

I would take a look at some versioning solutions to help make that easier if you plan to work that way and not right off the server.

Just my 2¢ for what its worth (probably only about 2 cents)

Hi Ben

Adobe been saying this for years and Don is right that the smb shares don't have the auto reconnect and causes quite a bit of hassle.

We use smb shares and tell people to work locally, but the designers don't like it because they don't want to have to relink things all the time. To get around this they work off the servers.

We have net app filers with lot of LUNS that take snapshots throughout the day. If someone's file gets corrupt at least we can revert instantly to a previous version.

Thanks guys!

We only backup nightly :(

I guess the mac guys use around 10 w2k8 servers globally, with around 100 shares. So I'd like to avoid the ExtremeZ-IP route if possible.

I'd rather get educate them in a better way of working across the estate. But if it works I should have a look.

I thought that too that's why I didn't want to go around extremeZ-ip as well.

@stevewood wrote:

In the 15 plus years I've been working in the pre-press/printing/advertising world, I've had very, very few issues with working off a network share (from simple AFP share to Xinet share to ExtremeZ-IP share).

Xinet, Helios, etc., is sold as a workflow solution that facilitates working off the server, whether it's built on AFP 3 or SMB. So your open document survives and remains open (read: user can continue to work). Without auto-reconnect, a user may have the option to Save-As (or whatever it's called now) to the Desktop (say goodbye to defined workflow), or they may not be able to (say goodbye to hours of work).

Depending on how these solutions are implemented, if a user works locally it may "break" the defined workflow (OPI, links, etc.).

SMB doesn't protect you if there is an anomaly, AFP 3 does (up to 120 seconds and your session ID is preserved). There simply is no comparison, although Wintel techs would argue otherwise (they simply don't understand the intricacies of the Mac platform - it's our job to "teach" them so they can accommodate our needs).

It's all about risk...I prefer to take the least risky route whenever possible. Thankfully I haven't had to deal directly with users in the past several years, instead we define workflows and processes and the groups enforce our recommendations (or risk impact to productivity).

YMMV...if it works for you, great. Those of us who've been burned know not to stick our neck out.

I spent 13 years doing production work, and the next 13 years supporting those environments. I transitioned to IT because the folks who "supported" our Macs were knuckleheads who didn't care and couldn't be bothered - the production folks were the squeaky wheel; the 10%. So if/when we lost productivity, it was a joke to them. Trust me, it took less than a year before they dotted line reported to me, and the ones who understood the critical nature of their job continued to work there - and those jokers, well, I hope they have fun flipping burgers. Sorry for the rant, but the issue of risk is a sensitive issue for high intensity production environments, and I hope everyone makes their decision based on continuity and not based on what's easiest for them.

Don

Hey Don, Sorry to bring up a ghostly horse and get to beating it, but what are you doing now that Apple no longer officially recommends AFP shares?

Hi Toby @Knighton, Apple doesn't make AFP the default but it is still quite alive.