C1 11 & Synology DS718+

NN636531089864297925UL

• February 01, 2018 18:26

Hi all, I've just rolled out a new Synology DS718+ for a client who is working with C1 11 and has all the updates for C1 11.

The client has been working directly from External Drive so moving to NAS is new for him. Everything is setup fine, I have run Black Magic speed tests to the shares using both AFP and SMB and i'm getting the full 100 to 120 MB/s read and write speeds to NAS. I ran a black magic disk test to his old External Drive he uses and that was 35 MB/s read and write so the new NAS is reading and writing much faster then the external drive he's been using previously.

The issue I am having is when opening and working within a session folder on the NAS it's very slow. I'm waiting around 10-15 seconds for the session to open over the NAS but the same session folder off the External drive opens immediately and working within the sessions is much quicker using direct attached storage even though the read + write speeds are 3 x faster on the NAS.

The Client works with sessions and not catalogues. All the session files / folders are stored directly on the NAS

Is there a limitation or issue with NAS and C1 11? Is direct attached storage still the best method for my client working with C1 11?

Thank you.

0

• 

  Jeffrey Jakucyk

  • February 01, 2018 18:43

  Seems that network attached storage just doesn't work that well with Capture One. My guess is that the overhead of network operations (whether inherently or due to Apple's SMB file sharing implementation) is what kills performance. It's better on Windows, but still not as good as direct-attached storage.
  
  I recently switched from a 2008 Mac Pro to a new iMac Pro. Since I have nearly 8TB of data on the old Mac Pro I'm using it for the time being as my repository of RAW files accessed via file sharing over Ethernet. It works OK, but it's not great. Rendering of the images on the iMac is actually a bit slower than if I was using Capture One on the old Mac Pro itself, despite being a gigabit connection. It's fine for seldom-accessed archives. Of course if I copy the RAW files to the iMac's internal SSD then there's no contest and it absolutely flies.
  
  The sessions-based workflow might not be the best for this type of setup either since all the data (the RAW photo files AND the adjustment sidecar files) will be on the NAS. With a catalog-based workflow (which is the only one I'm familiar with honestly), the catalog can reside on the fast internal hard drive, and as long as preview size is kept to a minimum it doesn't grow much. The benefit is that adjustments can be made even if the photo access is laggy since all the adjustment data and organizational information is in the catalog file itself. It's still not ideal, but it at least limits network I/O to reading only the photo files, not the photo files and the sidecar data.

  0
• 

  SFA

  • February 01, 2018 21:38

  I have no idea about Mac connections but using sessions with a PC (Win 7) and a recently purchased Synology 218+ is initially slower than internals SSDs (no surprise) and external USB3 drives (once they have "woken up" from a low power state.)
  
  Up and running the NAS is not too bad especially considering that the drives are only 5400rpm platters as it is intended mainly as archival storage with active sessions being kept locally.
  
  I left the NAS configured with Synology's own formatting system.
  
  One thing I noticed when doing some test loads from existing disks using the USB3 copy facility is just how much the available speed is maximised for large files and slows to a crawl for small files. Sessions have a lot of small files.
  
  With Windows once a session has been opened and the working files have been loaded into memory cache editing can be quite responsive despite my slow spinning disks but there is almost certainly a possibility to defeat the C1 "predict ahead" functionality if jumping around in a session rather than working through it. The effect of that on the shuffling of small files to and from the NAS is likely quite obvious compared with internal SSDs or any other internal storage or non-sleeping USB3 (or better) connected device.
  
  I should mention that I am connecting to the NAS via Wi-Fi, the NAS being cabled to a Wi-Fi extender device. Clearly one might expect better using a good cable connection. However the read and write speeds obtainable when working with a large number of small files are, so far as I can discover, always going to be very compromised compared to what is possible when running with large files.
  
  That said it works pretty well as a mapped drive once up and running.
  
  Also I have not yet had much chance to dig into the tuning and operational refinement options. that's something I hope to be working in in the coming days.
  
  Grant

  0
• 

  Permanently deleted user

  • February 02, 2018 21:24

  From my experience there is one thing you should do to improve the performance on OSX: disable AFP and rely on the SMB protocol.
  For whatever reason SMB is much faster and also more stable although it's quite a bit more work to get it to work (you have to connect to the server-IP in the Finder) and it's less comfortable for "normal" access via the Finder, but the lags are significantly reduced and the transfer speed is significantly improved and somewhat close to what you might expect on Windows.
  
  I would also suggest to only work on a cabled network with your NAS, WiFi is comfortable but simply too slow and lags a lot.
  
  Though I have no experience with Catalogs on OSX, I have recently worked a lot with one on Windows and I can tell you for a fact that it's much faster to work on a local Catalog which includes the files on a NAS than having the catalog stored on the NAS as well. But you can set your catalog to create backups on the NAS.
  
  I strongly advise you not to use Session on a NAS - they are much too slow and cumbersome and very unreliable; jjakucyk is absolutely right about his remark about those many small files killing the performance.

  0

Post is closed for comments.