duch

Members
  • Posts

    26
  • Joined

  • Last visited

  • Days Won

    1

Posts posted by duch

  1.  

    Here's just one example of how to justify the cost of BT Sync:  

     

    - let's say an individual had 2 laptops and a desktop and they want to use a backup service (such as Carbonite)

    - on the desktop you could create a folder c:\backups\laptop01 & c:\backups\laptop02

    - sync the documents and whichever folders you wanted on each laptop to the appropriate folders

    - purchase a Carbonite backup subscription for the desktop

    - you now have a Carbonite backup of all 3 systems for the price of 1 

     

    Ugliest workflow ever.

  2. https://web.archive.org/web/20130811154750/http://forum.bittorrent.com/topic/17782-bittorrent-sync-faq-unofficial/#entry44650

     

     

     

    Will BitTorrent Sync remain free, or will they start charging for it once it comes out of beta?
    BitTorrent Sync will remain free! smile.png

    From BitTorrent:
    "If tomorrow we want to charge you $100 for 10Kb transferred, stop everything related to the app or try to force you not to use BitTorrent Sync, we just physically can't achieve that!
    BitTorrent Sync will work tomorrow exactly like it works today, no matter what we will do. And it will work exactly like today even 10 years from now, of course, if we will have computers in future smile.png
    (Source)

    ...and in publicising the start of the "beta" phase on 17 July 2013, the team commented: "And don’t worry. BitTorrent Sync is still free, simple to use, and secure. Pretty awesome, huh?"
  3. @advhound

    Also note, that Sync requires around 1Kb of memory per file / folder being synchronized. As RPI is pretty limited in memory (only 512Mb), it means that it will definitely have the limitation of ~500K of files total (that is not counting space necessary for OS to run on your Pi).

     

    Even without accounting for the OS the limit will probably be way lower, on my Pi btsync needs 40 to 50 MB of memory to sync 29k files.

     

    Furthermore the 512 MB of memory of the Pi B or B+ can't be wholy used by system and/or user processes, a fraction of it has to be dedicated to the GPU (which is required in the boot process). The minimum is supposed to be 16 MB but I had trouble with values lower than 32 MB.

  4. Can you post specific CPU usage stats? On my NAS, 1.4 increased CPU usage significantly over 1.3.109. Was there a big improvement with 1.4.93 over the previous 1.4's?

     

    Using iostat, prior to BT Sync update, avg-cpu idle was stable around 68% after 24h. I updated four days ago and iostat now report 74% idle, and it is still (slowly) decreasing. I'll do a one hour period iostat and report.

     

    The btsync process itself is idling between 0.0 and 0.9 % of the Raspberry Pi CPU according to htop, from what I remember 1.3.109 was way higher, around 6-10%.

     

    I went straight from 1.3.109 to 1.4.93, so I can't compare with previous 1.4 releases.

     

    I'm wondering if the lower CPU usage I'm experiencing might be related to the xattrs fix (no matter what I did to prevent the linux box from trying to sync them, they still seemed to trigger some activity).

  5. I too switched to 1.4.93 a few days ago as it fixed the port issue. Pretty much the same configuration as travellyan (2 mac desktops, 2 mac laptops, one -cheap- always on linux box, one iPad as a bonus). Around 35 GB of data synced, part backup, part cloud storage and part app preference syncing.

     

    I cleaned every .SyncIgnore of xattrs related stuff on every shared folder of all 5 computers before doing a clean install on the first laptop. It seemed to work fine, so I updated the others.

     

    Everything works, no more xattrs induced headaches, no beef with the interface (I don't get why so many people went completely berserk with it). RAM usage increased slightly on the Linux Box (Raspberry B running Archlinux arm) but idle CPU usage dropped considerably compared to 1.3.109.

  6. I upgraded my 1.3.x systems to 1.4.x and soon found out that, while 1.3.x never had let me down, 1.4.x proved to be unreliable. Peers got lost, synchs got stuck. Since I used btsync a.o. for backup purposes I did not have another option than downgrading, the hard way, to 1.3.x on my most critical systems.

    Currently the 1.3.x systems are whistling nicely. I have one system that it still running 1.4.x and I use it mostly for testing and submitting bug reports.

    Unlike many other users I do not have mental problems with the UI. It's the bugs in the functionality that bug me.

     

    [...]

    Once 1.4.x will be as stable as the old 1.3.x I'll probably be more than happy to try again.

     

    The same here, I've got no issues with the UI, but the core functionality seems way too much full of bug and unreliable in 1.4.x, so I'm still on 1.3.109 everywhere.

     

    Regarding UI, I'd actually like the idea of having a real web UI accessible through a browser (like the linux version) on every platform.

  7. The dot in front of a folder or file name in a Linux / OSx / etc system is used to make it hidden (it won't show up by default on your desktop environnement and using ls without the -a or -A argument on the command line won't display it either).

     

    .DS_Store are files automatically created by OS X to store custom attributes of a folder (icons positions and such). They start with a . so they're hidden on Unix based systems, but they'll show up if the disk is mounted on a Windows system.