bradmurray

Members
  • Posts

    15
  • Joined

  • Last visited

Posts posted by bradmurray

  1. I have this problem all of the time. For example if I modify a bunch of metadata in MP3 files and they have to be re-synced I will get this on a bunch of random ones. The only solution I have found is to stop both clients, manually copy the file, delete the *.SyncPart and *.!sync files in the destination folder, and restart syncing. I usually pause the client on my macbook before I do the batch mods and resume it when I'm done so I don't think it is an issue with it trying to pick open files.

  2. On the Android and iPhone clients it would be nice to be able to set the "use cellular data" option at the share level. I have a share that contains large video files that I don't want chewing my bandwidth, but I'd like to keep everything else up-to-date all of the time.

  3. Surely the first letter does not control the secret's function.

    It doesn't control it, but it does designate the type of key. Changing A to B will not change the function as it will invalidate the rest of the key. You can see this by comparing the full key for a folder you already have to the read-only key that it gives you if you ask for it. It's just a human-readable thing. They do the same thing with credit cards. Visas always start with 4 (MC with 5, AmEx 3 Discover 6). Changing the first number doesn't change the card type. It invalidates the number because there is a built-in checksum.

  4. I'd vote for a more rsync like behavior of the temp file while syncing, i.e for file file.ext, make the file.ext!sync a copy, and move to the original file (i.e overwriting/replacing it). This would be a huge benefit when syncing webpage content, as there'll be no 404 while a sync of a file is in progress.

    Along this line (requested by somebody else) would be the option to have this copy temp-file in a separate folder and not alongside the file, in order not to expose partially synced files.

    I do see the purpose of the current behavior for many scenarios though (i.e preventing someone from concurrently modifying the file, saving disk space as you don't end up using twice the space during transfer, etc.), so it would be very nice if this could be configurable on a per-folder basis.

    Second on this. I have a directory that is running on BTSync between two computers, but it is also being mirror to an online backup company. The mirroring program is always trying to mirror these !.sync files. If I had a temp folder on the same physical volume where all of the !.sync files could go while they are in progress it would make it a lot cleaner with this and many other apps. I have planned on using BTSync to publish web code, but this would cause a problem there as well. The temp directory should name files as the full destination path with the slashes converted to ~ or something like that so if it gets stuck you know where the file is supposed to go.

  5. If we had decent logging and a record of every device that connected and disconnected from a share we would at least be able to monitor what is going on and maybe even put decent alerts on any time a new device connects. Heck, you could have an option to require a manual acceptance of a new device into a folder by at least one of the existing participants.

  6. I have a 3.5GB file that I am trying to sync from Windows 7 (1.1.27) to Android (1.1.7 on a Galaxy S4 going to the external SD card) and it has been running for a few hours. The Windows client says that it is running mostly over 3MB/s so it should have finished in about 15 minutes. On the phone the KB/s status keeps popping up and disappearing so it looks like nothing is happening there, but Windows continues to show progress. The folder that is syncing is very small and only has one other file in it that is a few hundred MB.

  7. I just ran into this problem in Windows/OSX as well. I am syncing a file where the full path is 250 chars. If you add on the .!sync extension for the temp file you get to the magic 256 length and the file just syncs over and over again. It is only a 1MB file. I tried closing sync and manually syncing the file in both locations and then starting up, but it seems to puke on that length of file name. When I shorten the name everything works fine.

  8. I don't see a problem with multiple clients running on the same machine as long as they are on different ports and sharing different directories. If you write the PID in the shared directory on launch you can keep a second instance from trying to sync the same folder. The only catch would be one session syncing /mnt/a/ and another trying to sync /mnt/a/b/ to it.

  9. I synced a directory of about 20GB (10K files) and there were two files that showed that they were each doing about 1.2MB/s for a while, but the files were only 8MB and 3MB in size. I let it run overnight figuring it was a display bug and the syncing was going on in the background even though each computer already had a mirror of the data. When I woke up the same two files were still going so I paused it. On the Windows side I had dupes of the files. One had the double quote character and the other had a  which looks like a bullet in explorer. On the Mac there was just one of each file with the double quote. I paused syncing and renamed them to 'inch' instead of the character and all is well. I have seen this problem with other syncing products in the past, but just thought I'd raise it.