Stardust

New Members
  • Content Count

    5
  • Joined

  • Last visited

About Stardust

  • Rank
    New User
  1. Ugh, just rechecked. It seems one of the peers was running a version from Jun 12 (2.0.124). That peer was not in use (ie. no updates made locally, but still on r/w sync). I guess this was the problem?
  2. A quit search of the forums did not turn up this issue, which is surprising as I've had it happen to my files frequently in the last 6 months or so. Specifically, it appends what appears to be logging information to files at random. An example can be found below. The first part is the real file (a rpm package spec file), but after the %doc line, random garbage appears. This corruption is generated on the host on which the file is altered, and propagates to every other host. <snip> rm $RPM_BUILD_ROOT%{_libdir}/*.la %cleanrm -rf $RPM_BUILD_ROOT %files%defattr(-, root, root)%{_libdir}/*.so.* %files devel%defattr(-, root, root)%{_libdir}/*.so%{_includedir}/meos++/*%doc %{_docdir}/%{name} [20150730 11:30:14.411] SyncFolderNotify: SyncFolderNotify: "0EAD2A2BDFBBDC5EE30141C3051CC4811CE3652C.SyncTemp", event = "IN_MODIFY"[20150730 11:30:14.411] [OnNotifyFileChange] "/home/nilss/work/HEAD/packages/.sync/0EAD2A2BDFBBDC5EE30141C3051CC4811CE3652C.SyncTemp"[20150730 11:30:14.411] SyncFolderNotify: SyncFolderNotify: "0EAD2A2BDFBBDC5EE30141C3051CC4811CE3652C.SyncTemp", event = "IN_MODIFY"[20150730 11:30:14.411] [OnNotifyFileChange] "/home/nilss/work/HEAD/packages/.sync/0EAD2A2BDFBBDC5EE30141C3051CC4811CE3652C.SyncTemp"[20150730 11:30:14.411] SyncFolderNotify: SyncFolderNotify: "0EAD2A2BDFBBDC5EE30141C3051CC4811CE3652C.SyncTemp", event = "IN_MODIFY"[20150730 11:30:14.411] [OnNotifyFileChange] "/home/nilss/work/HEAD/packages/.sync/0EAD2A2BDFBBDC5EE30141C3051CC4811CE3652C.SyncTemp"[20150730 11:30:14.411] SyncFolderNotify: SyncFolderNotify: "0EAD2A2BDFBBDC5EE30141C3051CC4811CE3652C.SyncTemp", event = "IN_MODIFY"[20150730 11:30:14.411] [OnNotifyFileChange] "/home/nilss/work/HEAD/packages/.sync/0EAD2A2BDFBBDC5EE30141C3051CC4811CE3652C.SyncTemp" <snip> BTW, do you really trigger a sync on IN_MODIFY and not IN_CLOSE_WRITE? That seems exessive... Can someone move this? It seems I've posted it on the wrong part of this forum.
  3. All machines are running NTP. It would have to be a serious clock drift problem if BTSync suddenly thinks that two week old files suddenly are never than files timestamped minutes ago. Just to make sure, I rechecked all my clocks: ssh host1 'date +"%F %T.%N"'; ssh host2 'date +"%F %T.%N"'; ssh host3 'date +"%F %T.%N"' 2015-04-30 13:15:37.3458029642015-04-30 13:15:37.4863563242015-04-30 13:15:37.489242791
  4. Update. I was able to restore everything overwritten from the .sync/Archive folder, so there is a good thing. What I didn't mention i my initial bug report was that these were old 1.4 shares. I disconnected all clients, and recreated the shares as 2.0(?) shares, lets see how that works out. Also, the setup is 3 peers, Opensuse 13.1 and 13.2 on a LAN, and a Fedora 21 peer outside that (firewall/NAT). The fedora 21 peer was the one falling out of sync all the time.
  5. Same problem. I have been using btsync since December (documents and music) and lately a to sync folder with source code. Big mistake. Btsync has never really seemed to sync properly, and stops syncing frequently. Today, I needed to look at some source code on a fedora box and noticed that it hadn't synced for more than two weeks. So I restarted btsync, without it helping. BTsync still claimed everything was in sync, when it clearly wasn't. I disconnected and reconnected the share leaving the files in place. Big mistake. Every old file was copied from the stale folder to the remote fresh one overwriting two weeks of work. Luckily most of it was commited to a git repo, but still loads of work in progress was lost. BTW, you keep referring to version numbers, but neither the download nor the output when starting it mentions it. Anyway, I'm sure it is the latest, since the web UI didn't report an update was available. Also, do you really expect me to "turn on debugging" and recreate the problem? I mad enough as it is.