• Content Count

  • Joined

  • Last visited

About iaw4

  • Rank
  1. I am very pleased with resilio-sync. there are some not-so-intuitive aspects in adding peers and troubleshooting that could be made clearer with better messages, but overall it is awesome. I use it on 2 Macos computers and 4 linux computers, with different folders shared. I am usually the only regular user on these computers and I have sudo access on them. all my resilio-ed directories are under my main user directories, something like /home/me/... or /Users/me/... . my uid is typically 501 under macos and 1001 under linux, because these are the defaults for the first created user. not always. what I want can best be described as "easy uid and permissions" setup. that is, the resilio-ed folders mimick the 'me' ownership and permission files for sure. files that belong to someone else (like 'root') should be synced, too, to the greatest degree possible, although ideally they would inherit setuid. the sync-ed results should retain the same time stamps, ownership (root or me; but not the numerical id), and the protection settings. I am ok with the resilio daemon being su (and in fact prefer this, because there may be files that are owned by root that I have dropped into my sync-ing folders, and which I would prefer to be synced, too); but the rslsync daemon should chown, chmod, and touch the sync-ed files appropriately to keep them, well, in sync. right now, I have been using the "global resilio" configuration, which means that rslsync is often the owner of synced files. most of the time, this works; and when it does not, I just chown and chmod the directory tree a few times until it works. messy, not pretty. for use cases such as mine---which I presume are fairly common---what is the best-practice recommended setup? advice appreciated. /iaw
  2. weirdly, nothing. and weirdly, it started to work again, again nothing. this is why a step-by-step guide to checking where it is stuck (or a shell script) would be great.
  3. ...but now I have a connection again, though from another peer. this is very confusing. the network was up all the time, so this was not it. A troubleshooting procedure would be a great help.
  4. a resilio sync peer that used to work. now it no longer does. it has a correct key. but the peers think it is offline now (they know it exists.) this is a linux machine. I have had sync failures in the past because of permissions, but because of the offline message, I do not think this is the case. looking at the rslsync , the last messages look like this: [20170518 19:31:48.889] Relay: Connection to the relay[] has been lost with error: 104 (Connection reset by peer) [20170518 19:31:48.914] D! 10SyncTcpReq[0x00007f26344b90a0]: EOF - error: 103 (Connection to the relay server has been lost) [20170518 19:31:48.915] PD[A568] [E94F] (t:0 a:2): failed to open tunnel[0x00007f26344b90a0] to [::]:0/RELAY - error: 103 (Connection to the relay server has been lost), enc: SRP, tunnels: 0 now, the linux machine also sits together behind a firewall with one of the peers, and this peer is free from the firewall. I am beginning to be embarrassed about sending megabyte long log files to the poor people at support... is there a step-by-step guide to troubleshooting on linux ? sort of like "first ping here, then telnet there, you should get this response, to check that your write permissions to the destination work, do this" etc. /iaw
  5. rather than guess whether I have everything set correctly, it would be nice if I could write $ resilio-check-permissions ~/myfolder/ on a local computer, and resilio-check would contact the daemon to try to write a file, check for its existence, and then delete it again, telling me whether permission problems prevented functioning. /iaw
  6. on examination, the Archive is fairly small. most of the space is consumed by .!sync files. what are they, and can they be removed or trimmed? Update: Deleting them and restarting resilio seems to just rebuild them. I am now back over 60GB already...
  7. I have the following odd situation: # du -h -d1 bitsync 160G bitsync/.sync ... various 194G bitsync can I remove the .sync directory? or the archive subdirectory in .sync? if I remove it on one machine, will it remove on the others? (I think I need to figure out a way to store the history of files elsewhere. I was also looking at a GUI way to turn on/off archiving; is it in advanced?) is there some general advice on how to handle archiving (e.g., and trim/remove/move files after x days automatically)? /iaw
  8. not yet. :-). good to know. alas, the gui should be a little more userfriendly in this respect and filter what is of relevance a bit better. it could show the names of the connected hosts, rather than just a number. it could display the last error if something failed (e.g., permissions, out of space, etc.). it could display the last few files synched if one double clicks on the directory. wading through admin sync files is not great. otoh, given that a lot of info is already there, enhancing the gui should not be hard.
  9. for me, the best aspect about resilio sync is that it is the *only* filesync solution that does not mess up symbolic links. for me, the worst aspect is that the warning and error messages are not verbose enough. when two folders do not sync, please tell us why. is it a permission issue? is it a directory existence issue? is it something else that is stranger? when it says "not enough free space on the drive," add how much there is and how much it needs, and on what host the problem lies. can I see on all hosts which of my hosts has the problem? (could be in a hidden .resilio-warning file?) when it works, it is fantastic.
  10. there is an old rsync confusion, where adding a "/" on the directory could mean totally different things. for rsync, one does a dry-run (-n) to make sure one has it right. a small confusion can easily end up with gigabytes of duplicates. for resilio sync, this is more difficult. suggestion: on an initial synch, I would suggest three things: IF the destination folder is non-empty, collect info on how much needs to be actually transferred. after the reconciliation calculation, confirm with the user whether (s)he really wants to transfer move 100,000 files in 1 TB. similar to 1, give information about how many files would be added/duplicated/removed/etc. perhaps give a few sample filenames where they would be before and after. pick a few unusual and unique filenames from the two directories, see if there is a similar file on the other volume, and if there is, most likely what the user wants to do is for these two copies to become one copy. something else is unlikely. (anything else that would result in two copies should request at least one confirmation.) I know this is extra code, but it's a special and important single use case. because all that changes is an extra warning panel (if a preference is checked), and the underlying behavior of resilio is not changed, this should be doable. regards, /iaw