This commit is contained in:
parent
ac90d4f570
commit
6c6a5dff0d
1 changed files with 29 additions and 0 deletions
|
@ -0,0 +1,29 @@
|
||||||
|
Hello everyone,
|
||||||
|
I need some help with this use-case: two local computers, paired with git annex, also sync in the cloud (box.com) to pull and push changes even when the other local computer is offline.
|
||||||
|
|
||||||
|
I've tried the following steps with the webapp, but the result isn't working so far:
|
||||||
|
|
||||||
|
1. Create repo in *Computer A* through the webapp.
|
||||||
|
2. Create repo in *Computer B* through the webapp.
|
||||||
|
3. In *Computer A*, add box.com repository and put it in the 'backup' group (Just to make sure that it gets everything but I've also tried re-doing the whole procedure with box.com set to 'transfer' (as suggested by the webapp), and still will not work).
|
||||||
|
4. Pair computer *A* and *B* over the local network.
|
||||||
|
5. Now also enable box.com in *B* (the cloud repo has appeared thanks to pairing)
|
||||||
|
|
||||||
|
So far, when everyone is online, everything syncs correctly. Creating "file1" in *A*, will sync it to all repositories as shown by "git annex whereis file1".
|
||||||
|
|
||||||
|
The problem is that this won't work for the following use-case, which I'm currently trying to solve:
|
||||||
|
|
||||||
|
(continued from the previous steps)
|
||||||
|
|
||||||
|
6. *B* goes offline.
|
||||||
|
7. *A* creates a file (or makes a change), which is correctly synced to box.com (*whereis* shows that it's there too). As expected, syncing with B fails because B is offline.
|
||||||
|
8. *A* goes offline.
|
||||||
|
9. *B* come back online.
|
||||||
|
|
||||||
|
The expected behavior here is that B gets the new file (or change) from box.com, since it's been pushed there by *A*, but this doesn't happen. Strangely, the new file doesn't get in *B* until *A* comes back online.
|
||||||
|
|
||||||
|
**What am I missing? Should I do something differently to get this behavior?**
|
||||||
|
|
||||||
|
(This use-case is important for me because I need to keep some binary files in sync. These files are frequently modified by either *B* or *A*, and very often one of the two repos is not online. The different versions can't be simply merged back because of the binary nature of the files.)
|
||||||
|
|
||||||
|
Thanks everyone for your help.
|
Loading…
Add table
Add a link
Reference in a new issue