summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorGravatar http://cstork.org/ <Chris_Stork@web>2013-03-20 17:21:26 +0000
committerGravatar admin <admin@branchable.com>2013-03-20 17:21:26 +0000
commit0fff2934c97b5d9ae01520a8707624b9d0a12fde (patch)
tree8fa99feaa157804c295911e4bac0def47b52900a
parentb8e7c3b00cee7348ecf432b9a76d7edbae38ab98 (diff)
-rw-r--r--doc/forum/Direct_special_remotes.mdwn4
1 files changed, 3 insertions, 1 deletions
diff --git a/doc/forum/Direct_special_remotes.mdwn b/doc/forum/Direct_special_remotes.mdwn
index b2c93653e..bf80d57f8 100644
--- a/doc/forum/Direct_special_remotes.mdwn
+++ b/doc/forum/Direct_special_remotes.mdwn
@@ -2,7 +2,7 @@ I have a NAS at home which I access both as AFP/SMB shares and thru ssh/rsync.
Two **non-special remote setups** would be:
-1. Mount the share and turn it into a direct mode git annex repo. Does anybody have experience with this? I'd suspect this to be very inefficient due to the use of all the files in .git over AFP/SMB. Configuration as a WORM backend seems to be advised?
+1. Mount the share and turn it into a direct mode git annex repo. Does anybody have experience with this? I'd suspect this to be very inefficient due to the use of all the files in .git over AFP/SMB. Configuration as a WORM backend seems to be advised? (Edit: Well, I just tried this and 'git annex init' failed as described in this [forum post](http://git-annex.branchable.com/forum/Can__39__t_init_git_annex/). So AFP/SMB seem to be non-starters. :-( Furthermore, AFP/SMB are immediately detected as crippled and set to direct mode automatically.)
2. Same as 1. but with a local GIT_DIR. This should work by having .git on the NAS link to the local GIT_DIR.
Alternatively, I could treat my NAS as a [**web special remote**](http://git-annex.branchable.com/tips/using_the_web_as_a_special_remote/). Some URL schemes come to mind:
@@ -22,3 +22,5 @@ All of these indicate a mismatch between my use case and web special remotes.
Hence my question: **Would something like a "direct special remote" make sense?**
As a starting point I'd look at a setup similar to 2. above, i.e., a remote "working copy" with local GIT_DIR. Except that instead of a whole local .git directory a branch in an existing .git dir might be more appropriate...
+
+--Chris