diff options
| -rw-r--r-- | doc/todo/speed_up_transfers_over_ssh+rsync_--_directly_reuse_the_same_connection__63__/comment_9_88dbe6602f273b7600776228ec099955._comment | 14 |
1 files changed, 14 insertions, 0 deletions
diff --git a/doc/todo/speed_up_transfers_over_ssh+rsync_--_directly_reuse_the_same_connection__63__/comment_9_88dbe6602f273b7600776228ec099955._comment b/doc/todo/speed_up_transfers_over_ssh+rsync_--_directly_reuse_the_same_connection__63__/comment_9_88dbe6602f273b7600776228ec099955._comment new file mode 100644 index 000000000..ae4e77279 --- /dev/null +++ b/doc/todo/speed_up_transfers_over_ssh+rsync_--_directly_reuse_the_same_connection__63__/comment_9_88dbe6602f273b7600776228ec099955._comment @@ -0,0 +1,14 @@ +[[!comment format=mdwn + username="joey" + subject="""comment 9""" + date="2018-03-06T17:29:16Z" + content=""" +I think that any further improvements here would need to be a fundamental +change, rather than shaving away fractions of a second from the above +benchmarks. Most of the remaining overhead seems to be outside of +git-annex's own code. + +I think it could use sftp, rather than rsync, when git-annex does not +need to resume transfer of a file. git-annex could keep sftp running, and +ask it for each file in turn, and avoid all the connection overheads. +"""]] |