summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorGravatar https://me.yahoo.com/a/FHnTlSBo1eCGJRwueeKeB6.RCaPbGMPr5jxx8A--#ce0d8 <Hamza@web>2013-11-20 18:17:39 +0000
committerGravatar admin <admin@branchable.com>2013-11-20 18:17:39 +0000
commitb616e0b8893194af34d92fd87f0d6aeaac2be08b (patch)
tree62c19b4a703f5345784e4f8bfcb0ad52f2c43e36
parent8dbffd39ffc30879bd1d5068114c18e341a8e8ae (diff)
-rw-r--r--doc/forum/s3_vs_ssh_Performance_Problems.mdwn8
1 files changed, 8 insertions, 0 deletions
diff --git a/doc/forum/s3_vs_ssh_Performance_Problems.mdwn b/doc/forum/s3_vs_ssh_Performance_Problems.mdwn
new file mode 100644
index 000000000..5176e196f
--- /dev/null
+++ b/doc/forum/s3_vs_ssh_Performance_Problems.mdwn
@@ -0,0 +1,8 @@
+I backup/sync my music using annex. I used to have one repo with 3 clones, one full copy on my vps, one full copy and one partial copy on my laptops. I decided to move all data to s3. Declared my vps repo dead, purged history (I do not care about history for this particular repo) pushed the git repo to a different computer (bare repo no data) and data to s3 (gpg encrypted). I've just finished uploading all files (around 200gb) couple files failed during the initial upload so I did a `git annex copy --quiet --to mys3 --fast` this command used to take 15 20 secs on my laptop when sending data to vps using ssh but now it took 2 hours to complete (pushing mem usage to 90%).
+
+I have one other repo (1.5gb in size 42k files using indirect mode, data gpg encrypted on s3) using the same setup except this repos content has always been on s3 i had the same behavior on this repo too. adding a file and running a copy to push content to s3 took couple hours even if I add a single 1kb file. I used to blame my hard drive, damn thing is slow but now I think this is related to s3. is there any workaround for this?
+
+Both machines are using,
+
+ git-annex version: 4.20130913-gd20a4f2
+ build flags: Assistant Webapp Pairing Testsuite S3 WebDAV Inotify DBus XMPP Feeds Quvi