summaryrefslogtreecommitdiff
path: root/doc
diff options
context:
space:
mode:
authorGravatar Joey Hess <joey@kitenet.net>2012-02-14 16:28:16 -0400
committerGravatar Joey Hess <joey@kitenet.net>2012-02-14 16:28:16 -0400
commite76988f6c24277770fe1b7143e245cadab1e0ba7 (patch)
tree1f30b0c0fcbf198468718bbdfacf8433297111e1 /doc
parent03c559f8d63f5f4e9f3472d7d2de7342a21c0a90 (diff)
add
Diffstat (limited to 'doc')
-rw-r--r--doc/todo/redundancy_stats_in_status.mdwn23
1 files changed, 23 insertions, 0 deletions
diff --git a/doc/todo/redundancy_stats_in_status.mdwn b/doc/todo/redundancy_stats_in_status.mdwn
new file mode 100644
index 000000000..56095fd33
--- /dev/null
+++ b/doc/todo/redundancy_stats_in_status.mdwn
@@ -0,0 +1,23 @@
+Currently, `git annex status` only shows the size of 1 copy of each file.
+If numcopies is being used for redundancy, much more disk can actually be
+in use than status shows.
+
+One idea:
+
+ known annex size: 2 terabytes (plus 4 terabytes of redundant copies)
+
+But, to get that number, it would have to walk every location log,
+counting how many copies currently exist of each file. That would make
+status a lot slower than it is.
+
+One option is to just put it at the end of the status:
+
+ redundancy: 300% (4 terabytes of copies)
+
+And ctrl-c if it's taking too long.
+
+Hmm, fsck looks at that same info. Maybe it could cache the redundancy
+level it discovers? Since fsck can be run incrementally, it would be tricky
+to get an overall number. And the number would tend to be stale, but
+then again it might also be nice if status shows how long ago the last fsck
+was.