From 661f452dfa518b509f31d5bee502e2f6afb7ec04 Mon Sep 17 00:00:00 2001 From: jgoerzen Date: Mon, 5 Sep 2022 01:01:59 +0000 Subject: [PATCH] Added a comment --- .../comment_12_e1f3b19d1dd62b1c9314e1e826fec54f._comment | 8 ++++++++ 1 file changed, 8 insertions(+) create mode 100644 doc/tips/Repositories_with_large_number_of_files/comment_12_e1f3b19d1dd62b1c9314e1e826fec54f._comment diff --git a/doc/tips/Repositories_with_large_number_of_files/comment_12_e1f3b19d1dd62b1c9314e1e826fec54f._comment b/doc/tips/Repositories_with_large_number_of_files/comment_12_e1f3b19d1dd62b1c9314e1e826fec54f._comment new file mode 100644 index 0000000000..8a85021b7f --- /dev/null +++ b/doc/tips/Repositories_with_large_number_of_files/comment_12_e1f3b19d1dd62b1c9314e1e826fec54f._comment @@ -0,0 +1,8 @@ +[[!comment format=mdwn + username="jgoerzen" + avatar="http://cdn.libravatar.org/avatar/090740822c9dcdb39ffe506b890981b4" + subject="comment 12" + date="2022-09-05T01:01:59Z" + content=""" +To expand: this took about 2 hours to run. A git annex sync from another git annex repo is a lot faster (a minute or so). It is the directory remote that's so slow. Examining with strace and lsof, I don't believe this is chechsumming. In fact, it seems to be mostly continuous reading from cidsdb/db, which is only 49MB in size and therefore certainly cached. The process is entirely CPU-bound. +"""]] -- 2.39.5