comment

diff --git a/doc/bugs/__34__failed_to_send_content_to_remote__34__/comment_11_539c272450ceb56badde88702afb4e4d._comment b/doc/bugs/__34__failed_to_send_content_to_remote__34__/comment_11_539c272450ceb56badde88702afb4e4d._comment
new file mode 100644 (file)
index 0000000..9cac6e0
--- /dev/null
+++ b/doc/bugs/__34__failed_to_send_content_to_remote__34__/comment_11_539c272450ceb56badde88702afb4e4d._comment
@@ -0,0 +1,23 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 11"""
+ date="2021-07-22T17:08:56Z"
+ content="""
+@mih so the new version's output confirms my suspicion that git-annex
+sees a change in the stat information of the file (mtime, size, inode)
+from before it started the transfer to when it finished.
+
+I could imagine eg, that somehow the filesystem is not preserving stable
+inode numbers, and so the inode might appear different, without anything
+having actually changed.
+
+It also seems possible that git-annex might somehow fail to stat the file
+either before or afterwards. In either case that would result in the same
+message. Maybe there's some way that the file gets deleted at the same time
+it's being transferred. Maybe the stat call fails for some other reason.
+
+A strace of git-annex seems like the next step, if you can reproduce this
+somewhat reliably. I suggest a `strace -v`, which will display the full
+results of the stat() calls. The stats of the object file done before
+and after are the only part of the strace that we should need.
+"""]]

diff --git a/doc/bugs/__34__failed_to_send_content_to_remote__34__/comment_12_67567cf57d9b7d57fb5e10ff7a54a99f._comment b/doc/bugs/__34__failed_to_send_content_to_remote__34__/comment_12_67567cf57d9b7d57fb5e10ff7a54a99f._comment
new file mode 100644 (file)
index 0000000..fee8fb6
--- /dev/null
+++ b/doc/bugs/__34__failed_to_send_content_to_remote__34__/comment_12_67567cf57d9b7d57fb5e10ff7a54a99f._comment
@@ -0,0 +1,17 @@
+[[!comment format=mdwn
+ username="joey"
+ subject="""comment 12"""
+ date="2021-07-22T17:26:02Z"
+ content="""
+I don't think it's a case of the file being deleted file it's being
+transferred. It is possible to do that, just `git annex drop` while
+a transfer is in progress. But that transfer fails in a different way:
+
+         .git/annex/objects/Pg/8g/SHA256E-s1048576000--81a66cc6944beade4f98430ad1993bfe98a773fd0db76dd7818b2486b7c33399/SHA256E-s1048576000--81a66cc6944beade4f98430ad1993bfe98a773fd0db76dd7818b2486b7c33399: getFileStatus: does not exist (No such file or directory)
+       
+         .git/annex/objects/Pg/8g/SHA256E-s1048576000--81a66cc6944beade4f98430ad1993bfe98a773fd0db76dd7818b2486b7c33399/SHA256E-s1048576000--81a66cc6944beade4f98430ad1993bfe98a773fd0db76dd7818b2486b7c33399: openBinaryFile: does not exist (No such file or directory)
+       
+         failed to send content to remote
+       
+         content not available
+"""]]