mirrors/git-annex
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages 1 Wiki Activity Actions

Merge branch 'master' of ssh://git-annex.branchable.com

Browse source
This commit is contained in:
Joey Hess 2014-04-07 18:31:16 -04:00
commit a1817e9760
4 changed files with 62 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

23
doc/bugs/Assistant_having_a_child_git_cat-file_--batch_do_the_same_thing_over_and_over_and_using_a_lot_of_memory/comment_3_1429ca784a03bc424b3537cbe0449421._comment Normal file
View file

@ -0,0 +1,23 @@
[[!comment format=mdwn
username="http://joeyh.name/"
ip="209.250.56.244"
subject="comment 3"
date="2014-04-07T21:55:19Z"
content="""
Unfortunately all I have been able to tell for sure from this log is that it seems that the expensive transfer scan is not running, and this is unlikely to be a repository auto-repair.
My best guess as to what might be going on is an update of the git-annex branch.
[2014-04-07 23:24:24 EEST] read: git [\"--git-dir=/storage/sarjat/annex-sarjat/.git\",\"--work-tree=/storage/sarjat/annex-sarjat\",\"write-tree\"]
This is prep for an index file commit, probably to the git-annex branch.
[2014-04-07 23:24:24 EEST] read: git [\"--git-dir=/storage/sarjat/annex-sarjat/.git\",\"--work-tree=/storage/sarjat/annex-sarjat\",\"rev-parse\",\"84068090af4bcd3d24f16d865ac07b0478f20ada:\"]
This is a getting the parent commit's tree.
The git-cat-file churn could then be a union merge reading the contents of the git-annex branch to union-merge it into the `.git/annex/index` (in `mergeIndex`). This would reuse the main git cat-file process.
That does not explain why it would need to read eg, SHA256E-s106800355--c70e31d511e7eec4881a15dfba521ea3d1fe14694968f81ae1819f1a2a93f9be.mp4.log 28 times.
Normally, during a union merge only files listed by `diff-index` need to be read, and it lists each file only once.
"""]]

12
doc/bugs/Assistant_having_a_child_git_cat-file_--batch_do_the_same_thing_over_and_over_and_using_a_lot_of_memory/comment_4_f9e65cf5598b4b14eeee1f41f46d4084._comment Normal file
View file

@ -0,0 +1,12 @@
[[!comment format=mdwn
username="http://joeyh.name/"
ip="209.250.56.244"
subject="comment 4"
date="2014-04-07T21:57:32Z"
content="""
Does `git log git-annex` show a commit that was made at 23:30?
Does it show a commit 84068090af4bcd3d24f16d865ac07b0478f20ada?
Is 84068090af4bcd3d24f16d865ac07b0478f20ada the parent of the 23:30 commit?
"""]]

19
doc/bugs/Assistant_having_a_child_git_cat-file_--batch_do_the_same_thing_over_and_over_and_using_a_lot_of_memory/comment_5_044ecac2d2e670e1ef69809c944093d1._comment Normal file
View file

@ -0,0 +1,19 @@
[[!comment format=mdwn
username="http://joeyh.name/"
ip="209.250.56.244"
subject="comment 5"
date="2014-04-07T22:12:14Z"
content="""
Is the repository using direct mode?
Another theory is that:
* test/hello appears
* watcher sees new symlink, tries to make a commit with it
* master branch already has that symlink
* this is why the write-tree is not followed by a commit-tree. The commit would have been empty.
If this is the case, then 84068090af4bcd3d24f16d865ac07b0478f20ada will be a ref on the master branch.
And all of the above is normal operation. But it does suggest, that if this repo is in direct mode, it might be running a direct mode work tree update around then. Which requires a lot of cat-file queries of the git-annex branch. And would certainly make repeated queries at least if the repository has duplicate copies of some files..
"""]]

8
doc/bugs/Assistant_having_a_child_git_cat-file_--batch_do_the_same_thing_over_and_over_and_using_a_lot_of_memory/comment_6_6f4f51e1583bed5e7e601e4f30f4207b._comment Normal file
View file

@ -0,0 +1,8 @@
[[!comment format=mdwn
username="http://joeyh.name/"
ip="209.250.56.244"
subject="comment 6"
date="2014-04-07T22:17:16Z"
content="""
Does the git log have any recent commits that were \"git-annex automatic merge conflict fix\" ?
"""]]