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 2019-03-19 13:56:46 -04:00
commit 56b1274d8c
No known key found for this signature in database
GPG key ID: DB12DB0FF05F8F38
2 changed files with 54 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

21
doc/bugs/git-annex_precommit_.__gets_stuck_after_likely_interrupted___96__annex_add__96___call/comment_2_f5e21a674c54c5dc4c8474ce1b4ffaf1._comment Normal file
View file

@ -0,0 +1,21 @@
[[!comment format=mdwn
username="yarikoptic"
avatar="http://cdn.libravatar.org/avatar/f11e9c84cb18d26a1748c33b48c924b4"
subject="reporting back"
date="2019-03-19T17:01:10Z"
content="""
Removing .lck, setting `annex.pidlock=true` worked! got just one complaint from git:
[[!format sh \"\"\"
$ git commit -m 'Committing staged changes after annex process was killed
>
> Also removed stale .git/annex/*.lck and now set annex.pidlock=True for the repo'
(recording state in git...)
git-annex: .git/annex/journal/e64_e4a_MD5E-s55076--a7352bf6fd37902fc1dfb6228f2e8a53.log: removeLink: does not exist (No such file or directory)
\"\"\"]]
I found that file (via `find * -lname *MD5E-s55076--a7352bf6fd37902fc1dfb6228f2e8a53*` FTR), it was not known to git yet, so I am now running `annex add` on it (takes awhile -- I guess repo is indeed got large + NFS) -- will report back when done/if there is more.
"""]]

33
doc/bugs/git-annex_precommit_.__gets_stuck_after_likely_interrupted___96__annex_add__96___call/comment_3_028eca36967a89e5f479802499232a90._comment Normal file
View file

@ -0,0 +1,33 @@
[[!comment format=mdwn
username="yarikoptic"
avatar="http://cdn.libravatar.org/avatar/f11e9c84cb18d26a1748c33b48c924b4"
subject="comment 3"
date="2019-03-19T17:41:12Z"
content="""
Managed to login to that node again, after hours (>3) of \"running\" `annex add` is still there with current process tree being
[[!format sh \"\"\"
PID TTY STAT TIME COMMAND
22168 pts/50 Ss 0:00 /bin/bash
15560 pts/50 S+ 0:00 \_ git annex add freesurfer/sub-000458/label/lh.BA4p_exvivo.thresh.label
15565 pts/50 Dl+ 0:18 \_ /ihome/crc/install/datalad/python3.7/bin/git-annex add freesurfer/sub-000458/label/lh.BA4p_exvivo.thresh.label
15591 pts/50 Z+ 0:00 \_ [git] <defunct>
15601 pts/50 S+ 0:00 \_ git --git-dir=.git --work-tree=. --literal-pathspecs check-attr -z --stdin annex.backend annex.numcopies annex.largefiles --
15625 pts/50 S+ 0:00 \_ git --git-dir=.git --work-tree=. --literal-pathspecs cat-file --batch
15626 pts/50 S+ 0:00 \_ git --git-dir=.git --work-tree=. --literal-pathspecs cat-file --batch-check=%(objectname) %(objecttype) %(objectsize)
15642 pts/50 Z+ 0:00 \_ [git] <defunct>
15650 pts/50 Z+ 0:00 \_ [git] <defunct>
15703 pts/50 S+ 0:02 \_ git --git-dir=.git --work-tree=. --literal-pathspecs hash-object -w --stdin-paths --no-filters
\"\"\"]]
and all the locks there, and I believe being generated by this `annex add` run:
[[!format sh \"\"\"
$ /bin/ls .git/annex/
index index.lck journal journal.lck misctmp objects pidlock precommit.lck sentinal sentinal.cache
\"\"\"]]
the process 15565 is slowly increasing its seconds (now 19)... is that a good sign (no lockup, just very slow IO) or not (it is locked up, and all the time corresponds just to the periodic querying of the lock)? if to strace -- which particular process? or should I interrupt and run the entire `annex add` under strace?
"""]]