Closed yarikoptic closed 2 years ago
@yarikoptic
attn @jwodder -- should be shown
There's an INFO-level log message "Moving assets for Dandiset %s to backup remote" that should be shown before the git-annex move
output. Is it not included in the e-mail?
I did fsck on one of those files and everything seems to be ok, move
also didn't move anything, odd
(dandisets) dandi@drogon:/mnt/backup/dandi/dandisets/000022$ git annex fsck --from dandi-dandisets-dropbox --key SHA256E-s1900372495--44f06eea609701c7f7d5bba7ff3483c756883a3e10b2e1e64b78f67ceed521ec.nwb
fsck SHA256E-s1900372495--44f06eea609701c7f7d5bba7ff3483c756883a3e10b2e1e64b78f67ceed521ec.nwb Total objects: 1 Total size: 953.674 MBytes (1000000000 Bytes)
Total objects: 1 Total size: 858.662 MBytes (900372495 Bytes)
(checksum...) ok
(recording state in git...)
(dandisets) dandi@drogon:/mnt/backup/dandi/dandisets/000022$ git annex move -J6 --all --to dandi-dandisets-dropbox --not --in dandi-dandisets-dropbox --and --in here
(dandisets) dandi@drogon:/mnt/backup/dandi/dandisets/000022$
There's an INFO-level log message "Moving assets for Dandiset %s to backup remote" that should be shown before the
git-annex move
output. Is it not included in the e-mail?
nope
@yarikoptic Looks like PR #133 wasn't pulled in somehow.
May be but I think it was pulled in, there is DEBUG entry for the populate dandi-dandisets-dropbox
invocation. I bounced you an email from yesterday... since then there were no new email,
without any obvious network activity, so I guess it is just stuck. git annex is 8.20211028-g1c76278
, and IIRC there were some fixes for parallel git annex get operation since then, but I am a bit reluctant to upgrade since most recent versions have their own issues... hm FWIW, running that annex find
in 000026 gives only a single file
(dandisets) dandi@drogon:/mnt/backup/dandi/dandisets/000026$ git annex find --not --in dandi-dandisets-dropbox --and --not --in here
derivatives/EPIC/sub-EXC022/sub-EXC022_ses-MRI_flip-4_VFA.nii.gz
so may be indeed somehow a race condition across this many get
jobs, eh. I will interrupt that backup run, do manual get, and reincarnate backup run manually to see where we would get
havent' seen anything like that recently so let's blissfully be happy and hope in our ignorance that the issue is not of a problem for us.
may be ran out of space??? from today's email:
so some did transfer ok but not the others ... who knows what dataset that is (attn @jwodder -- should be shown)... figured 000022.
similar message happened in other logs today as well...