If it's not too hard, we should just install it everywhere we install git, just for consistency. But at the very least it should be installed on all scap-related hosts (masters & targets)
|Resolved||awight||T187217 [Epic] Support word2vec for production ORES models|
|Resolved||awight||T188446 Package word2vec binaries|
|Resolved||Halfak||T171619 [Epic] ORES should use a git large file plugin for storing serialized binaries|
|Resolved||awight||T181678 Plan migration of ORES repos to git-lfs|
|Resolved||None||T176324 Scoring platform team FY18 Q2|
|Resolved||Halfak||T183198 Scoring Platform FY18 Q3|
|Resolved||awight||T176336 Deploy drafttopic model to production ORES|
|Resolved||mmodell||T180627 Support git-lfs in scap|
|Resolved||awight||T181010 [Spike] Write reports about why Ext:ORES is helping cause server 500s and write tasks to fix|
|Resolved||awight||T181071 Cache ORES virtualenv within versioned source|
|Resolved||fgiunchedi||T189306 SCAP: Upload debian package version 3.7.7-1|
|Resolved||awight||T180628 Install git-lfs client (at least on scap targets & masters)|
A direct backporting and rebuilding is impossible as it turns out. There just way too many dependencies that are present in buster and not in stretch. That being said, it's a go app so it compiles to not really dynamically linked executable (aside from libpthread and libc) so the same package provided for buster works on stretch as well (and I did a quick test of git lfs install and git lfs track to verify).
I 'll try and have it uploaded to stretch-wikimedia by the end of the day.
Why do the deployers issue git-lfs commands on the masters? Is it required per the process ? or just beneficial ? or nothing of the above ? Should we jot down the workflow (and even get it working in beta or some other environment) before we start trying it in production ?
@akosiaris: tl;dr I can't think of any reason that we must have git-lfs on masters, I've only got vague hand-wavy notions about it being convenient.
I don't know if the workflow requires it, I was really just imagining that it'd be convenient for debugging. Now that you mention it, I agree, it would be a good idea to figure out the workflow and get it written down somewhere. Trying it on beta is probably a good idea as well.
I've double-checked and it appears that we are only calling git-lfs/git-fat from one place in deploy-local, so git-lfs definitely isn't required (by scap) to be installed on masters.
Maybe @demon or @awight have a better idea than I do about the workflow. I'm only just getting familiar with git-lfs myself and I have never used git-fat (which is essentially the same workflow, afaik)..
I can't think of any compelling reason why it would be *required* on the masters...but I could see it as being useful if you want to actually have your large file contents on the master if you're debugging/curious about something...?