I am new to using DVC and just exploring it. I am trying to pull data from s3 that was pushed by another person on my team. But I am getting this error:
WARNING: Some of the cache files do not exist neither locally nor on remote. Missing cache files:
name: head_test_file.csv, md5: 45db668193ba44228d61115b1d0304fe
WARNING: Cache '45db668193ba44228d61115b1d0304fe' not found. File 'head_test_file.csv' won't be created.
No changes.
ERROR: failed to pull data from the cloud - Checkout failed for following targets:
head_test_file.csv
Did you forget to fetch?
My mistake. I ran dvc add but I had missed to run dvc push. Running this fixed it.
When I ran dvc add it did create my_file.csv.dvc file but it was not pushed. Hence when I was trying to pull it saw the .dvc file but got confused.
It was a simple solution, but it took a while for me to figure out. Since this is a new tool, asking and answering my own question just in case if someone else makes the same mistake.
You may want to run dvc install which installs a Git hook to automate dvc push before git push (:
Push: While publishing changes to the Git remote with git push, its easy to forget that the dvc push command is necessary to upload new or updated data files and directories tracked by DVC to remote storage.
This hook automates dvc push.
Related
I want to finally start using version control for my R projects. I started following this guide, created an empty repo, but the problem started when I got to the command line part.
First I did:
git remote add origin https://github.com/username/repo_name.git
But after I type:
git pull origin master
I get this error:
fatal: Couldn't find remote ref master
Now, if I try the following:
git remote -v
I get:
origin https://github.com/username/repo_name.git (fetch)
origin https://github.com/username/repo_name.git (push)
So something is working. I am new to git and not sure about where the problem is.
As an attempt to fix the problem, I added one file for the repository (file.R) but apart from that, the repository is empty. I get the same error as before.
And if I try git branch -r on the terminal I don't get anything in return.
If you started by creating an empty repo in GitHub, then the default branch is no longer master. The default branch is now main.
If you started by creating a new repo locally using git init, then check what your default branch is using git branch. This should show you either a main or master branch.
It may be easier to create the repo in GitHub with a simple README.md and .gitignore, then clone that repo to another folder. Move your code into that repo, add and commit everything, then push your code to GitHub.
I am trying to set up a "simple" git workflow for a Wordpress installation that has a local version, a staging version and a production version. All I want to do is make changes locally and push from local to staging and from local to production. At first I thought this would be a simple task. I have been able to initialize the repository, add the two remotes "staging" and "production", initialize git on the remotes and push changes from my local version to the staging and production servers normally using the commands:
git add .
git commit -m "some changes"
git push staging master
git push production master
However, at some point during my work something changed, and while I am still able to push to Staging, now I am unable to push to the Production server without getting the error:
! [remote rejected] master -> master (Working directory has unstaged changes)
When I do "git status" it says:
On branch master
nothing to commit, working tree clean
After reading the answers to several SIMILAR BUT DIFFERENT questions on Stack Overflow I have tried the following:
git pull staging master
git pull staging master --rebase
git pull production master
git pull production master --rebase
I also tried executing this command on the remote servers
git config --local receive.denyCurrentBranch updateInstead
I have already completely re-created the servers and repositories a few times just to re-install git entirely from scratch, but this problem keeps happening after a while and at this point Git is actually HURTING my workflow instead of helping it. If anyone has any insight into what my mistake is, it would be much appreciated!
I had similar problems, pushing to a non-bare remote repo where I wanted the working copy files to be checked out immediately.
My remote was configured with receive.denyCurrentBranch updateInstead, but it still refused to accept pushes at unpredictable times.
Git 2.4 added a push-to-checkout hook, which can override such push failures.
I wrote a short script based on the example in the githooks documentation.
#!/bin/sh
set -ex
git read-tree --reset -u HEAD "$1"
I installed this script in .git/hooks/push-to-checkout in my remote repo.
Note that this script overwrites the working copy in the remote — it does not attempt to merge any changes to those files.
That's because I want the working copy to simply reflect the files in the repo.
Making a git bare repo is the best practice to push to.
You could push to a non-bare one... but only if you are not modifying files on the destination side while you are pushing file from the source side. See push-to-deploy.
But the best practice remains to add a post-receive hook (as in my other answer) in order to checkout in an actual folder all the files you have received.
So, I did a stupid rm -rf on the folder where the complete phabricator folder was present.
The whole phabricator database is still intact though.
I cloned the required repos on the same old location:
somewhere/ $ git clone https://github.com/phacility/libphutil.git
somewhere/ $ git clone https://github.com/phacility/arcanist.git
somewhere/ $ git clone https://github.com/phacility/phabricator.git
Apache was already configured during previous install.
I then ran:
./bin/storage upgrade
After which I went to the address which pointed to phabricator folder. Now I get the following error:
1146: Table 'phabricator_user.user_cache' doesn't exist
How do I resolve it? Or in general, what's the best way to reinstall phabricator using the old database?
Thanks
Well, if you still have the database, make a mysqldump from the data (export the db data - you should have this by default - a cron job, running a backup script on another backup machine/usb/hard/cloud)
Do a fresh reinstall on phabricator(EVEN on whole LAMP).
Import the previous backup.sql you did.
After setting the user/passwd/host/port/ in the "path_to_phab/conf/local/local.json" via the command line or simply editing the file, try to run the
./bin/storage upgrade
This should work fine if you have the storage engine set to mysql db (not-recommended). If you have a different storage engine (like hdd) try to restore data reproducing the path to where you have data in phab`s fresh installation conf files along with mysql import.
I've used Git successfully on this machine in the past but suddenly I can no longer push my commits to the Github repo. The last change to the Git toolchain that I made was to install Git 1.8.5.2, in addition to the Github for Windows client. RStudio could not find Git unless I'd already started the Github client so I decided to simply install a stand-alone Git client and change the RStudio Git path.
Error message (RStudio):
error: cannot spawn rpostback-askpass: No such file or directory
fatal: could not read Username for 'https://github.com': No such file or directory
Troubleshooting:
I can commit all projects.
I can pull new projects.
I cannot push any projects, I receive the same error message every time.
I cannot push with Github or RStudio.
Reinstalling /uninstalling Git / Github does not resolve the issue.
Setup:
This is an R project, with RStudio as my IDE / Git GUI.
I'm using Git 1.8.5.2 for Windows 7.
Let me know if there's any more information that you need.
Update 1:
Git GUI tells me that:
Error: hook execution requires sh (not in PATH).
Let's see if I can fix that...
Found something that might help from here: https://github.com/STAT545-UBC/Discussion/issues/93
in RStudio, click on the "Tools" menu and select "Shell"
Run the following command: git push -u origin master
it might ask you for your git username and password. Supply this information, make sure it is correct
hopefully the push is successful, then you can close the window
Now make some more edits to some file so that you have new content to push
click on the "push" button in RStudio and this time the push should work
Found a different suggested solution here: https://github.com/OHI-Science/ohicore/issues/104
git config --global credential.helper osxkeychain
How do I clone and run auth branch with my app, installed at [SRH/meteor/myapp], in addition to having master meteor installed in [usr/local/meteor]?
After having cloned the auth branch, when I attempt to add a package to my app (via meteor add accounts-ui), an error states the package is not available. Clearly, it is looking for the package at: [usr/local/meteor], while I am attempting to run meteor from [SRH/meteor].
Is it OK to have meteor installed at two locations? Obviously, I am confused.
More specifically, how do I clone the auth branch directly to [usr/local] location? Note: I am a newbie to git cloning. Took me a long time to find a tip on how to clone a branch, using:
git clone -b auth git://github.com/meteor/meteor.git
Thanks!
You can do one of two things.
Clone the auth branch to a location (as you have described) and then run meteor via
../../path/to/cloned/auth/branch/meteor run
Use meteorite which was basically created to do this for you.