I created a git repo in Windows 7 on a NTFS partition and when opening it in Linux (Ubuntu 12 x64, dual-boot setup) I get the index file open failed error. How can I figure out what's wrong? The partition is mounted read-write and I've never had any other problems. Does git store data in a different format Windows vs. Linux and I need to do either a clone or some conversion? I'd really like to be able to work on the same repo in both OSs without cloning around...
Clarification: I also get cat: index: Input/output error
when running the command cat index in the .git dir, so it is a NTFS related problem... but I've never had it before untill using git in a cross-systems way and I've run other apps from NTFS parts and copied files around...
The .git/index file is a binary file, which describes the current workdir. Perhaps a git fsck is able to fix it up (move the one you have out of the way to make sure it isn't lost while you fool around, or make any expertiments on a copy of the repository). You might try to clone the repository locally, the clone might get a good copy of the file, which you could then copy over the broken one.
Possibly permission problems? Backup what is relevant, defragment the drive, run hardware checks (it might be a broken/breaking disk!).
Either your Linux NTFS driver is broken, or you have filesystem corruption, or both. Reboot to Windows and run the disk checking utility, then see how things stand when it finishes.
Related
I recently updated a chroot on an old Chromebook from Ubuntu bionic to focal. The chroot has encryption enabled.
I usually work with Git repositories and other files within the Chrome's Downloads folder and haven't had any issues with this previously.
Since the update though, I found I was unable to run things like git clone -- I get an error saying cannot create worktree dir: no such file found. I looked around and found people had similar problems but there's been no clear solution.
Then I decided to look inside one of the existing folders within Downloads and noticed a problem there...
I can open a repo within my Downloads folder on ChromeOS and see all files as I used to.
I can enter-chroot and run ls on the same folder and see all files as I used to there too.
But when I launch the chroot/crouton (I used xfce4), and try to ls the folder from within the terminal, or even look at the folder contents from a UI window, the contents of the repo look encrypted -- as in all the filenames have changed to strings of equal-length and apparently random characters.
It's almost as if encryption is working in reverse -- so my files are unencrypted outside the crouton, but as soon as I go into the xfce UI, they're encrypted and there's no decryption happening. But that's speculation on my part...
Any ideas as to what is going on here? And how I can continue to work within crouton?
It seems this is to do with the fact that Chrome OS encrypts files and that something had happened since I updated Crouton (rather than my updating Ubuntu from Bionic to Focal).
I realised this was a bigger issue when even command line tools like tar and git (which I'd installed) weren't working.
When I tried to unpack a download of Firefox with tar xjf I got an error saying "Required key not available". Some searching around that led me to issue #3261 on the Crouton Github repo.
The solution for me was:
Ensure /etc/pam.d/su-l was writable. (I did ls -l /etc/pam.d/su-l to check but ultimately used sudo...)
Edit the file /etc/pam.d/su-l. (I used sudo vi /etc/pam.d/su-l to ensure the file wasn't read-only in that instant, and because I had no other text editor options but vi available.)
Comment out the line session optional pam_keyinit.so force revoke. (So it should read # session optional pam_keyinit.so force revoke.
Save the file.
Restart the chroot.
I'm opening a new worktree in my local repository in Windows, and can't compile it in my Linux virtual machine, because the ".git" contains a full path starting with "C:/Git" that the virtual machine does not recognize.
I have a git repository that I often need to cross-compile on both Windows and Linux.
I originally cloned the repository in my windows system and mounted my "C:" drive as a shared file system in a virtual Ubuntu machine.
This usually works perfectly fine.
When attempting to open a new worktree in git, the linked repository doesn't have it's own ".git" folder, but rather a link to the original repository's ".git" folder, and this is saved as an absolute path, in the following format:
gitdir: C:/Git/...
When attempting to run:
git rev-parse
I get the following message:
fatal: not a git repository: /mnt/hgfs/WindowsDriveC/Git/WorktreeDir/C:/Git/OriginalGitDir/.git/worktrees/WorktreeDir
Due to this limitation, several of my scripts fail.
Is there any way to "trick" Linux into recognizing the full path and properly identifying the original git repo?
Typically, doing this kind of thing is discouraged because it can lead to corruption, and as you've seen the Windows paths don't play well with the Linux paths.
However, that being said, there are a couple of options. One is to set up the worktree under Cygwin or Windows Subsystem for Linux and share it. That will still result in a path that doesn't exist on the Linux VM, but it would be possible to create a symlink from /cygdrive/c or /mnt/c to /mnt/hgfs/WindowsDriveC and then the rest of the stuff would work.
Another option is just to use Windows Subsystem for Linux, which means that you can do your compilation and produce Windows binaries without leaving Windows. You'd produce Linux binaries, but you wouldn't have to worry about sharing with the VM. The worktree would need to be created under WSL, though.
A final option is to edit the .git file in the worktree to use a relative path to the main repository. This is unsupported and might break, but then again it might work for your needs. Be sure to use forward slashes for that purpose.
I have a dynamic linker which is /lib64/libc.so.6
I stupidly renamed it to /lib64/libc.so.6.old and now NO commands work.
I cannot do ls or mv to rename it back.
I can run ldconfig but it says permission denied and I cannot run sudo or su - What on earth can I do to fix this? I am running Oracle Linux redhat 6.7
LD_PRELOAD=/lib64/libc.so.6.old mv /lib64/libc.so.6.old /lib64/libc.so.6
Start from a recovery/install iso and rename the file back.
If you can't reboot or don't have physical access to the machine you could try to install a compiled version of BusyBox https://busybox.net/FAQ.html#getting_started and use its su and mv commands. Since BusyBox is statically linked it should work without libc.so.
Go to single user mode, mount the file system with rw, since you know the location of the renamed file move /lib64/libc.so.6.old /lib64/libc.so.6
I would also propose a workaround with a mount point as already mentioned by #wildplasser.
You can make majority of command line tools working again if you have a mounted directory on your broken host. If you are lucky to have one then all you need is to upload the libc-x.yz.so (which you can take from another host of from Internet) on the share, rename it there to libc.so.6 and add the mounted directory to LD_LIBRARY_PATH.
If the version x.yz is the same as for the one which you thoughtlessly moved then the commands like ls, cp etc. will work again in the console where you set LD_LIBRARY_PATH. You should not logout from this console, because you won't be able to login again.
! Be aware that setuid command line tools won't work (see https://askubuntu.com/a/1029363/832810). Unfortunately "sudo" is one of them, this is why you won't be able to put back easily your long-suffering .so (unless you have a root# console). However it gives you a possibility to save all data and finish all actions before you do some hard restore.
If you managed to do the above-mentioned trick and you have enough time you can try to build a statically linked version of "sudo" as suggested on https://askubuntu.com/a/1030475/832810 (probably even build on another host and copy through NFS) and move the .so back using it.
I have a Windows PC and on the same PC I have a Lubuntu VM inside the VMWare player.
I share my Windows folders so I can see them from the Lubunutu VM.
My problem is when I clone a Git repo on Windows (using Tortoise Git) and then try to access it from Lubuntu (using Git from command line) all files appear modified although I have not changed them.
I know that reason for this is that Windows and Linux handle the new lines differently.
My question is how can I configure my Git installations on Windows and on Lubuntu so I do not have this problem?
Update:
As suggested (by Craig Estey) this does not seem to be CRLF problem. I tried cloning a repo in my Linux VM on the shared directory and got following error:
fatal: 'origin' does not appear to be a git repository
fatal: Could not read from remote repository.
Please make sure you have the correct access rights
and the repository exists.`
I have no problem cloning the same repo on the VM's local drive.
Any idea what could be the reason?
Note: The main reason I want to do this is to keep my VM's drive size small and use shared drive as much as possible. My VM's local drive is on SSD drive and shared drive is on much larger HDD.
Suppose you put a repository into a Windows Shared folder.
And mount(Mount shared folder on Ubuntu) that shared folder on Ubuntu, so you can clone that repository. (Talking about you can not clone, it must be another problem. I tested it, good for me.)
When you cd to that repository in that mount, you see all files are modified on Ubuntu, then it should be the EOL problem.
On Windows the autocrlf is true by default(assume you are using Git for Windows).
When you clone a repository on Windows, the files will be checkout with CRLF EOL.
But, On Ubuntu the autocrlf is false by default. Using git in that repository which is mount on Ubuntu, git expects the EOL is LF. But, the EOL is still CRLF. That's why git treats all file as modified on Ubuntu.
To fix this problem, make sure both OS to use the same autocrlf value, or using .gitattributes to control the EOL.
For example, using false value:
On windows,
Delete all files in working tree of that repository.
(Note: if you have local changes, commit them first.)
Run git config core.autocrlf false for that repository only. Or
Right click in that repository, click TortoiseGit -> Setting
In Settings dialog, go Git node, select local and un-check the AutoCrlf
Apply the change
Perform git reset hard to get all files back with correct EOLs.
Open TortoiseGit Log Message dialog
Right click on current branch and perform Reset "<current branch name>" to this
Choose "Hard" option
On Ubuntu, you should not see all files modified.
For using .gitattributes:
* -crlf
I can do checkouts, update and such. But when I try to commit changes, SVN gives me the following error:
Can't open file '/svn/p/pokemonium/code/db/txn-current-lock': Permission denied
I am using Windows 7 x64 SP1 with latest version of TortoiseSVN.
UAC is off, my account has read and write access, etc.
I can commit fine to other svn repositorys.
For me it ended up being a permissions issue on the server. I have my repo on a linux box, and ssh in to use svnadmin. For convenience sake, I had executed my create repository command as root. I was looking to get source I had on my Windows box into the repo, so was using TortoiseSVN to set up trunk/branches/tags. The directory containing the repo on the server was owned by root, and Tortoise was coming in as apache. I chowned the directory on the server to apache:apache, and it all went smoothly.
chown apache:apache -R my_repo_root
This is a server configuration issue. On windows host Visual SVN server runs under NETWORK SERVICE account by default. I solved this problem by granting full access rights to the repository folder to this account. Another option is switching Visual SVN service to the SYSTEM account, but that could pose a potential security risk.
Try this.
Make a back up copy of your working copy (just to be safe).
Make another copy your entire working copy off somewhere else.
Take the copy and delete all of the SVN folders out of it
Delete your working copy and do a fresh checkout
now copy/paste your corrupted working copy over your fresh checkout.
it is critical for this to work that you have completely removed ever _svn or .svn folder from your corrupted working copy before you perform the copy/paste.
This will leave you (hopefully) with a working copy that shows (!) on all the files you had modified since your last commit. And fixes your lock issue.
I had the same problem after I re-installed Windows 7 and just copied the SVN Repository from the old Windows to the new one.
After trying the steps that Mr. Manager proposed, the problem was still not fixed in my case.
After making sure that the permissions was setup correctly for the SVN Repository folder I just deleted the file 'txn-current-lock' in the /db folder of the project. That fixed it for me. From thereon I could commit my project again.
I had faced same issue on Unix box
Restarting the Apache service of the SVN server solved myproblem.
-f httpd.conf -k stop
-f httpd.conf -k start
In my own case, my linux server had been restarted after a power loss. The file system remained mounted as read-only since some journal repairs had been made. Rebooting the machine restored full function.
permissions worked for me too
error
repo/db/txn-current-lock: Permission denied
fix
chown apache:apache -R my_repo_root