Git Deploy - "file write error (No space left on device)"
Git was trying to unpack some objects, but was not able to do so. As the error suggests this might be because of a just "no space left on device", which might also mean "no inodes left on device".
Please check with df -h
and df -hi
. (Assuming gnu-coreutils are present, as in any good linux system.)
You did not say what command you ran and where you found that error. The command might interact with another server, hence it might not be obvious what "device" is talked about.
The error message looks like copied from a log file, in this case it is likely the "device" is on the same host as the log file.
Related videos on Youtube
yexela
Updated on June 04, 2022Comments
-
yexela almost 2 years
[Wed, 01 Dec 2010 12:17:07 -0800] DEBUG: STDERR: error: file write error (No space left on device) fatal: unable to write sha1 file fatal: unpack-objects failed
I have seen a couple of discussions on this but haven't found a fix. I have two apps on my instance and never had this problem before. My recent commit was not very large either (database migration and controller addition). Any advice on what to do here? The volumes on my instance are not out-of-space.
-
owl over 13 yearsActually by the looks of things, they are out of space. What's the git command that's running when this happens?
-
KingCrunch over 13 yearsMaybe no write permission (anymore)?
-
Alexander Aleksandrovič Klimov over 13 yearstmp file system out of space?
-
cdhowie over 13 yearsOne of two things has happened: you have actually run out of disk space, or your repository has grown over 4GB and you are trying to compress it into one pack on a FAT32 filesystem. The former scenario is much more likely...
-
Ibolit about 13 yearsHi. I have the same problem, and i have quite a lot of free space, and the file system on my server is UFS2 (it is a FreeBSD server). The permissions seem to be alright, and it all used to work for quite a while...
-
twalberg over 11 yearsYou can also get "No space left on device" if you're using a file system that creates a fixed number of inodes (or similar constructs) at file system creation time, and those object are exhausted, even though there's still free space left (
df -t
can be used to check this). In that case you will need to recreate the file system (backup/restore data) with more inodes or switch to a filesystem type that allocates them dynamically. -
boddhisattva about 11 yearsThis happened on our prod box, while doing a git fetch. Freeing up some disk space fixed it for us.
-