7 ms·

Couldn't `mv` or `cp` from the temp file to `/etc/passwd` be interrupted as well? I think the only way to do it atomically is a temporary file on the same files

Couldn't `mv` or `cp` from the temp file to `/etc/passwd` be interrupted as well? I think the only way to do it atomically is a temporary file on the same filesystem as `/etc`, followed by a rename. On most systems `/tmp` will be a different filesystem from `/etc`.

AdamJacobMuller 5y ago
mv can't, or, more correctly the rename system call can not. rename is an atomic operation from any modern filesystem's perspective, you're not writing new data, you're simply changing the name of the existing file, it either succeeds or fails. Keep in mind that if you're doing this, mv (the command line tool) as opposed to the `rename` system call, falls back to copying if the source and destination files are on different filesystems since you can not really mv a file across filesystems! In order to have truly atomic writes you need to: open a new file on the same filesystem as your destination file write contents call fsync call rename call sync (if you care about the file rename itself never being reverted). This is some very naive golang code (from when I barely knew golang) for doing this which has been running in production since I wrote it without a single issue: https://github.com/AdamJacobMuller/atomicxt/blob/master/file.go https://github.com/AdamJacobMuller/atomicxt/blob/master/file...
- JJMcJ 5y ago
  Not clear on need for fsync and sync. Are those for networked like NTFS or just as security against crashes. Logically on a single system there would be no effect assuming error free filesystem operation. Unless I'm missing something.
  - AnssiH 5y ago
    Without the fsync() before rename(), on system crash, you can end up with the rename having been executed but the data of the new file not yet written to stable storage, losing the data. ext4 on Linux (since 2009) special-cases rename() when overwriting an existing file so that it works safely even without fsync() (https://lwn.net/Articles/322823/ https://lwn.net/Articles/322823/), but that is not guaranteed by all other implementations and filesystems. The sync() at the end is indeed not needed for the atomicity, it just allows you to know that after its completion the rename will not "roll back" anymore on a crash. IIRC you can also use fsync() on the parent directory to achieve this, avoiding sync() / syncfs().
    - AdamJacobMuller 5y ago
      > ext4 on Linux (since 2009) special-cases rename This is interesting. The linked git entry (https://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.git/commit/?id=dbc85aa9f11d8c13c15527d43a3def8d7beffdc8 https://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.g...) from the LWN article says "Notice: this object is not reachable from any branch." Did this never get merged because I definitely saw this issue in production well after 2009. I guess it either got changed, or, a different patch applied but perhaps this https://github.com/torvalds/linux/blob/master/fs/ext4/namei.c#L3765-L3766 https://github.com/torvalds/linux/blob/master/fs/ext4/namei.... does it?
      - AnssiH 5y ago
        The patch just got rebased, here's the one that was actually applied in master for v2.6.30: https://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.git/commit/?id=8750c6d5fcbd3342b3d908d157f81d345c5325a7 https://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.g... And yes, the code you highlighted is exactly this special-case in its current form. The mount option "noauto_da_alloc" can be used to disable these software-not-calling-fsync safety features.
  - rcoveson 5y ago
    I'd like to know why as well. The inclusion of the fsync before the rename implies to me that the filesystem isn't expected to preserve order between write and rename. It could commit a rename before committing _past_ writes, which could leave your /etc/passwd broken after an outage at a certain time. I can't tell whether that's the case or not from cursory googling (everybody just talks about read-after-write consistency). Maybe it varies by filesystem? The final sync is just there for durability, not atomicity, like you say.
    - AdamJacobMuller 5y ago
      > the filesystem isn't expected to preserve order between write and rename Correct. The rename can succeed while the write of the file you just renamed gets rolled back.
sedatk 5y ago
You can use `/etc/passwd.new` as a temporary file to avoid the problems you mentioned. In the worst case, you'll have an orphaned passwd.new file, but /etc/passwd is guaranteed to remain intact.

AdamJacobMuller 5y ago

mv can't, or, more correctly the rename system call can not. rename is an atomic operation from any modern filesystem's perspective, you're not writing new data, you're simply changing the name of the existing file, it either succeeds or fails. Keep in mind that if you're doing this, mv (the command line tool) as opposed to the `rename` system call, falls back to copying if the source and destination files are on different filesystems since you can not really mv a file across filesystems! In order to have truly atomic writes you need to: open a new file on the same filesystem as your destination file write contents call fsync call rename call sync (if you care about the file rename itself never being reverted). This is some very naive golang code (from when I barely knew golang) for doing this which has been running in production since I wrote it without a single issue: https://github.com/AdamJacobMuller/atomicxt/blob/master/file.go https://github.com/AdamJacobMuller/atomicxt/blob/master/file...

JJMcJ 5y ago
Not clear on need for fsync and sync. Are those for networked like NTFS or just as security against crashes. Logically on a single system there would be no effect assuming error free filesystem operation. Unless I'm missing something.
- AnssiH 5y ago
  Without the fsync() before rename(), on system crash, you can end up with the rename having been executed but the data of the new file not yet written to stable storage, losing the data. ext4 on Linux (since 2009) special-cases rename() when overwriting an existing file so that it works safely even without fsync() (https://lwn.net/Articles/322823/ https://lwn.net/Articles/322823/), but that is not guaranteed by all other implementations and filesystems. The sync() at the end is indeed not needed for the atomicity, it just allows you to know that after its completion the rename will not "roll back" anymore on a crash. IIRC you can also use fsync() on the parent directory to achieve this, avoiding sync() / syncfs().
  - AdamJacobMuller 5y ago
    > ext4 on Linux (since 2009) special-cases rename This is interesting. The linked git entry (https://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.git/commit/?id=dbc85aa9f11d8c13c15527d43a3def8d7beffdc8 https://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.g...) from the LWN article says "Notice: this object is not reachable from any branch." Did this never get merged because I definitely saw this issue in production well after 2009. I guess it either got changed, or, a different patch applied but perhaps this https://github.com/torvalds/linux/blob/master/fs/ext4/namei.c#L3765-L3766 https://github.com/torvalds/linux/blob/master/fs/ext4/namei.... does it?
    - AnssiH 5y ago
      The patch just got rebased, here's the one that was actually applied in master for v2.6.30: https://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.git/commit/?id=8750c6d5fcbd3342b3d908d157f81d345c5325a7 https://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.g... And yes, the code you highlighted is exactly this special-case in its current form. The mount option "noauto_da_alloc" can be used to disable these software-not-calling-fsync safety features.
- rcoveson 5y ago
  I'd like to know why as well. The inclusion of the fsync before the rename implies to me that the filesystem isn't expected to preserve order between write and rename. It could commit a rename before committing _past_ writes, which could leave your /etc/passwd broken after an outage at a certain time. I can't tell whether that's the case or not from cursory googling (everybody just talks about read-after-write consistency). Maybe it varies by filesystem? The final sync is just there for durability, not atomicity, like you say.
  - AdamJacobMuller 5y ago
    > the filesystem isn't expected to preserve order between write and rename Correct. The rename can succeed while the write of the file you just renamed gets rolled back.

sedatk 5y ago

You can use `/etc/passwd.new` as a temporary file to avoid the problems you mentioned. In the worst case, you'll have an orphaned passwd.new file, but /etc/passwd is guaranteed to remain intact.