open files with O_NOATIME if possible

[c4rlo]

The Linux manpage for open(2) describes O_NOATIME thus:

   O_NOATIME (since Linux 2.6.8)
          Do  not  update the file last access time (st_atime in the inode) when the file is read(2).  This flag
          is intended for use by indexing or backup programs, where its use can significantly reduce the  amount
          of  disk  activity.  This flag may not be effective on all filesystems.  One example is NFS, where the
          server maintains the access time.

Change attic to provide this flag where possible.

Unrelatedly, fix a typo in a comment.

[skarekrow]

Does this impede usage on other systems? Solaris, *BSD's and such?
On 9/12/2014 4:33 PM, c4rlo wrote:

The Linux manpage for |open(2)| describes |O_NOATIME| thus:

| O_NOATIME (since Linux 2.6.8)
Do not update the file last access time (st_atime in the inode) when the file is read(2). This flag
is intended for use by indexing or backup programs, where its use can significantly reduce the amount
of disk activity. This flag may not be effective on all filesystems. One example is NFS, where the
server maintains the access time.
|

Change attic to provide this flag where possible.

Unrelatedly, fix a typo in a comment.

---

    You can merge this Pull Request by running

git pull https://github.com/c4rlo/attic noatime

Or view, comment on, or merge it at:

#113

    Commit Summary

• open files with O_NOATIME if possible

  File Changes
  

• M attic/archive.py
  https://github.com/jborg/attic/pull/113/files#diff-0 (43)

  Patch Links:
  

• https://github.com/jborg/attic/pull/113.patch

• https://github.com/jborg/attic/pull/113.diff

—
Reply to this email directly or view it on GitHub
#113.

[c4rlo]

It should do. If os.O_NOATIME doesn't exist, I fall back to the old behaviour.

[c4rlo]

I've actually been trying to test this under FreeBSD. Unfortunately, building cython fails with a segfault.

But like I said, it should work under all systems.

[skarekrow]

Ok that's great to hear. Odd you're having segfaults building cython, but obviously that's not related to this change. If he merges it, I use FreeBSD as my home server, so I'll know if something doesn't function hopefully.

[anarcat]

this seems to be an important change that should be merged in...

[meskio]

I agree this is an important change that should be merged. I test it and it works fine in my system and I'm using it.

@jborg can I do something to help on getting this pull-req merged?

[jborg]

Looks good, thanks!

[anarcat]

awesome!!! thank you so much!