[MIR] eatmydata

Bug #1126328 reported by Phillip Susi on 2013-02-15
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
debootstrap (Ubuntu)
Undecided
Unassigned
libeatmydata (Ubuntu)
Undecided
Unassigned

Bug Description

Please promote eatmydata to main.

The reason for this request is to support bug #1126314, which requests debootstrap to make use of eatmydata to make installs and debootstrapping chroots 5 times faster. This is a very simple LD_PRELOAD library that overrides fsync() and related calls to be a noop. As such, it is extremely unlikely to have any security issues.

Michael Terry (mterry) wrote :

The package itself is fine. There is a bug in Debian that will affect some builds:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=667965

Also, there have been some upstream releases that haven't been packaged in Debian. Debian hasn't seen an update in two years. But since there aren't any super-bad bugs, in Ubuntu or Debian, I suppose it's fine. The maintainer himself is still active. Just hasn't updated this package for some reason.

And a bug subscriber in Ubuntu would be nice.

Approved.

Changed in libeatmydata (Ubuntu):
status: New → Fix Committed
Phillip Susi (psusi) wrote :

Good point about the package needing updated. I think that bug is in gnulib, not eatmydata. When fsync() doesn't actually do anything, it shouldn't have to report an error that would otherwise happen if it DID try to sync. I have emailed gnulib about it and will file a bug report to update the eatmydata package.

Scott Moser (smoser) wrote :

psusi, do you have any more information on this?
I'm interested in getting eatmydata into main this cycle also (bug 1236531).

Phillip Susi (psusi) wrote :

I don't think it was ever fixed and the gnulib guys said eatmydata should use fcntl to check if the fd is valid or not before returning the correct status.

Matthias Klose (doko) wrote :

Override component to main
libeatmydata 26-2 in trusty: universe/utils -> main
eatmydata 26-2 in trusty amd64: universe/utils/optional/100% -> main
eatmydata 26-2 in trusty arm64: universe/utils/optional/100% -> main
eatmydata 26-2 in trusty armhf: universe/utils/optional/100% -> main
eatmydata 26-2 in trusty i386: universe/utils/optional/100% -> main
eatmydata 26-2 in trusty powerpc: universe/utils/optional/100% -> main
eatmydata 26-2 in trusty ppc64el: universe/utils/optional/100% -> main
7 publications overridden.

Changed in libeatmydata (Ubuntu):
status: Fix Committed → Fix Released

I don't think deboostrap needs changes

Changed in debootstrap (Ubuntu):
status: New → Incomplete
Jeremy Bicha (jbicha) wrote :

I'm closing the debootstrap task so that this bug doesn't show up on the MIR review list. There are other bugs requesting that debootstrap pull in libeatmydata.

Changed in debootstrap (Ubuntu):
status: Incomplete → Invalid
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.