[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#899303: marked as done ([Feature Request] Add option to use libeatmydata)

Your message dated Fri, 5 Apr 2019 15:14:47 +0200
with message-id <20190405131447.GE10131@crossbow>
and subject line Re: Bug#899303: [Feature Request] Add option to use libeatmydata
has caused the Debian Bug report #899303,
regarding [Feature Request] Add option to use libeatmydata
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@bugs.debian.org
immediately.)


-- 
899303: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=899303
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems
--- Begin Message --- 
Package: apt
Version: 1.6.1

Dear APT developers,

I'd like to request for an APT feature, where one could add an option
via either /etc/apt/apt.conf or command line argument specifing whether
to use libeatmydata1 to silent the fsync call and friends.

Eatmydata is used to speed up debootstrap speed for pbuilder. Despite
the risk that data may go into an incosistent state on error, this 
library brings a literally insane speed up.

I believe there will be users who are willing to take the risks and
boost the APT speed.

Do you agree to add this feature? If you don't have enough time to
implement this, please let me know. I'd be happy to contribute a patch.

Thanks.

Attachment: signature.asc
Description: PGP signature


--- End Message ---
--- Begin Message --- 
Hi,

On Tue, May 22, 2018 at 10:52:47AM +0000, Lumin wrote:
> I'd like to request for an APT feature, where one could add an option
> via either /etc/apt/apt.conf or command line argument specifing whether
> to use libeatmydata1 to silent the fsync call and friends.

as APT isn't doing much in terms of syncing that would effect runtime
I suspect we are talking about the moment dpkg is called?

There are options to define how dpkg is called, so it is possible to do
what you are asking for already by warping the dpkg call in eatmydata.
If it is a good idea through…

A package upgrade can involve the (re)starting of background processes
which can potentially (depending on the tools involved) inherent the
data eating which can be rather surprising for users as they thought it
would only apply to apt/dpkg and not outlife them via services.

dpkg has the option --force-unsafe-io which disables most syncs without
having the effect eatmydata has on processes started from it – you can
add that via apt options to the dpkg calls it makes as well.


So between all these options and the semi-hidden risks involved I don't
see a good reason to add yet another option to make that potentially
dangerous undertaking seem more appealing.

I am therefore closing this feature request as semi-refused,
semi-already-implemented.


Best regards

David Kalnischkies

Attachment: signature.asc
Description: PGP signature


--- End Message ---
Reply to: