I had a closer look at this proposal, and given that this is a cross
distribution challenge, it seem better to me to change the default
behaviour of zfs, instead of creating a configuration file during
installation on the Linux distributions that remember to do so.
What about modifying the zfs sou
[Craig Sanders]
> that's how conffiles are supposed to work. they'll get a question only
> when the package-supplied conffile is both:
I suspect you misunderstand me. The conffile mechanism is designed to
work well with system administrators manually editing files. Their
manual edits are not aut
On Sun, Sep 11, 2016 at 07:12:49AM +0200, Petter Reinholdtsen wrote:
> [Craig Sanders]
> > good idea. at the least, we should include an example in the
> > package. e.g. I made the following file for my systems, which I copy
> > to various machines and modify to suit.
>
> Make sense, as long as it
[Craig Sanders]
> good idea. at the least, we should include an example in the package.
> e.g. I made the following file for my systems, which I copy to various
> machines and modify to suit.
Make sense, as long as it isn't stored under /etc/ where any chnages to
the examples will trigger a conffi
On Thu, Sep 08, 2016 at 09:59:56AM +, Lumin wrote:
> I suggest that a little enhancement should be made to postinst
> -- let the user input a zfs_arc_max value and then put it into
> /etc/modprobe.d/zfs.conf. If user input nothing then nothing will
> happen.
good idea. at the least, we should
Package: zfsutils-linux
Version: 0.6.5.7-2
Severity: minor
Hi,
We never know on what machine will the user deploy ZFS, and
it may be expensive to run zfs under the default arc config.
For example, my laptop has 4G memory and nearly half of
it would be gone after zfs being loaded.
With reference
6 matches
Mail list logo
