irqbalance as ubuntu-server dep...

[Frédéric Dreier]

Hi,

I tried to tuned some new servers with 10GB NICs, and found on the
internet several references where irqbalance may be negative even for
servers.

http://bit.ly/SatTsH

I did not investigate further. But you may be interested to.

Regards


2012/7/12 Scott Kitterman <ubuntu at kitterman.com>:
> On Thursday, July 12, 2012 11:26:17 AM Daniel J Blueman wrote:
>> On 10 July 2012 22:34, Scott Kitterman <ubuntu at kitterman.com> wrote:
>> > On Tuesday, July 10, 2012 09:47:33 PM Daniel J Blueman wrote:
>> >> Hi Ubuntu devs,
>> >>
>> >> I've been wanting to raise the question about if ubuntu-standard
>> >> should depend on irqbalance, since the benefits of irqbalance in a
>> >> typical non-server environment are questionable.
>> >>
>> >> In fact, I see it clocking up time, causing more context switches and
>> >> wakeups. This is clearly a step backward for mobile and even desktop
>> >> users. The documentation [1] shows the use-cases it was designed for,
>> >> ie heavy server load.
>> >>
>> >> It would be far better as a dependency in an Ubuntu server
>> >> installation. What do you guys think?
>> >
>> > My recollection is that it was added to the seeds for the server use case,
>> > so I think this makes sense.
>>
>> https://bugs.launchpad.net/ubuntu/+source/ubuntu-meta/+bug/1023020
>>
>> Is a debdiff suitable for the ubuntu-standard metapackage, or should I
>> present a bzr branch?
>
> Ubuntu metapackages are generated from seeds.  See:
>
> https://wiki.ubuntu.com/SeedManagement
>
> This is trivial to change, but I'd like to see a bit more agreement than my
> vague recollection before doing it.
>
> Scott K
>
> --
> Ubuntu-devel-discuss mailing list
> Ubuntu-devel-discuss at lists.ubuntu.com
> Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss