The main performance thing is using an external SATA SSD, rather than
onboard flash which is usually extremely slow. The Cubietrucks are
reasonable for this.
Justin
On 12 November 2015 at 21:08, Stefan Xenon wrote:
> Hi!
> As a newbie I'm a bit confused which boards and CPUs work with
> MirageO
On 6 September 2015 at 12:10, Gareth Rushgrove wrote:
> Hi
>
> I have a vague plan of automating the build-out of a box to run many
> unikernels, probably on Packet*.
>
> Does anyone have an up-to-date set of steps for installing Xen,
> including with a suitable network setup and that doesn't requ
On 25 August 2015 at 12:46, Thomas Leonard wrote:
> On 22 August 2015 at 18:37, Gareth Rushgrove wrote:
>> On 22 August 2015 at 12:34, Thomas Leonard wrote:
>>> On 21 August 2015 at 17:07, Gareth Rushgrove
>>> wrote:
I'd managed to get a bunch of throughts out of how head an into blog
>>>
On 3 August 2015 at 15:26, Dave Scott wrote:
> I was wondering if we should move away from crunch, and use some other method
> for static data. Mirage already supports static data
It is certainly useful for some applications not to use RAM.
> - from Irmin
> - from BLOCK devices formatted with F
On 9 July 2015 at 15:04, David Sheets wrote:
> Off the top of my head I notice two things:
>
> 1. The int64 and tuple representations make a timestamp use 2*word +
> 12 bytes if I count correctly. That is 20 bytes on a 32-bit machine
> and 28 bytes on a 64-bit machine to represent 12 bytes of data
On 11 June 2015 at 15:37, Anil Madhavapeddy wrote:
> On 9 Jun 2015, at 04:04, Vincent Bernardoff wrote:
>>
>> On 07/06/2015 19:45, Nicolas Ojeda Bar wrote:
>>> Forwarding in case it is of interest.
>>>
>>> Best wishes,
>>> Nicolas
>>
>> This is really cool, thanks!
>>
>> I was wondering, since yo
On 11 June 2015 at 15:25, Anil Madhavapeddy wrote:
>
>> On 9 Jun 2015, at 17:47, Justin Cormack wrote:
>>
>> On 9 June 2015 at 11:16, Justin Cormack wrote:
>>> On 9 June 2015 at 09:28, Thomas Gazagnaire wrote:
>>>>>> or just inside
On 10 June 2015 at 07:50, Gareth Rushgrove wrote:
>> I made a native Musl OCaml build that you can use with
>> docker run -i -t justincormack/alpine-pkgsrc-ocaml /bin/sh
>>
>
> Awesome, do publish how when you get a moment.
It is a pkgsrc bootstrap and build from source on top of Alpine Linux - s
On 9 June 2015 at 11:16, Justin Cormack wrote:
> On 9 June 2015 at 09:28, Thomas Gazagnaire wrote:
>>>> or just inside a scratch container and get the following error:
>>>>
>>>>System error: no such file or directory
>>>>
>>>&
On 9 June 2015 at 09:28, Thomas Gazagnaire wrote:
>>> or just inside a scratch container and get the following error:
>>>
>>>System error: no such file or directory
>>>
>>> So, questions:
>>>
>>> * Is the unix unikernel statically linked? ie. is the above ever going to
>>> work?
>>> * What ab
On 8 June 2015 at 21:55, Gareth Rushgrove wrote:
> I'm experimenting with unix unikernels and docker, specifically
> looking at the smallest container I can get running a unix unikernel.
> I have a few questions:
>
> I've built a unix unikernel on Ubuntu (14.04). The resulting binary
> runs succes
On 6 June 2015 at 23:54, Anil Madhavapeddy wrote:
> On 6 Jun 2015, at 15:49, Justin Cormack
> wrote:
>
>
>
> On Jun 6, 2015 4:56 PM, "Anil Madhavapeddy" wrote:
>>
>> I'm just testing out tuntap.1.3 for release on Linux, and am having some
>> t
On Jun 6, 2015 4:56 PM, "Anil Madhavapeddy" wrote:
>
> I'm just testing out tuntap.1.3 for release on Linux, and am having some
trouble getting it to work. Just wanted to check if others are using it
ok. I've done:
>
> $ sudo tunctl
> $ sudo ifconfig tap0 10.0.0.1 netmask 255.255.255.0 u
On 6 May 2015 at 10:40, Mindy wrote:
> - improve and formalize (create?) error communication for bad state
> transitions, syntactically OK but semantically bogus packets, etc
The only useful error communication I can think of is counters of
types of error, which you can use when deliberately te
On 5 May 2015 at 09:48, Anil Madhavapeddy wrote:
> On 5 May 2015, at 01:47, Justin Cormack wrote:
>>
>> On 5 May 2015 at 09:35, Anil Madhavapeddy wrote:
>>> On 4 May 2015, at 14:50, Justin Cormack
>>> wrote:
>>>>
>>>> I have se
On 5 May 2015 at 09:35, Anil Madhavapeddy wrote:
> On 4 May 2015, at 14:50, Justin Cormack wrote:
>>
>> I have sent a pull request to fix the requires root issue - it was
>> trying to bring the interface up which requires root, so long as it is
>> already up this is
On 29 April 2015 at 15:56, Mindy wrote:
>>> Third, when I managed to get one of the network samples building, for
>>> example "ping", I couldn't get the unikernel running locally with a
>>> tun/tap
>>> device I created. Even though I use tapctl to create a tap0 device owned
>>> by
>>> my user acco
On 4 May 2015 at 15:10, Anil Madhavapeddy wrote:
> On 4 May 2015, at 07:03, Gareth Rushgrove wrote:
>>
>> I'm tinkering with some automation to make getting started with Mirage
>> easier and ran into a few issues I thought I'd mention.
>>
>> While following the docs at: http://openmirage.org/wiki
On 29 April 2015 at 18:33, Nick Betteridge wrote:
> modprobe: FATAL: Module tun not found.
>
> Doing a :
>
> sudo tunctl
>
> gives:
>
> Failed to open '/dev/net/tun' : No such device
>
>
> Does anyone have any experience on this?
Yes, the precompiled kernels for Arm are generally missing lots of
On 29 April 2015 at 10:35, Thomas Gazagnaire wrote:
> Third, when I managed to get one of the network samples building, for
> example "ping", I couldn't get the unikernel running locally with a tun/tap
> device I created. Even though I use tapctl to create a tap0 device owned by
> my user account,
On 17 April 2015 at 21:49, Nick Hardiman wrote:
> What is the advantage of Scaleways, apart from the processor? What kind of
> centralized hosting do you want?
For Arm builds and so on they have proved rather more reliable than my
in-house hosting, with good performance (quad core, SSD).
> I’m g
21 matches
Mail list logo
