Re: Proposal to replace/extend current armhf builders
Wookey <email@example.com> writes:
> +++ Konstantinos Margaritis [2013-11-13 23:05 +0200]:
>> Hi all,
>> Here is some food for thought for the minidebconf that starts tomorrow
>> in Cambridge . Unfortunately I will not make it there, though I wish
>> I did, but I'm in the process of job searching at the moment and
>> could not afford the expense.
> We miss you markos.
>> Upgrade 5
>> of the iMX53s with 5 Odroid-XU and be done with armhf builders for a
>> long time.
>> This is a counter-proposal to what Hector Oron (zumbi) suggested: that
>> we buy or get donations for real server class hardware (arm64).
> Or arm32, for just the two 32-bit ports.
> This was discussed somewhat today, and we all agree that server-grade
> hardware would be best. James from Boston told us that half a chassis
> full would be about $8000 (or was that £8000?) Either way we can't afford it.
> Arm64 kit will be even more expensive/unobtanium for a while. Although I
> expect the situation to be very different laster next year.
>> 1. Terribly expensive: a single board costs ~7k USD. And it needs the
>> chassis which is also expensive. So if you take redundancy in mind, you
>> need 2 setups in 2 locations, so at least 14k just for the boards and
>> I'd not be surprised if the chassis costs more than 5k itself. So in
>> total ~25k.
> It's expensive, but not that bad. Or are you talking about 64-bit
> hardware? Has anyoneannounced prices for that yet?
I guess he was talking about systems like highbank or midway
>> 3. Liability. Right now armhf builders (iMX53 Loco boards) are in ARM HQ
>> and York University, if I'm not mistaken. If something happens and that
>> equipment breaks due to eg. some fault in the electrical wiring and a
>> board burns, the boards are very cheap to replace and there is no
>> issue there. However with $10k+ worth of equipment I'm very doubtful if
>> ARM or York or anyone really would accept to host these systems and
>> accept liability in case of fault.
> I think that's nonsense. If we had a box we could find somewhere to house it.
>> So, my counter-proposal to that is that we get instead some cheap easy
>> to replace boards like the Arndale  or the Odroid-XU . Personally
>> I'd prefer the XU as it's better equipped, but I'd go with either
>> choice if people think it's better.
>> Finally, last but not least, there is the issue of mainline kernel
>> support for exynos5 and the odroid/arndale boards.
> There was very strong resistance in the room to the use of buildds
> without debian kernel support, because it's a major hassle and security
> risk. That rules out odroid for the time being.
I won't say it's bad or good but I'll say again what Markos told in a
previous mail: running buildd with non-mainline kernels already happened
in the past (while waiting support being merged).
>> If people *really* don't want non-mainline kernels, then we could go
>> for the utilite boards, I would suggest the Pro version:
> Or Wanda or the Nitrogen6x we've just kindly been offered.
> What's the mainline status of those?
I remember seeing some patches for wanda but I can't say the same for
utilite or nitrogen. I don't remember how hard it would be to push
support for them upstream / what's missing.
>> Anyway, apologies for my long email,
> No need to apologise - it's a very useful summary and useful basis for discussion.
> I think we need to bash out some criteria for deciding during the
> mini-debconf. i.e deciding how we decide (or just make a decision if possible).
> e.g hold out for server-grade kit for a while - if so how long? (0 months, 3 months, 6months, longer?)
there are technicals issues but we have to take care about how much
money can be used to buy hw imho.
> Is debian kernel an absolute requirement, or are we prepared to risk a
> custom kernel if we think it'll only be for 6 months?
it depends also on what are the changes needed:
- either hw support patches
- or only a missing dts.
my 2 cents,