Re: Bits from the Release Team (Jessie freeze info)

2013-10-23 Thread Thorsten Glaser
Steven Chamberlain dixit:

>Come to think of it, it must take a day or more for m68k to rebuild
>eglibc.  This is a more serious problem than resources needed by

Kernel takes a day now (on the fastest VMs), eglibc 3 days,
gcc 5 days (since gcj got folded into it; add another day or
so once gnat will also be folded).

>Jenkins.  We can't ask them to rebuild their entire toolchain each night!

No OpenJDK either (can probably be fixed, but zero is sloow).

Additionally, with only, say, 256 or 768 MiB physmem, running
additional software on the buildds is something you do not want,
considering how much RAM building some stuff takes (I had to use
about 5 GiB of swap to link Webkit, and imagine just how much
paging that involves, also in terms of time). Building GCC isn’t
exactly resource-saving. (Even running apt/dpkg isn’t due to the
sheer size of the archive, though Guillem kindly reduced memory
usage in the upcoming dpkg upload.)

I think with my “better SCC proposal” we could have a sliding
scale for this, but I’d oppose using something OpenJDK-based
for that (think of mipsel, too). Especially as simple mksh
scripts would take care of the job too (including CGI for web
export ;).

bye,
//mirabilos
-- 
Solange man keine schmutzigen Tricks macht, und ich meine *wirklich*
schmutzige Tricks, wie bei einer doppelt verketteten Liste beide
Pointer XORen und in nur einem Word speichern, funktioniert Boehm ganz
hervorragend.   -- Andreas Bogk über boehm-gc in d.a.s.r


--
To UNSUBSCRIBE, email to debian-sparc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/pine.bsm.4.64l.1310231242320.29...@herc.mirbsd.org



Re: Bits from the Release Team (Jessie freeze info)

2013-10-23 Thread Steven Chamberlain
On 23/10/13 12:55, Stewart Smith wrote:
> Geert Uytterhoeven  writes:
>> On Wed, Oct 23, 2013 at 12:36 AM, Stewart Smith
>>  wrote:
>>> Jenkins can have slaves on remote hosts, via SSH. It runs a small java
>>> app there, so as long as the arch has a JVM then you're pretty right.
>>
>> For whatever definition of small. I've seen it consuming 1 GiB of
>> memory...
> 
> with 'm68k' in your email address your definition of small is likely
> much different than my "many years in large scale databases" small :)

Come to think of it, it must take a day or more for m68k to rebuild
eglibc.  This is a more serious problem than resources needed by
Jenkins.  We can't ask them to rebuild their entire toolchain each night!

For the goal of software freedom, it shouldn't be too difficult for
anyone to do that, though.  We should be trying to make it easier.

Maybe it would be permissible for the toolchain test suite to run on a
faster platform, and cross-compile, or use any sort of emulation
available in Debian free packages.

If it were technically feasible for each Debian port to rebuild its
toolchain and some essential packages, at least once per week, I think
that would be an accomplishment.  And the smaller the initial set of
packages required to boostrap the process, the better.

Regards,
-- 
Steven Chamberlain
ste...@pyro.eu.org



signature.asc
Description: OpenPGP digital signature


Re: Bits from the Release Team (Jessie freeze info)

2013-10-23 Thread Stewart Smith
Geert Uytterhoeven  writes:
> On Wed, Oct 23, 2013 at 12:36 AM, Stewart Smith
>  wrote:
>> Jenkins can have slaves on remote hosts, via SSH. It runs a small java
>> app there, so as long as the arch has a JVM then you're pretty right.
>
> For whatever definition of small. I've seen it consuming 1 GiB of
> memory...

with 'm68k' in your email address your definition of small is likely
much different than my "many years in large scale databases" small :)

That being said... I haven't recently seen a slave jenkins java process
more than one or two hundred mb.

This is (of course) absolutely insane, as is the 4-6GB jenkins master
process. However, dollars per GB of memory is suitably low that it's not
worth me fixing it, instead it just sits there annoying me as it could
undoubtedly be better


-- 
Stewart Smith


pgpBBYdyNkP40.pgp
Description: PGP signature


Re: Bits from the Release Team (Jessie freeze info)

2013-10-23 Thread Steven Chamberlain
On 22/10/13 23:36, Stewart Smith wrote:
> Jenkins can have slaves on remote hosts, via SSH. It runs a small java
> app there, so as long as the arch has a JVM then you're pretty right.

That may be useful to set up on some arches, for things where Jenkins
needs direct control over CPU-intensive tasks.  Building and testing
d-i, for example.

But for this, I would imagine only the test suite needs to run over SSH,
and the master Jenkins instance just has to process the output?

For the proposed test suite to be as accessible as possible to a
new/upcoming port, the barrier to using it ought to be very low.  A
working JVM is quite a lot to ask, the current openjdk-7 is not even
built for mipsel in more.  mipsel buildds and porterboxes had only 1GB
RAM maximum until now, and that is heavily used already for their
current tasks.

Regards,
-- 
Steven Chamberlain
ste...@pyro.eu.org



signature.asc
Description: OpenPGP digital signature


Re: Bits from the Release Team (Jessie freeze info)

2013-10-23 Thread Britt Dodd
Small is 64m ram not 256m. I just woke up and was catching up on things. My
apologies.
On Oct 23, 2013 7:20 AM, "Britt Dodd"  wrote:

> I run Jenkins at my job. Small is around 256mb. Plus the Jenkins server
> can sit on a high-memory machine and the agent just sit on a 68k box doing
> builds. Small is like 64M ram. You Amiga/Atari guys seem to have oodles of
> ram to work with Lol.
> On Oct 23, 2013 2:45 AM, "Geert Uytterhoeven" 
> wrote:
>
>> On Wed, Oct 23, 2013 at 12:36 AM, Stewart Smith
>>  wrote:
>> > Jenkins can have slaves on remote hosts, via SSH. It runs a small java
>> > app there, so as long as the arch has a JVM then you're pretty right.
>>
>> For whatever definition of small. I've seen it consuming 1 GiB of
>> memory...
>>
>> Gr{oetje,eeting}s,
>>
>> Geert
>>
>> --
>> Geert Uytterhoeven -- There's lots of Linux beyond ia32 --
>> ge...@linux-m68k.org
>>
>> In personal conversations with technical people, I call myself a hacker.
>> But
>> when I'm talking to journalists I just say "programmer" or something like
>> that.
>> -- Linus Torvalds
>>
>>
>> --
>> To UNSUBSCRIBE, email to debian-68k-requ...@lists.debian.org
>> with a subject of "unsubscribe". Trouble? Contact
>> listmas...@lists.debian.org
>> Archive:
>> http://lists.debian.org/camuhmdvz3jwmdujds762z-cnhv4z5c9wuuf5rkanarqbsdx...@mail.gmail.com
>>
>>


Re: Bits from the Release Team (Jessie freeze info)

2013-10-23 Thread Britt Dodd
I run Jenkins at my job. Small is around 256mb. Plus the Jenkins server can
sit on a high-memory machine and the agent just sit on a 68k box doing
builds. Small is like 64M ram. You Amiga/Atari guys seem to have oodles of
ram to work with Lol.
On Oct 23, 2013 2:45 AM, "Geert Uytterhoeven"  wrote:

> On Wed, Oct 23, 2013 at 12:36 AM, Stewart Smith
>  wrote:
> > Jenkins can have slaves on remote hosts, via SSH. It runs a small java
> > app there, so as long as the arch has a JVM then you're pretty right.
>
> For whatever definition of small. I've seen it consuming 1 GiB of memory...
>
> Gr{oetje,eeting}s,
>
> Geert
>
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 --
> ge...@linux-m68k.org
>
> In personal conversations with technical people, I call myself a hacker.
> But
> when I'm talking to journalists I just say "programmer" or something like
> that.
> -- Linus Torvalds
>
>
> --
> To UNSUBSCRIBE, email to debian-68k-requ...@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact
> listmas...@lists.debian.org
> Archive:
> http://lists.debian.org/camuhmdvz3jwmdujds762z-cnhv4z5c9wuuf5rkanarqbsdx...@mail.gmail.com
>
>