Re: Bug#661283: cmor: FTBFS on kfreebsd-amd64

[Steven Chamberlain <steven@pyro.eu.org>]

reopen 661283
done

On 03/03/12 14:11, Julien Cristau wrote:
> I gave it back and it built, so closing.

Hi,

Doesn't this want fixing more permanently?  It got queued for a rebuild
for some reason and has been failing again since:

https://buildd.debian.org/status/logs.php?pkg=cmor&arch=kfreebsd-amd64

The test suite with ~700MiB process images are presumably asking a bit
too much of the kfreebsd-amd64 buildds (and way too much for
kfreebsd-i386, #598745).  That's odd though if they have 6GiB RAM in total.

I suspect a limit is be being hit, something similar to kern.maxdsiz
although that one is per-process so wouldn't explain why this happens
only sometimes.

I'm curious anyway what are the kern.maxdsiz and kern.dfldsiz on the
buildds (I'm guessing 1GiB on amd64) and is it permissible for them to
be raised (to maybe 2GiB) if some package like this one required it?
(I'm also wondering about the openjdk-7 failures here.)

Or, should/could the cmor test suite be slimmed down to suit?

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