Re: HPPA and lenny
Peter Palfrader schrieb:
> while discussing the state of our infrastructure among some members of DSA
> and the security team we once more realized how badly hppa was doing. First
> concerns about this architecture were raised at least half a year ago, and it
> doesn't appear as if we are any closer to resolving them. It doesn't look
> good at all.
> We do have two autobuilders, but one of them is currently down for whatever
> reason (and the rsys excuse for remote management is a joke):
> | [17208343.620000] awk (pid 1061): Protection id trap (code 27) at 00000000400b4763
> | [17208343.712000] Backtrace:
> | [17208343.740000]
> (that's the complete output.)
> This may or may not be related to the issue where everytime something
> even remotely related to ruby gets close to the system they lock up.
> At least the (random) crashing issue has to be fixed if hppa wants
> to release with lenny.
> We already had this discussion in June and again in October but I don't think
> anything came from it.
> Having little to no buildd redundancy is currently hurting us even for
> etch as security updates are missing for at least perl.
> (Also, the porter machine is currently down. We cannot reach it over
> the network and the serial console is just dead.)
> With my DSA hat on I can say that we will not be able to provide porter
> machines and buildd systems if they are not capable of running reliably.
> Again, if you want to release with lenny *you* will have to fix that.
I just installed a old hppa (c100) machine with etch. Maybe the machine
could help as an autobuilder.
What do you think?