[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Vancouver meeting - clarifications



* Andreas Barth 

| Well, that was one of the examples where we pay a price for more
| architectures. Of course, the testing migration script is not all, and
| this problem can be solved, but I think we should not forget that we pay
| a price - even if at the end, we think the price is acceptable.

Sure.

| (And, BTW, newraff is a quite mature box. Of course, there is always
| more and better hardware available, but newraff is already a very good
| machine. And, we want to give the testing migration script more tasks,
| like handling of the udebs, which puts load off from human being and on
| into a script. But that increases the ressources the script uses.)

If throwing hardware at a problem solves the problem and we're talking
about reasonable amounts of hardware and this helps the release, I
think it's an entirely appropriate thing to do.  It helps the
volunteers use their time well.  We haven't gotten donations to let
them sit around in a bank account, unused.  We've gotten them to
improve the distribution.  If that means buying hardware, fine, we'll
do that.

(As I'm not a RM/RA and don't have access to newraff, I didn't know
newraff ran out of memory due to the testing scripts -- in fact, I
thought it was doing very well, given that auric used to have problems
completing the testing scripts at all (IIRC), while aj said the
scripts take about 20 minutes on newraff now (he posted this in
January).)

I think identifying those kinds of bottlenecks is important and will
help us make releases happen faster.

-- 
Tollef Fog Heen                                                        ,''`.
UNIX is user friendly, it's just picky about who its friends are      : :' :
                                                                      `. `' 
                                                                        `-  



Reply to: