Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-22 Thread Wouter Verhelst
On Mon, Mar 21, 2005 at 01:44:37PM -0800, Steve Langasek wrote: > On Mon, Mar 21, 2005 at 06:23:30PM +0100, Wouter Verhelst wrote: > > On Sat, Mar 19, 2005 at 02:12:50PM -0800, Steve Langasek wrote: > > > TTBOMK, even m68k has one buildd admin per buildd > > > False. There are some of us who curre

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-21 Thread Steve Langasek
On Mon, Mar 21, 2005 at 06:23:30PM +0100, Wouter Verhelst wrote: > On Sat, Mar 19, 2005 at 02:12:50PM -0800, Steve Langasek wrote: > > TTBOMK, even m68k has one buildd admin per buildd > False. There are some of us who currently don't maintain more than one > buildd host, but with the exception of

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-21 Thread Wouter Verhelst
On Sat, Mar 19, 2005 at 02:12:50PM -0800, Steve Langasek wrote: > TTBOMK, even m68k has one buildd admin per buildd False. There are some of us who currently don't maintain more than one buildd host, but with the exception of Roman, we all have (or have had) more than one buildd host under our res

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-21 Thread Wouter Verhelst
On Sat, Mar 19, 2005 at 03:17:33AM -0800, Steve Langasek wrote: > > - at least two buildd administrators > > > This allows the buildd administrator to take vacations, etc. > > This is at odds with what I've heard from some buildd maintainers that > having multiple buildd maintainers makes it hard

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-20 Thread Stephen Frost
* Andreas Barth ([EMAIL PROTECTED]) wrote: > * Steve Langasek ([EMAIL PROTECTED]) [050319 12:35]: > > On Wed, Mar 16, 2005 at 12:20:34AM -0800, Blars Blarson wrote: > > > - at least two buildd administrators > > > > > This allows the buildd administrator to take vacations, etc. > > > This is at o

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-20 Thread Andreas Barth
* Steve Langasek ([EMAIL PROTECTED]) [050319 12:35]: > On Wed, Mar 16, 2005 at 12:20:34AM -0800, Blars Blarson wrote: > > - at least two buildd administrators > > > This allows the buildd administrator to take vacations, etc. > This is at odds with what I've heard from some buildd maintainers tha

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-19 Thread Matthias Urlichs
Hi, Steve Langasek wrote: >> TTBOMK, m68k has no such problem. > > TTBOMK, even m68k has one buildd admin per buildd -- the most they > generalley have in terms of buildd admin redundancy is that if the admin > for a machine that has built a certain package is unavailable, another > admin can was

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-19 Thread Steve Langasek
On Sat, Mar 19, 2005 at 04:37:05PM +0100, Matthias Urlichs wrote: > Hi, Steve Langasek wrote: > >> This allows the buildd administrator to take vacations, etc. > > This is at odds with what I've heard from some buildd maintainers that > > having multiple buildd maintainers makes it hard to avoid

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-19 Thread Matthias Urlichs
Hi, Steve Langasek wrote: >> This allows the buildd administrator to take vacations, etc. > > This is at odds with what I've heard from some buildd maintainers that > having multiple buildd maintainers makes it hard to avoid stepping on one > another's feet, I assume that that's a problem if the

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-19 Thread Steve Langasek
On Wed, Mar 16, 2005 at 12:20:34AM -0800, Blars Blarson wrote: > In article <[EMAIL PROTECTED]> [EMAIL PROTECTED] writes: > >- the release architecture must have N+1 buildds where N is the number > > required to keep up with the volume of uploaded packages > If we are going to require redundancy,

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-16 Thread Ingo Juergensmann
On Wed, Mar 16, 2005 at 04:37:13PM -0500, Roberto C. Sanchez wrote: > >For example both public debian m68k machines are located on the same window > >sill at the Univ. of Duesseldorf. IMHO not the best place to position > >important infrastructure. > I agree. A sturdy table, or even a shelf or se

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-16 Thread Roberto C. Sanchez
Ingo Juergensmann wrote: On Wed, Mar 16, 2005 at 12:20:34AM -0800, Blars Blarson wrote: If we are going to require redundancy, I think we should do it better and add: - at least two buildd administrators *nod* - systems located in at least two different facilities (different cities and backbone

Re: Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-16 Thread Ingo Juergensmann
On Wed, Mar 16, 2005 at 12:20:34AM -0800, Blars Blarson wrote: > If we are going to require redundancy, I think we should do it better > and add: > - at least two buildd administrators *nod* > - systems located in at least two different facilities (different > cities and backbones if at all po

Buildd redundancy (was Re: Bits (Nybbles?) from the Vancouver...)

2005-03-16 Thread Blars Blarson
In article <[EMAIL PROTECTED]> [EMAIL PROTECTED] writes: >- the release architecture must have N+1 buildds where N is the number > required to keep up with the volume of uploaded packages If we are going to require redundancy, I think we should do it better and add: - at least two buildd adminis