[HCoop-Discuss] why aren't we [admins] letting users log in? (was Re: Ruffled kerfuffle)

[Adam Megacz]

Look, I'll keep this simple: AFS+Kerberos is up and working and ready
to support user storage and ssh logins to mire.  Right now.  Period.

The people who are complaining about the current delay being due to
AFS are the exact same people who are stonewalling my request to open
those accounts.

Does anybody notice a coincidence here?

  - a



Nathan Kennedy <ntk at hcoop.net> writes:
> Davor Ocelic wrote:
>> On Mon, Apr 30, 2007 at 03:54:56PM -0700, Adam Megacz wrote:
>>   
>>>>>> No, Adam, we had delays due to AFS that happened before you joined
>>>>>> as well. 
>>>>>>           
>>>>> Perhaps.  But if that is the case you should describe the explanation
>>>>> for the delay as "previous AFS administrator" not "AFS".
>>>>>         
>>>> No. The problem was not AFS itself, but the hacking of all kinds of 
>>>> things to work with AFS.
>>>>       
>>> Can you be more specific?
>>>     
>>
>>   
>>> For the record, Apache has been the only thing that has taken more
>>> than an hour or two to get working with AFS.
>>>
>>> This "blame it on AFS" thing has become a political scapegoat.
>>>     
>>
>> You are mixing some things up here. Noone is explaining or blaming
>> AFS for anything.
>>
>> I have just stated the fact.
> This is true.  I'm not being "I told you so" about this, but I had 
> barely heard of AFS before it was chosen for the new server, and I had 
> reservations given everyone's lack of substantial deployment of it, 
> although of course I deferred to those who were actually going to be 
> working with it.  I also am not blaming the technology, but it 
> definitely is not an out-of-the-box solution.  I've pulled many a hair 
> over NFS before, but at least I understand it pretty well.  The majority 
> of the time spent on Peer1 configuration has been related to AFS in some 
> way, whether it is Apache integration, kernel compiles, kerberos 
> wackiness, etc.  There was also one instance where we were getting 
> kernel errors that turned out to be a recently patched bug.
>
> I am glad it's coming together (and Adam M has done his part), and I 
> certainly hope we'll reap AFS's benefits, but if I were going to do this 
> deployment over I wouldl have used NFS.  (Well, if we were doing it over 
> we could get AFS working very quickly, but you know what I mean).
> -----
> Now, as to those members who are now claiming that the board has pulled 
> one over the membership with this "enforced" and "overengineered" Peer1 
> migration, I have to assume you haven't been tuned in.  Number one, 
> every board member ran both this year and last year on the platform of 
> rolling out a better setup at a better hosting provider.  Any member was 
> free to voice their concerns about this, or run for a director seat 
> under the "Stop the Migration" party.  Nobody did.
>
> Secondly, the process of choosing a hosting provider was as transparent 
> as it could possibly be.  Everything was discussed on the mailing list, 
> on open and announced IRC meetings (logs are still available at 
> hcoop.net/board/ if you don't believe me), with candidates suggested, 
> listed, and culled on the wiki.  The board and others involved with the 
> process did all we could to maximize member participation.  So please 
> don't come by at this late date, after we've all put so much effort, 
> time, and money into this, and say that you want no part in it.  Either 
> be a part of the cooperative or don't.
>
> I am guessing that by this time next year, things will have been working 
> so well at Peer1 and with all the new members who have joined, everyone 
> will have forgotten this whole controversy and taken it for granted.
>
> -ntk

-- 
PGP/GPG: 5C9F F366 C9CF 2145 E770  B1B8 EFB1 462D A146 C380