> Here we have another, horrifying example of NetInfo sucking in a major > fashion. 10-15 hours to boot your machine. 18 hours to reload your > passwd database -- updating a 24,000 user flat file on 40 machines > takes about two orders of magnitude less time!
Well Mark, if you hate NS so much, why do you still use it, just for the pleasure of complaining?
In article <1995Jan27.122651.1...@rdbois.uucp> s...@rdbois.fdn.org
(serge_ruby) writes: > > Here we have another, horrifying example of NetInfo sucking in a major > > fashion. 10-15 hours to boot your machine. 18 hours to reload your > > passwd database -- updating a 24,000 user flat file on 40 machines > > takes about two orders of magnitude less time!
> Well Mark, if you hate NS so much, why do you still use it, just for the > pleasure of complaining?
No. The university purchased NeXTs for the dorms, and Mark is the fellow who got the job of administering them, along with the Macs and other dorm computers. It wasn't his decision to use NextStep. The university has now dumped the NeXTs in the main computing area, but there are still a bunch of them left in the dorms. As I see it, Mark is trying to do his job the best he can and is running into NextStep specific problems which are quite understandably annoying.
>>Here we have another, horrifying example of NetInfo sucking in a major >>fashion. 10-15 hours to boot your machine. 18 hours to reload your >>passwd database -- updating a 24,000 user flat file on 40 machines >>takes about two orders of magnitude less time!
>Well Mark, if you hate NS so much, why do you still use it, just for the >pleasure of complaining?
Mark doesn't hate NS any more than I do. He's only pushing what he things is the best technology route for NeXT to take on some (several) issues, one of which is netinfo. Mark has been around for a while, and his opinions are very good to hear (experience does mean something, after all). Too often in the rush to make new and cool things people forget, or dismiss opinions that come from experience, often because they seem unreasonable, or invalid, without the benefit of that experience to be one's guide.
So, when Mark complains,he's only doing it because he cares about NEXTSTEP and how well it works, and thinks that it can become better. Now, you should be able to go home feeling all warm and fuzzy inside ;-).
- darcy
(For the record, I don't hate NS either! I looove it. Sometimes it makes me frustrated, like it's a child that's happy getting 85% on a report card, when it's fully capable of bringing home 95%). -- (prog (senseFood (prog (prog (senseFood move eat) eat (senseFood move eat)) rotRight eat) (prog move (prog (prog eat (senseFood move rotRight) eat) loop) rotRight)) (prog (senseFood (senseFood (prog (prog (senseFood move move) eat (senseFood move eat)) rotRight eat) (senseFood (prog (prog (senseFood move
On Fri, 27 Jan 1995, serge_ruby wrote: > In article > <Pine.NXT.3.92.950126200840.12093A-100...@Tomobiki-Cho.CAC.Washington.EDU> Mark > Crispin <m...@CAC.Washington.EDU> writes: > > Well, well, well.
> > Here we have another, horrifying example of NetInfo sucking in a major > > fashion. 10-15 hours to boot your machine. 18 hours to reload your > > passwd database -- updating a 24,000 user flat file on 40 machines > > takes about two orders of magnitude less time!
> Well Mark, if you hate NS so much, why do you still use it, just for the > pleasure of complaining?
I hate NetInfo, you mindless idiot, ***NETINFO***.
Look at the little squiggles: NetInfo does not look all that much like NEXTSTEP
> > Here we have another, horrifying example of NetInfo sucking in > > a major fashion. 10-15 hours to boot your machine. 18 hours > > to reload your passwd database -- updating a 24,000 user flat > > file on 40 machines takes about two orders of magnitude less > > time!
> Well Mark, if you hate NS so much, why do you still use it, just > for the pleasure of complaining?
Don't confuse Mark with Mike Dahmus. Mark actually tends to know what he's talking about. Mark also doesn't hate NeXTSTEP, he complains about obnoxious problems because he thinks they are obnoxious. Certainly the description of the above problem (which is a real problem that happened to a real NeXTSTEP user) qualifies as being obnoxious.
Now, the above problem might also be specific to release 3.0 of NeXTSTEP, and things might not be anywhere near as bad in the latest release.
-- Garance Alistair Drosehn = g...@eclipse.its.rpi.edu ITS Systems Programmer (handles NeXT-type mail) Rensselaer Polytechnic Institute; Troy NY USA
Netinfo is far from perfect and has real performance problems with scalability, it seems to me.
However, it should be pointed out that all the serious problems Mark has mentioned so far (that I have noticed) do not really bring down small clusters. In many (most) situations, NetInfo works pretty well. On the plus side, I think NetInfo allows one to setup the system (esp. networkside) pretty quickly even if you don't have unix experience. I have seen this happen a number of times. For most newcomers to Unix-ish systems, the barrier for administration (especially network) is formidably high, or seems so, and Netinfo is great for these guys. These guys won't be administering big networks anyway (well I hope not!). I have talked to a number of academics who want systems networked quickly and well but don't want to be bothered by it. (and don't have the $$ to hire pro administrator.) These guys are very happy using NetInfo and have never encountered (noticed?) the aforementioned performance problems with NetInfo.
Btw, I have never encountered serious problems with NetInfo since I have administered only a small NeXT cluster. These days I administer HP's, a NeXT, linux and sparcs (again a small number, less than 10) and to tell you the truth, I sometimes miss NetInfo.
What should be done --- and I think (hope?) that this is what Mark is clamoring for --- is to have the option of completely circumventing NetInfo if one wants. The choice thing.
This way, I think everyone will be happy.
Happy NeXTing. -- Kenichiro Aoki (k...@phys.titech.ac.jp), Department of Physics, Tokyo Institute of Technology, Oh-okayama, Meguro-ku, Tokyo, JAPAN ... posting from kyoto....
> What should be done --- and I think (hope?) that this is what > Mark is clamoring for --- is to have the option of completely > circumventing NetInfo if one wants. The choice thing.
Yup, that's exactly what I want. Which means that the DNS has to work even if NetInfo is disabled. Which means that fstab has to work in the standard Unix way even if NetInfo is disabled.
This used to work in 0.9, but after 1.0, NetInfo became mandatory.
