my odyssey with openafs / kerberos for windows continues.

recall from last time, i had managed to actually get openafs 1.4.0 + KfW 3.0.0 working. nominally... more working than not, at least. 95% working, say.

now, also recalling from last time, i mentioned a strange issue: after you install KfW / openafs for windows, for some reason, you must sit at the novell login window before trying to log in for approximately 3 minutes or so, or else the novell client will complain that "the tree or server cannot be found".

do others out there who are using novell with openafs encounter this problem, of having to wait at the novell login prompt?

now also, this only seems to occur the first time a user tries to log on after the machine has been booted up. say you cold start your computer, get to the novell prompt, wait as long as is necessary, and then log in successfully. you can log out, and then log in again without having to wait at the prompt the second time around! why?

so my questions are, basically

(1) why is there the 3-minute pause between when "tree or server not found" and you can actually log in successfully? is this pause a necessary compromise when using openafs with novell? is there any way to mitigate this? it sounds hokey to go to a user and say "well, you need to wait three or four minutes before trying to log in just because  there is some issue with all this software" when they are not having this problem with their existing (transarc?) afs 1.2.x clients..

(2) does anyone have any good techniques for diagnosing these sort of pre-logon problems with windows? are there some log files that i could peek at that i'm not aware of? event viewer is doing nothing for me.

i really want to try and reduce the waiting that the user must do at the login prompt because i feel that it is fairly unreasonable to expect people to "wait five minutes" arbitrarily and it will probably generate a storm of user complaints, resistance, and so forth

i'm not expecting much in the way of responses but i'll take any suggestions i can get -- this sort of deeper program debugging is a little outside of my typical experience.

thanks for the continued tolerance,


sean caron
associate systems administrator
university of michigan school of public health
[EMAIL PROTECTED]
1-734-763-4206


Reply via email to