Message-Id: <199709111552.IAA09187@mailhost2.cac.washington.edu> From: "James W DeRoest" <deroest@cac.washington.edu> To: "AST Webpage" <astweb@u.washington.edu> Subject: SMB authentication meeting minutes, 9/8 Date: Thu, 11 Sep 1997 08:51:58 -0700
-----Original Message----- From: Yonah Karp <yonah@u.washington.edu> Date: Wednesday, September 10, 1997 5:30 PM Subject: SMB authentication meeting minutes, 9/8
Minutes are below. Please inform the group of updates and current issues and needs so that we can give input (and help) as needed.
We should meet again next week. I'm seeing free blocks for most of you for 10:00 or 10:30 or 4:00 Thursday 9/18. Please let me know individually what times will work for you.
Thanks.
Yonah
====
09/08/97 -- SMB authentication
Present: Adam Ith, Brad Greer, David Cox, Steve Jones, Ryan Campbell, Mark McNair, Yonah Karp, Chad Warren, Oren Sreebny, Betsy Williams
Status:
Mac client (using ftp) is working OK, using the setup on simba02 and authorizing via the Kerberos database. Logging out is an issue, but it's a social engineering rather than a technical issue (users aren't used to logging out).
CS will put on 70 Mac's by the end of the week to exercise scaling needs. There will be ~150 Mac's which will use it once we are in full production:
70 CRC 50 Commons 20-30 ACC
PC client testing is also working. 2 are in low use currently. To exercise scaling problems, CS will put on 30 ACC PC's by the end of the week. There will be ~300 PC's which will use it once we are in full production:
60 CRC 200 Commons 30 ACC
Kelp is a local server in the CRC (an NT box). Kelp needs more testing. CS will do this.
Steve Jones will change the name of NT login box (currently "AC Lab") to "UW Lab".
We discussed the issue of people who don't have access but need to run "new" (or to print). Ideas included "guest" accounts, PC's set up in an NT domain with a different policy, Kerberos principals with access to no UA accounts. CS (David Cox & Ryan) will decide what works best and let us know what they've decided.
Everyone logging in at once in the morning in the labs will be a fairly good test of load.
We talked about logging issues. How many logins/ unit time, people/ unit time, time between logins were measures of interest. (We can't get average session time because there is no "logout event.") Steve is keeping a wtmp log file. He will work with David Cox on this issue.
Oren mentioned the need to be consistent with what we keep (with green, bank and other NDC machines; UA data on anthony; and in the labs).
If the Kerberos server goes down, then lab personnel should call Operations and have them contact unix_guru. Both operations and the gurus need to know this is a critical service. Also, the Wins server could be able to talk to "secondary" with a simple approach -- using a list of IP #'s, which is gotten at each login. Ryan/ Mark/ David will make this happen and test it.
Steve will get bud into service and doing what simba02 is currently doing and CS will start using bud rather than simba02. [This has already happened as of 09/10.]
Steve will work on getting a second "bud" into service.