lyons@loligo.cc.fsu.edu (Jim Lyons) (09/01/89)
We have been having a problem at our site with the automatic log on feature of AppleShare. The problem is that when a user tries to set automatic log on from the volume selection dialog box in Chooser, an alert says, "The AppleShare Prep file needed some minor repairs. Some startup information may be lost." The file it then creates is garbage compared to a prep file that works; the former just has a dozen or so (non-text) bytes and the latter has some sort of structure which includes plain-text strings for the user name and password, the zone(s), and volume name(s). Where did we get a prep file that works, you ask? Well, we have have Mac IIs which all have Apple's EtherTalk board and Mac Pluses on LocalTalk; a Kinetics FastPath bridges the two. Turns out this problem was only occurring on the IIs; on the Plusses, it works fine. (We are using System 6.0.3, EtherTalk 1.2, and AppleShare 2.0.1 on all machines.) Hold on, it gets wierder. Further investigation showed that if a Mac II was switched over to LocalTalk, AppleShare then created a good prep file! We also found you could create a good prep file on a Plus and then copy it into the System Folder of a II on the Ethernet, and it would work correctly; however, if you tried to change the startup information, the prep file would be garbaged again. The problems apparantly started when we reconfigured the server volumes. We noticed that after the change, users who had been using the automatic log on feature with just the name being saved would get the dialog box during boot-up asking for their password; they would enter it, but after boot-up, no server volume would be mounted. In looking at these prep files we discovered, interestingly, that the prep file kept information about previous configurations of the server; we saw in them names of test servers we had set up briefly, and volumes that had existed some time in the past but were no longer present. We have reported this problem to Apple Tech Support via our local Apple S.E. If I find out what's going on, I'll post it. In the meantime, if anyone can corroborate this problem or has any information or ideas about this, please let me know or post it to the net. Jim Lyons \ Supercomputer Computations Research Institute 904/644-4274 \ Florida State University lyons@gw.scri.fsu.edu \ Tallahassee, FL 32306-4052