From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Wed, 27 Mar 1996 22:56:43 -0500 From: Thomas Riemer triemer@babbitt.bernstein.com Subject: CPU server again... Topicbox-Message-UUID: 419b80ae-eac8-11e9-9e20-41e7f4b1d025 Message-ID: <19960328035643.t8jmrHMOqINQy7x6d_vXU7hoqBKznXVc0GW8kQ4qXpc@z> Ok, with a lot more investigation, I'm starting to get an inkling as to what is going on. I've been booting the CPU server via IL. I'm not sure how it figures out that its supposed to boot /386/9cpu instead of /386/9pc, but anyway, it has been mounting its root directory from the file server. It has not been bothering to bind the hard drive devices. This means that even though I have two disks in place /dev/hd0disk, /dev/hd1disk simply don't exist, include /dev/nvram ... Essentially anything that has to do with the local hard drives. Small surprize that you can't read the /dev/nvram device if it doesn't exist! (I don't quite believe that I didn't notice this earlier...) Now, I know that when I boot up a standalone 9pc kernel... The disk partitions exist for nvram. Let me note that the cpu kernel (9cpu presumeably) is the one off the CD-ROM dist. I looked at the manual stuff about "rootdir" (plan9.ini man page) - is this the case where one would actually use the "rootdir" option? Again - this CPU server is a PC. Let me ask the dumb question - has anyone else run a CPU server on IDE disks only (on a PC - whereelse does one find ide)? I'm thinking, may the CPU code is only aware of scsi disks. I should also note that the like, bind -a '#H0' /dev seems to do nothing... I don't happen to have a spare scsi in the apartment at the moment, otherwise I'd try booting off the scsi drive... Any ideas would be much appreciated. Ah, yes... for those interested in fileservers... The fs that I finally got running was the fs compiled off the CD-ROM. The fs's distributed on plan9.att.com and the binary produced by the installation program - didn't cut it. Granted I made the one little change to lower the number of files as mentioned here on the list - and maybe that was all it was. I'm starting to think the CPU authentication never worked at all... and somehow I faked myself out. -Tom BTW - its been a long, long time since I struggled with an operating system like this... last time was an AT&T 3B2 - my first Unix box. Come to think of it... it was always complaining about NVRAM too. mmm... haven't booted that puppy up recently. ---- Where theory and reality meet. ---- Thomas Riemer, triemer@wesleyan.edu