Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!ukma!gatech!bloom-beacon!spdcc!dyer From: dyer@spdcc.COM (Steve Dyer) Newsgroups: comp.sys.ibm.pc.rt,comp.unix.i386 Subject: Re: Why minidisks? Message-ID: <3794@ursa-major.SPDCC.COM> Date: 20 Jul 89 15:58:49 GMT References: <672@msa3b.UUCP> Reply-To: dyer@ursa-major.spdcc.COM (Steve Dyer) Organization: S.P. Dyer Computer Consulting, Cambridge MA Lines: 38 In article <672@msa3b.UUCP> kevin@msa3b.UUCP (Kevin P. Kleinfelter) writes: >Why does AIX partition my lovely 300MB disk into "minidisks?" >I understand the need for a paging minidisk (partition) and a dump >minidisk (partition). Also, I can make a good argument for a >separate root minidisk. The thing I don't understand is why have > / > /u > /machinename > /machinename/tmp > /whatever-else-you-create-to-fill-rest-of-300MB >Is there any reason why I should not just merge all minidisks except >for root, paging, and dump into one "/usr"? >(P.S. If it matters, this is AIX PS/2 1.1.) AIX RT doesn't require minidisks of the form /machinename. AIX PS/2 does, and both / and /machinename are implicitly mounted at boot time. This is undoubtedly an attribute of the yet-to-be-released TCF facility in AIX PS/2 and AIX 370, wherein a collection of PS/2s and/or 370's all share the same extended filesystem, including a common root. Files formerly found in /etc which are "private" to a particular machine have been moved to /machinename. On most UNIX systems these days, /tmp is a smallish partition of its own, and under TCF, placing it under /machinename/tmp makes sense, since having a global shared /tmp is risky, given that most UNIX systems presume that /tmp is private (e.g., composite filenames consisting of a string+PID will no longer be unique.) /tmp is a symlink to /machinename/tmp. Similarly, the per-machine files which used to reside on /etc (and elsewhere) are now symlinks to files on /machinename. TCF hasn't been released yet, but much of the code (and all of the design) to support it is already present in AIX PS/2 1.1. -- Steve Dyer dyer@ursa-major.spdcc.com aka {ima,harvard,rayssd,linus,m2c}!spdcc!dyer dyer@arktouros.mit.edu