Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!wuarchive!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!aplcen!unmvax!nmt.edu!nraoaoc From: nraoaoc@nmt.edu (NRAO Array Operations Center) Newsgroups: comp.sys.dec Subject: Re: Can I safely remove a SCSI device from a running DEC3100 Message-ID: <1990Aug10.232210.10390@nmt.edu> Date: 10 Aug 90 23:22:10 GMT References: <1990Aug10.173348.4125@spock.UUCP> Sender: rmilner@zia.aoc.nrao.edu (Ruth Milner) Reply-To: rmilner@zia.aoc.nrao.edu (Ruth Milner) Organization: National Radio Astronomy Observatory, Socorro NM Lines: 25 In article <1990Aug10.173348.4125@spock.UUCP> lancelot@spock.UUCP (Thor Lancelot Simon) writes: >We swap TK50s around all the time here - it works fine as long as the machine >you are putting it on had a tape on it at boot time or it won't be recognized. I didn't think TK50's were SCSI devices, though I don't know about on 3100's. What you describe sounds awfully like the way VMS treats them. On Suns I have found that you *can* disconnect SCSI devices so long as: a) nobody tries to access any SCSI device on the bus while you are working, b) you have an easy way of moving the terminator if necessary The second point is important, since you don't want to leave the SCSI bus unterminated. External or inline terminators make this easy, or if the device to be removed is in the middle of a daisy chain. If the disks your system serves are SCSI, though, forget it. The only way you can guarantee point a) is to unexport them all, in which case you might as well just shut the system down anyway. Disclaimer: no promises, it has worked for me. -- Ruth Milner Systems Manager NRAO/VLA Socorro NM rmilner@zia.aoc.nrao.edu