Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!iuvax!purdue!decwrl!shelby!portia!forel!karish From: karish@forel.stanford.edu (Chuck Karish) Newsgroups: comp.sys.ibm.pc.rt Subject: Re: RT Baseband (Ethernet) adapter Keywords: TCP/IP network rt pc x windows Message-ID: <3824@portia.Stanford.EDU> Date: 21 Jul 89 05:13:54 GMT References: <12775@well.UUCP> Sender: USENET News System Reply-To: karish@forel.stanford.edu (Chuck Karish) Distribution: comp Organization: When necessary Lines: 86 In article <12775@well.UUCP> gors@well.UUCP (Gordon Stewart) wrote: >One particular problem is that, when attempting to "ping" or otherwise >make network contact with another host that is not on the net, I get >a diagnostic message: > > net0: transmit timeout -- device shutdown !@#$!@!!!! > >This is plain annoying! After this happens, the easiest thing to do >is reboot! The network is "unreachable" after this! Did you take the sample token ring entry out of /etc/net? I suppose you know by now, but I doubt that your baseband adapter's jumpers are the whole problem. I don't think you'd be able to talk to the net at all with default settings on both the card and in the software. The card comes jumpered to use memory starting at 98000 hex; the default in /etc/ddi (used by devices) is 80000. >I would love to hear from anyone regarding: > > - TCP/IP services on the RT (running at the same time as X11) They get along fine here. I don't know anything about XONDOS. >An seemingly unrelated but vexing problem is that, upon rebooting (after >a shutdown, of course) the f***ing login: doesn't recognize any login >names (including 'root'). The only thing to do is reboot again, interrupt, >which puts us in maintenance mode, and exit -- strangely, when multi-user >mode is started again, it recognizes my login !!!! hooray! but what is >happening?? What's in your /etc/.init.state file? It should be one character (and a newline), probably `a' for `automatic multi-user' or `m' for `maintenance' (single-user mode). This letter gets passed as an argument to /etc/rc to determine the run level on boot. Documented in chapter 3 of the Operating System Technical Reference (assuming you're running an RT). >Anyone else regret going with AIX?? Not me. Our systems are fast and reliable. The Berkeley stuff that's been integrated into the system is really a part of it, not a pile of scabs and patches, and it's true to the way the stuff works on BSD-based systems (except for symbolic links, but I don't really want to get into that now... ). *-*-*-*-* Things to do to get a network node going: - Edit /etc/hosts. You don't need a `localhost' entry for loopback; the local system is just another name in the list. - Edit /etc/hosts.equiv. - Edit /etc/rc.tcpip (to set the host name at boot time) - Edit /etc/rc to uncomment the invocation of /etc/rc.tcpip - Edit /etc/net. Comment out or remove the token ing entry unless you're running token ring. - Run `devices' to add net0 (driver for the baseband adapter). - Use the `devices' command to make ptys, some with gettys and some without. If you need more than 16 ptys, bump the value of `ptybuffers' in /etc/master and make a new kernel, and edit /etc/ddi/pty to make more blank entries for use by devices. IBM people say they're going to provide a batch program to create devices; for now, it's all interactive and somewhat tedious. - To run X on an RT, you need at least 8 megs of memory. For running X, there are some resource limits in the kernel that should be bumped upward. The ones I changed were ptybuffers (from 16 to 32) charlists (64 to160) maxminor (in the `uptc' and `upts' stanzas; 16 to 32) procs (in `sysparms'; 60 to 90) filetab (250 to 350) inodetab (250 to 350) Chuck Karish {decwrl,hpda}!mindcrf!karish (415) 493-7277 karish@forel.stanford.edu