Return to BSD News archive
Xref: sserve comp.sys.sun.misc:11608 comp.sys.sun.admin:24302 alt.sys.sun:8493 comp.unix.bsd:13264 comp.unix.admin:16440 Newsgroups: comp.sys.sun.misc,comp.sys.sun.admin,alt.sys.sun,comp.unix.bsd,comp.unix.admin Path: sserve!newshost.anu.edu.au!munnari.oz.au!constellation!news.uoknor.edu!ns1.nodak.edu!netnews.nwnet.net!news.uoregon.edu!cs.uoregon.edu!sgiblab!uhog.mit.edu!europa.eng.gtefsd.com!uunet!mnemosyne.cs.du.edu!nyx!croten From: croten@nyx.cs.du.edu (Charles Roten) Subject: Bring me the head of /var/yp/YP.noname ! :-) :-( Message-ID: <1994Jan13.225644.6459@mnemosyne.cs.du.edu> Summary: _What_ is blowing this directory away on boot up ?!? Heeeelp !! X-Disclaimer: Nyx is a public access Unix system run by the University of Denver for the Denver community. The University has neither control over nor responsibility for the opinions of users. Keywords: Arrrgh !! !*%$@+!!!! Sender: usenet@mnemosyne.cs.du.edu (netnews admin account) Organization: Nyx, Public Access Unix at U. of Denver Math/CS dept. Date: Thu, 13 Jan 94 22:56:44 GMT Lines: 48 Late last week, a friend of mine with a Sun 386i/150 sent up a distress rocket. It seems that his wife had been doing some work on his machine, as root. After this, the machine would not reboot in multi-user mode. She did report she had mistakenly used the 'hostname' command and had inadvertantly changed the system name. But the file /etc/hostname.le0 did not exist on his (standalone) system, and I could not detect any trace of this change during my later maintainace work on his system disk. BTW, his machine (like my own Sun 386i/250) has NIS (Yellow Pages) set up during the boot process (the default with Sun 386is), but is configured as a standalone. Yeah, I know, NIS should be disabled on both our boxes .. sigh. :-( _Believe_ me, that is priority one at this point. After the odd red herring or two, it finally turned out that the NIS daemon ypbind was choking during the boot, because ypserv did not get loaded, because the /var/yp/YP.noname directory had been _blown_ _away_!! Thus, the boot failed to finish. Now during my work with his system, I had his system disk mounted, partition by partition, as an auxiliary disk on my box. So I just copied my /var/yp/YP.noname directory to his system disk. Yeah, I know, it goes in /files/var/localhost/yp, because of the way Sun set up the file system .. sigh. Then I shut down, and swapped his system disk for mine. The following reboot went OK. So I returned his disk to his system. _That_ reboot went OK. But after a later Stop-a shutdown, the /var/yp/YP.noname directory was somehow blown away _again_! This does _not_ seem to me to be a coincidence. I have recreated the directory, this time using mkdir instead of copying from my own system disk. Then, using his system disk as the system disk in my box, I ran a make in /var/yp. I then replicated the sequence of events which lead to the second destruction of /var/yp/YP.noname. The system rebooted normally afterwards. /var/yp/YP.noname remained intact. But until I understand _precisely_ what went wrong, I will be unsure of the stability of my friend's system. Can some kind soul enlighten me as to what went wrong here ?? If you reply by email, please do so to croten@eosdata.gsfc.nasa.gov, rather than to my nyx account. Thanks in advance. -- Charles D. Roten | Hughes STX Inc. croten@nyx.cs.du.edu | NASA GSFC (Hurrah DAAC!) croten@eosdata.gsfc.nasa.gov | (301) 286-4413 (w), (301) 317-0782 (h)