From: Digestifier To: Linux-Admin@senator-bedfellow.mit.edu Reply-To: Linux-Admin@senator-bedfellow.mit.edu Date: Mon, 29 Nov 93 20:35:32 EST Subject: Linux-Admin Digest #193 Linux-Admin Digest #193, Volume #1 Mon, 29 Nov 93 20:35:32 EST Contents: Patches to lilo 0.12 to avoid infinite recursion (Norman Ramsey) Re: /etc/securetty? (su from tty) (Christopher Shaulis) Re: trouble building kernel 0.99 pl 13 with CD-ROM fs (Christopher Shaulis) getty_ps: uugetty w/scheduling... (Gary Moyer) cat /dev/null > /dev/rmt0 dangerous? (Joern Jensen) Re: cat /dev/null > /dev/rmt0 dangerous? (K J MacDonald) Re: Watcher... (Daniel Z. Davidson) Re: /etc/securetty? (su from tty) (Matthew Donadio) Working NIS anywhere? (Kjetil Torgrim Homme) Re: Difference between cua and ttyS (Intellection) Disk troubles (Shekhar Mahadevan) Re: getty_ps: uugetty w/scheduling... (John Paul Morrison) Can't telnet to my machine (Michael C Landauer) Re: Can't telnet to my machine (David Barr) ---------------------------------------------------------------------------- From: norman@flaubert.bellcore.com (Norman Ramsey) Subject: Patches to lilo 0.12 to avoid infinite recursion Date: Mon, 29 Nov 1993 06:24:31 GMT I found a simple cure, namely not to follow symbolic links. There may be other problems lurking elsewhere, your mileage may vary, etc, but this beats the /dev/fd -> /proc/self/fd problem. *** device.c.dist Mon Nov 29 00:54:06 1993 --- device.c Mon Nov 29 01:21:07 1993 *************** *** 31,38 **** *(start = strchr(parent,0)) = '/'; while (dir = readdir(dp)) { strcpy(start+1,dir->d_name); ! if (stat(parent,&dev->st) < 0) ! die("stat %s: %s",parent,strerror(errno)); if (S_ISBLK(dev->st.st_mode) && dev->st.st_rdev == number) return 1; if (S_ISDIR(dev->st.st_mode) && strcmp(dir->d_name,".") && strcmp(dir->d_name,"..") && scan_dir(dev,parent,number)) return 1; --- 31,38 ---- *(start = strchr(parent,0)) = '/'; while (dir = readdir(dp)) { strcpy(start+1,dir->d_name); ! if (lstat(parent,&dev->st) < 0) ! die("lstat %s: %s",parent,strerror(errno)); if (S_ISBLK(dev->st.st_mode) && dev->st.st_rdev == number) return 1; if (S_ISDIR(dev->st.st_mode) && strcmp(dir->d_name,".") && strcmp(dir->d_name,"..") && scan_dir(dev,parent,number)) return 1; *** geometry.c.dist Mon Nov 29 01:17:11 1993 --- geometry.c Mon Nov 29 01:20:53 1993 *************** *** 82,88 **** die("opendir %s: %s",parent,strerror(errno)); while (dir = readdir(dp)) { sprintf(name,"%s/%s",parent,dir->d_name); ! if (stat(name,&st) < 0) die("stat %s: %s",name,strerror(errno)); if (S_ISBLK(st.st_mode) && MAJOR(st.st_rdev) == major && (MINOR(st.st_rdev) & (increment-1)) == 0) { this = MINOR(st.st_rdev)/increment+1; --- 82,88 ---- die("opendir %s: %s",parent,strerror(errno)); while (dir = readdir(dp)) { sprintf(name,"%s/%s",parent,dir->d_name); ! if (lstat(name,&st) < 0) die("lstat %s: %s",name,strerror(errno)); if (S_ISBLK(st.st_mode) && MAJOR(st.st_rdev) == major && (MINOR(st.st_rdev) & (increment-1)) == 0) { this = MINOR(st.st_rdev)/increment+1; ------------------------------ From: cjs@netcom.com (Christopher Shaulis) Subject: Re: /etc/securetty? (su from tty) Date: Mon, 29 Nov 1993 07:19:46 GMT philp@universe.digex.net (Phil Perucci) writes: >How do you set permissions so you can only "su root" from a >local console? I DON'T want remote logins to be able to >switch to root. I thought it was /etc/securetty... >I removed all entries except "console" from /etc/securetty, but >remote logins can still su to root. SOB! That does work.. I would have figured some interaction with /etc/securetty myself. If I were in your shoes then my thinking would probably go along the lines of "in a secure enviroment users, with the exception of adminstration, have no business changing IDs without going through login/out procedures".. and I'd change su's owner, group, and permissions accordingly. That or I'd make use of all this wonderful source code to add /etc/securetty checking and post a patch to the net. :) Christopher cjs@netcom.com ------------------------------ From: cjs@netcom.com (Christopher Shaulis) Subject: Re: trouble building kernel 0.99 pl 13 with CD-ROM fs Date: Mon, 29 Nov 1993 07:34:46 GMT norman@flaubert.bellcore.com (Norman Ramsey) writes: >I downloaded and successfully compiled the kernel release 0.99 >patchlevel 13, but the kernel doesn't have the iso9660 CD-ROM file >system even though I asked for that at `make config' time, and the >file .config contains the following line: >CONFIG_ISO9660_FS = CONFIG_ISO9660_FS Did you make the kernel once before without ISO 9660 support? Did you install the new sources over the old ones? Try doing a 'make clean'. Then do a make depend, config, and zimage in that order. My thinking is that you've got old stuff from the previous kernel sitting around. >I was also surprised that the kernel seemed to do an fsck of the root >file system at boot time; I brought down the system cleanly, and the >boot disk I made from the Yggdrasil distribution doesn't do such >checking. I assume that ROOT=CURRENT in the makefile means use >whatever the root file system is now (in my case /dev/sda1), but >perhaps it means something else? When your root file system is ext2fs, it doesn't appear to be marked clean at shut-down. When I used the Yggdrasil CD as root they always were but as soon as I switched it to the hard drive, I started having to sit through efscks every time I booted up. My solution was to remove the commands from /etc/rc because there was obviously nothing wrong. Go ahead and set your ROOT= to /dev/sda1 or whatever you like. Make sure you get what you want. Christopher cjs@netcom.com ------------------------------ From: gary@mojo.uu.pdx.edu (Gary Moyer) Subject: getty_ps: uugetty w/scheduling... Date: 29 Nov 1993 00:31:04 -0800 I've been (vainly) trying to get uugetty to work with scheduling with no luck whatsoever. I've grabbed the latest sources off sunsite (2.07b) and compiled it. I'm using the supplied /etc/default/uugetty. file (w/ tweaks only to the time schedule). With the uugetty running I see activity on the (via the panel) but answer never goes on. Here's the script (w/ comments removed): /etc/default/uugetty.cua2: ALTLOCK=cua2 SCHED=0:2:17-0:8:00 1:0:00-1:8:00 2:0:00-2:8:00 3:0:00-3:8:00 4:0:00-4:8:00 5:00:00-5:8:00 6:18:45-6:18:50 OFF="" \d+++\dAT\r OK\r\n ATH0\r OK\r\n AT\sM0\sE1\sQ0\sV1\sX4\sS0=0\r OK\r\n TIMEOUT=60 INIT="" \d+++\dAT\r OK\r\n ATH0\r OK\r\n AT\sM0\sE1\sQ0\sV1\sX4\sS0=1\r OK\r\n DELAY=1 Oh yeah, modem is a Practical Peripherals 9600 Hayes compatible. Has anyone got this working ? Thanks! -g.m. -- =============================================================================== Gary Moyer gary@mojo.uu.pdx.edu, gary@rodney.cs.pdx.edu [LINUX] gary@acacia.cs.pdx.edu [386BSD] moyerg@cs.pdx.edu "Linux: more than an OS -- it's a social movement" ------------------------------ From: jornj@edb.tih.no (Joern Jensen) Subject: cat /dev/null > /dev/rmt0 dangerous? Date: Mon, 29 Nov 1993 08:58:21 GMT We have this scsci-tapedrive here, and I needed to rewind my tape automatically in a script. (I haven't compiled mt yet..) Now, when I use cat /dev/null > /dev/rmt0, will that do any harm to the data on the tape? (I guess it will, if it's in the middle of a f.x. a tar-file). Is there any better way to do this in an 'easy' way with the standard utils or commands? //jornj ------------------------------ From: kenny@festival.ed.ac.uk (K J MacDonald) Subject: Re: cat /dev/null > /dev/rmt0 dangerous? Date: Mon, 29 Nov 1993 09:25:30 GMT Joern Jensen (jornj@edb.tih.no) wrote: : We have this scsci-tapedrive here, and I needed to rewind my tape : automatically in a script. (I haven't compiled mt yet..) : Now, when I use cat /dev/null > /dev/rmt0, will that do any harm to : the data on the tape? (I guess it will, if it's in the middle of a : f.x. a tar-file). : Is there any better way to do this in an 'easy' way with the standard : utils or commands? The 'mt' command has a whole collection of useful things you can do to a tape, includeing rewind it. `mt -f /dev/rmt0 rewind` Hope this helps :) Kenny. -- ============================================================================== Kenneth MacDonald E-mail: kenny@ed Dept. of Geology & Geophysics "Allow me to introduce myself, Major Dennis University of Edinburgh Bloodnok, International Christmas Pudding ------------------------------ From: danield@nevada.edu (Daniel Z. Davidson) Subject: Re: Watcher... Date: Mon, 29 Nov 1993 10:42:43 GMT In article <2d37ro$qk2@jacobs.jacobs.mn.org> root@jacobs.mn.org (Mike Horwath) writes: >Why do you need to 'spy' on them? I have been Followong this thread, and I guess it's time to add my two cents worth. "Spying" is not the only use for such a program. When I ran a networked bbs and users had problems, they would often page me and ask for help in doing somehting with the BBS. I found it much easier for them to show me a problem (especialy whne It was a bug in my BBS software that I couldn't re-produce) than try and tell me about it. You may not want to 'spy' in them. However, sometimes it is useful to see what a person is doing so you can help them with a problem. >Can't a user have some privacy? Or you are so worried what the user is >doing that you need to watch so you can reset your modem if he does >something questionable? (oh wow, memory protection and permissions are >NOT enough, we have to 'watch' the user and keep him in check) Sure, the user can have privacy all they want. I have never been one to deney that to users. However, It is nice to be able to watch and help users when they ask for help. Daniel ------------------------------ From: donadio@mxd120.rh.psu.edu (Matthew Donadio) Subject: Re: /etc/securetty? (su from tty) Date: 29 Nov 1993 17:22:37 GMT Christopher Shaulis (cjs@netcom.com) wrote: : SOB! That does work.. I would have figured some interaction with : /etc/securetty myself. If I were in your shoes then my thinking would : probably go along the lines of "in a secure enviroment users, with the : exception of adminstration, have no business changing IDs without : going through login/out procedures".. and I'd change su's owner, group, : and permissions accordingly. That or I'd make use of all this wonderful : source code to add /etc/securetty checking and post a patch to the net. :) I _believe_ that su is behaving correctly. There are times when you might need to su remotely, such as when you hang console or X. There is a version of su floating around. that checks to see what groups you belong to and will not let you su to a user that is a memeber of group that you aren't. I think it is in with Rik Faith's util stuff. So you can make all trusted users and root members of this group, and if a non-trusted user knows the root password and tries to su root, then he will get an error message "Cannot su to groups you are not a member of." -- Beaker aka Matt Donadio | Life is short, --- __ o __~o __ o donadio@mxd120.rh.psu.edu | ride like ---- _`\<, _`\<, _`\<, --- Penn State Cycling ---| the wind. --- ( )/( ) ( )/( ) ( )/( ) ------------------------------ From: kjetilho@ifi.uio.no (Kjetil Torgrim Homme) Crossposted-To: comp.os.linux.development Subject: Working NIS anywhere? Date: 29 Nov 1993 19:03:51 +0100 We're trying to set up a cluster of PC running Linux here at the University. We started out with SLS 1.03, and have upgraded the kernel to 0.99.13s, libc to 4.4.4, and installed the binaries from the yp-linux package (the lib-source in yp-linux seems to have been folded into libc by now). gcc is 2.4.5. Okay, so things are working great for an hour or so. Then we get things like: > bash$ ypmatch kjetilho passwd > yp_match: clnt_call: RPC: Timed out > clntudp_create: RPC: Port mapper failure - RPC: Timed out NFS still works. ypbind is running, and gets answered every minute by the NIS-server. ypwhich works correctly. However, portmapper produces a lot of zombies (we've mostly tried /etc/portmap, not /etc/rpc.portmap, but the latter has not been any better). I've tried to kill all net-related processes, rm /var/yp/binding/* and restarting them (ie. portmap, inetd, named, ypbind - almost anything except init and getty :-) I've looked at NYS, but it isn't near a finished state. Peter Erikson suggested I replaced yp-linux code (partially) with NYS, but this isn't trivial, so that would be my absolutely last option. So... Are anyone using NIS on their Linux-machines? Any clues to offer? Kjetil T. ------------------------------ From: intell@metronet.com (Intellection) Subject: Re: Difference between cua and ttyS Date: Mon, 29 Nov 1993 18:35:53 GMT root@jacobs.mn.org (Mike Horwath) writes: >ttyS? and cua? are different like this: >ttyS? block on an open call if carrier detect is low, meaning that the modem >is in an idle state and no one is attached. >cua? will not block and is the device to use when you wish to dial out, like >in UUCP using uucico, or kermit, or minicomm. There are two additional facts which confuse people. First, this behavior of the terminal driver is new. Some people may be running the old terminal driver and not getting what you describe. Second, some programs are written to code around the ttyS? block by opening O_NBLOCK. I believe from what I've read that the old getty did this. Therefore even somebody with the new terminal driver but old serial line software may not be getting what you describe. Moral: Anybody who is running into conflicts on the serial line ought to be sure they have the latest versions of software. Anybody who is writing for the serial line ought to be sure they are using the new driver as it's intended to be used. No finger pointing, just a reminder. Regards, Ed ------------------------------ From: mahadev@Erc.MsState.Edu (Shekhar Mahadevan) Subject: Disk troubles Date: 29 Nov 93 13:53:10 I am running Linux 0.99pl13r on my 486. A few days ago, my filesystem started to misbehave. Whatever space was freed was not becoming available for use. ie. No matter how much space I free, I don't have a block to write on. I tried to fix this problem using e2fsck (I even tried the latest version available from tytso@mit.edu.) No luck. Can somebody tell me how to fix this problem without having to go to extreme measures like backing up and/or reinstalling? Thanks Shekhar -- ------------------------------ From: jmorriso@rflab.ee.ubc.ca (John Paul Morrison) Subject: Re: getty_ps: uugetty w/scheduling... Date: 29 Nov 1993 21:26:03 GMT In article <1993Nov28.104754.9143@mojo.uu.pdx.edu>, Gary Moyer wrote: >I've been (vainly) trying to get uugetty to work with scheduling with no >luck whatsoever. I've grabbed the latest sources off sunsite (2.07b) and >compiled it. > >I'm using the supplied /etc/default/uugetty. file (w/ tweaks only >to the time schedule). With the uugetty running I see activity on the >(via the panel) but answer never goes on. Here's the script (w/ comments >removed): Quite simple, the getty scheduling is broken. I sent patches to the author several months ago, but I don't think there have been any updates to gettyps in a long time. (last month I had to make more changes, daylight savings!). Scheduling with gettyps will work if you live in England ;-) I think if you set the schedule to GMT it will work. You might try this patch: *** /tmp/getty_ps/main.c Tue Apr 20 09:51:44 1993 --- main.c Fri Nov 12 18:30:03 1993 *************** *** 168,178 **** { time_t t_cur, t_base, t_begin, t_end; int s_dow, s_hr, s_min, e_dow, e_hr, e_min, count; /* set up the time base */ ! (void) time(&t_cur); t_base = t_cur - ((t_cur + 4 * 86400) % 604800); /* parse the sched line */ --- 168,187 ---- { time_t t_cur, t_base, t_begin, t_end; int s_dow, s_hr, s_min, e_dow, e_hr, e_min, count; + long int tz_s; + struct tm *tm; /* set up the time base */ ! (void) time(&t_cur); ! tm = localtime (&t_cur); ! debug3(D_SCH, "time is now: %d\n", t_cur); ! ! /* tz_s = timezone - (daylight ? 3600 : 0) ; */ ! tz_s = timezone - (tm->tm_isdst ? 3600 : 0) ; ! t_base = t_cur - ((t_cur + 4 * 86400) % 604800); + debug3(D_SCH, "t_base is now: %d\n", t_base); /* parse the sched line */ *************** *** 181,192 **** while(sscanf(nextword(p, &count), "%d:%d:%d-%d:%d:%d", &s_dow, &s_hr, &s_min, &e_dow, &e_hr, &e_min) == 6) { p += count; ! t_begin = t_base + (s_dow % 7) * 86400 + s_hr * 3600 ! + s_min * 60; ! t_end = t_base + (e_dow % 7) * 86400 + e_hr * 3600 ! + e_min * 60; ! /* hadle week overlaps */ if(t_end < t_begin) { --- 190,201 ---- while(sscanf(nextword(p, &count), "%d:%d:%d-%d:%d:%d", &s_dow, &s_hr, &s_min, &e_dow, &e_hr, &e_min) == 6) { p += count; ! t_begin = t_base + ((s_dow % 7) * 86400 + s_hr * 3600 ! + s_min * 60 + tz_s) % 604800; ! t_end = t_base + ((e_dow % 7) * 86400 + e_hr * 3600 ! + e_min * 60 + tz_s) % 604800; ! /* handle week overlaps */ if(t_end < t_begin) { *************** *** 483,489 **** /* set the line owned by root */ ! (void) chmod(devname, 0666); if (! stat(devname, &st)) (void) chown(devname, 0, st.st_gid); --- 492,498 ---- /* set the line owned by root */ ! (void) chmod(devname, 0660); if (! stat(devname, &st)) (void) chown(devname, 0, st.st_gid); >-- >=============================================================================== > Gary Moyer gary@mojo.uu.pdx.edu, gary@rodney.cs.pdx.edu [LINUX] > gary@acacia.cs.pdx.edu [386BSD] moyerg@cs.pdx.edu > "Linux: more than an OS -- it's a social movement" -- ___________________________________________________________________________ John Paul Morrison University of British Columbia, Canada Electrical Engineering jmorriso@rflab.ee.ubc.ca VE7JPM Day 35 of Canada held hostage: the socialist Liberal tax & spend tyranny ------------------------------ From: mlandaue@ux4.cso.uiuc.edu (Michael C Landauer) Subject: Can't telnet to my machine Date: 29 Nov 1993 21:41:23 GMT I'm having trouble telnetting to my machine. I can't even do telnet 127.0.0.1 I can do it once it I run: telnetd -debug & How do I get telnetd to load at bootup, and stay loaded for more than one connection? -- Michael Landauer | IMPEACH THE SGA NOW!!!!!! mlandaue@ux4.cso.uiuc.edu | ------------------------------ From: barr@pop.psu.edu (David Barr) Crossposted-To: comp.os.linux.help Subject: Re: Can't telnet to my machine Date: 29 Nov 1993 22:09:35 GMT In article <2ddqa3$ne0@vixen.cso.uiuc.edu>, Michael C Landauer wrote: >I'm having trouble telnetting to my machine. I can't even do >telnet 127.0.0.1 > >I can do it once it I run: >telnetd -debug & Check your /etc/inetd.conf. It should have something like this in it: telnet stream tcp nowait root /usr/etc/telnetd telnetd Also make sure that you are running inetd. (typically out of /etc/rc.net or rc.local) followups redirected to the appropriate newsgroup. --Dave -- "No man is good enough to govern another man without that other's consent" - Abraham Lincoln ------------------------------ ** FOR YOUR REFERENCE ** The service address, to which questions about the list itself and requests to be added to or deleted from it should be directed, is: Internet: Linux-Admin-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux.admin) via: Internet: Linux-Admin@NEWS-DIGESTS.MIT.EDU Linux may be obtained via one of these FTP sites: nic.funet.fi pub/OS/Linux tsx-11.mit.edu pub/linux sunsite.unc.edu pub/Linux End of Linux-Admin Digest ******************************