From: Digestifier To: Linux-Admin@senator-bedfellow.mit.edu Reply-To: Linux-Admin@senator-bedfellow.mit.edu Date: Tue, 13 Sep 94 20:14:02 EDT Subject: Linux-Admin Digest #60 Linux-Admin Digest #60, Volume #2 Tue, 13 Sep 94 20:14:02 EDT Contents: Re: talk problem (Cord Hockemeyer) Re: DOMM 4 Linux /X is OUT !!! (Mark Lord) Re: 2.90 sound driver problem in 1.1.45 (Hannu Savolainen) xdm and sessreg: wtmp/utmp (Jim Ockers) Re: Removing LILO ? How? (Timothy Murphy) Re: Lilo wish (Mats 'MaDsen' Wikholm) Re: Is there any Accounting in Linux?? (Morten Steinvik) Re: Everyone's MOUNT - WARNING! (Morten Steinvik) Re: Two Hostnames: Possible? (Eelco H. Essenberg) What would be faster Dx-50 or DX2-66? Bus and memory performance. (Andreas Helke) LINUX Security Patch FTP Site??? (Joe Rach) Rstat patch (for xmeter) (Brad Cain) Re: PPP/IP Forwarding Problem (Vassili Leonov) Re: Remote Shell from linux to Solaris? (David Sears) ---------------------------------------------------------------------------- From: cord@msheinz.psych.nat.tu-bs.de (Cord Hockemeyer) Subject: Re: talk problem Date: Tue, 13 Sep 1994 14:11:06 GMT Dennis Duffner (duffy@dduff@dduff.ppci.com) wrote: | I have a problem with talk. | I cannot talk to a user I know on another system, even though they have | been able to talk me, I can't reply. | I get this: Connection Refused. (111) | What's this all about anyhow? I can't find anything on this anywhere. Quite simple: His host refuses to answer your finger request. :) A lot of systems don't serve finger requests because this was a security hole (with bad consequences) some time ago. Cord -- Cord Hockemeyer, TU Braunschweig, Abt. Math. u. Sozialpsychologie email: C.Hockemeyer@tu-bs.de / i3160503@dbstu1.bitnet ------------------------------ From: mlord@bnr.ca (Mark Lord) Subject: Re: DOMM 4 Linux /X is OUT !!! Date: 13 Sep 1994 14:32:02 GMT In article psmith@wellfleet.com writes: >Yikes! 8M ram + 8M swap is *not* enough to play DOOM! I have 16M ram >and I'm *still* swapping quite a bit--but DOOM is still pretty speedy at >normal resolution (not doubled, etc.) I have like 20M swap or >something. Also I'm playing the registered WAD off my DOS partition; I That's your problem. Copy the WAD to an ext2 partition and run it from there. On my DX2/66+16MB+ATIGUP system, DOOM is *fast* in -2 (pixel doubling) mode! -- mlord@bnr.ca Mark Lord BNR Ottawa,Canada 613-763-7482 ------------------------------ From: hannu@voxware.pp.fi (Hannu Savolainen) Subject: Re: 2.90 sound driver problem in 1.1.45 Date: Tue, 13 Sep 1994 18:56:42 GMT flipper@pentagon.io.com (Peter Sierant) writes: > I've been using the 1.1.19 kernel for several months without a >problem. Sound worked great for au files, mods, and cd audio, etc. >Last night I upgraded to 1.1.45/ 2.90 sound, and now if I >cat welcome.au > /dev/audio, It plays but gives a write io error, then >repeats part of the file, then quits. Sound card is a PAS16, machine is This occurs when there is an IRQ conflict. Propably the IRQ for PAS16 is in use by some other device on your system. Hannu -- ============================= Hannu Savolainen hannu@voxware.pp.fi "Don't use Windows since there is a door!" ------------------------------ From: ockers@mecad5.maem.umr.edu (Jim Ockers) Subject: xdm and sessreg: wtmp/utmp Date: Tue, 13 Sep 1994 05:57:20 GMT Reply-To: ockers@umr.edu Why can't I get sessreg to register X clients in utmp and wtmp??? My userlist always shows up as being like this: User tty From login@ idle JCPU PCPU what ockers ttyp0 11:49pm something As you can see, the From field is blank. I think that there should be a :0.0 in there. Also, the wtmp does not even show anybody logged in! This is bad. When someone connects to my xdm and logs in, I would like to have some record of that! I am running kernel 1.0.9, standard slackware 2.0 distribution. Nothing too fancy. XFree86 version 2.1.1 . SysVInit. I looked at the sessreg(1) man pages. It claims to handle utmp and wtmp for non-init clients. It does not. This is not covered in any FAQ that I could find. According to the man pages, sessreg acts differently for BSD or SYSV xdms. I am guessing that the SYSV stuff applies to me. The BSD problem seems to be the indexing of utmp, wherein each entry has a slot number. sessreg looks at the Xservers file and the /etc/ttys file, and takes (# of entries in /etc/ttys)+1=(the utmp slot for this session). The 1 there is for the X session, so that it can be added to utmp. My /etc/ttys is not updated to show the ttys that are currently in use! It contains 10 lines which are the first 6 VCs and the first 4 ptys. It has not been modified since I installed linux. This screws up sessreg; it says "lseek: invalid argument" since sessreg is trying to update the 11th slot in utmp. Of course there is no 11th slot in utmp. Even if there were, it would probably be the wrong slot. sessreg -s specifies a slot number for sessreg to assign the session to. sessreg does not return an error message if you assign a slot that exists in utmp; but it also does not modify utmp or wtmp in any way! Also, in the SYSV environment sessreg is supposed to ignore the BSD-specific flags. Since it is not ignoring them (lseek?) I must be in a BSD environ- ment. Right? And if I leave off the BSD-specific flags, I get the lseek error. My /usr/lib/X11/xdm/Xservers file looks fine. The first non-commented line it contains (and the last line in the file) is :0 local /usr/X386/bin/X This is from the man pages: "If for some strange reason your system uses a file other than /etc/ttys to manage init, the -t option can direct sessreg to look elsewhere for a count of terminal sessions." My system apparently does not use /etc/ttys to manage init. What does it use? Any ideas? I put sessreg -a $USER in the file specified by the DisplayManager._0.startup resource. Likewise, sessreg -d $USER is in the file specified by the reset resource. It has no visible effect on anything. (I know those scripts are executing properly because they do other things.) If you email me your ideas on how to make this work, I'll post a summary/ solution. Thanks so much. -- Jim (ockers@umr.edu) Ask me about Linux! | | PGP public key available upon request, or from my web pages. ---+---+--- GE d? p c++ l++ u++ e++ m++(--) s n-(+) h+ f !g w+ t+ r !y | X | home page ---+---+--- | | ------------------------------ From: tim@maths.tcd.ie (Timothy Murphy) Subject: Re: Removing LILO ? How? Date: 8 Sep 1994 17:15:08 +0100 laud@cs.curtin.edu.au (Daniel Lau) writes: >I currently have LILO running as the boot manager. But now I would >like to remove that. I believe once I removed it, there will be no boot >manager, thus my MS-DOS partition won't be reached? As everyone has said, the command fdisk/mbr within DOS is the answer. THIS SHOULD BE STATED clearly and loudly in the setup program, when the MBR is created. -- Timothy Murphy e-mail: tim@maths.tcd.ie tel: +353-1-2842366 s-mail: School of Mathematics, Trinity College, Dublin 2, Ireland ------------------------------ From: mwikholm@at8.abo.fi (Mats 'MaDsen' Wikholm) Subject: Re: Lilo wish Date: Tue, 13 Sep 1994 20:08:27 GMT In article <1994Sep13.013746.13885@galileo.cc.rochester.edu> mflt_cif@uhura.cc.rochester.edu (Pile Smurf) writes: >I have my Lilo set up to ask whether I want to run DOS or Linux at boot-up. >The problem is, if I crash the machine remotely, and I manange to re-boot >it, it will stay at the Lilo prompt until I can get back to my room to hit >enter (to run linux). Is there a way to set up Lilo so that after a minute, >it will automatically choose the "default" choice? You might ofcourse have different version than I but this is what my lilo.conf looks like: =====lilo.conf===== boot = /dev/hda compact delay = 5 vga = 0 ramdisk = 0 root = /dev/hdb1 image = /vmlinuz label = linux other = /dev/hda1 table = /dev/hda label = msdos =====lilo.conf===== This automagically boots Linux. If I want to boot msloss I have to press ctrl or alt in order to choose between them. -- . . . . mwikholm@at8.abo.fi / win-nt from the people who invented edlin . . . . frantzgatan 3 E 25 / apples have meant trouble since eden . . . . 20380 abo finland / Linux, the way to get rid of boot viruses . . . . @358.(9)21.377.363 / this .signature is released as shareware . . . . -- . . . mwikholm@at8.abo.fi / win-nt from the people who invented edlin . . . . frantzgatan 3 E 25 / apples have meant trouble since eden . . . . 20380 abo finland / Linux, the way to get rid of boot viruses . . . . @358.(9)21.377.363 / this .signature is released as shareware . . . . ------------------------------ From: mortenst@colargol.edb.tih.no (Morten Steinvik) Subject: Re: Is there any Accounting in Linux?? Date: 13 Sep 1994 21:10:37 GMT Pete Deuel (deuelpm@craft.camp.clarkson.edu) wrote: : In article tony@teleport.com (Tony Schwartz) writes: : >allow the Linux system to record who is on, when, how long, etc?? The file /etc/utmp keeps track of who is on 'right now', you can reach its data via the "who"-command. : Or, if you're looking for something simpler, the command "last -3 user" will : give you the last 3 logins of user, telling you where they logged in from and : for how long... It uses a file called "wtmp" and if someone knows of a way to : grep through it (it's in some weird format), you could extract the info you Correctly the file /var/adm/wtmp keeps track of who has logged in when and for how long and use the "last"-command to read it (the "weird" format). Simply "last" gives a list of all logins since last deletion of the file, combinate it with "grep" and for example wordcount. Try "last | grep user| wc -l" to see how many times user has been on... Did this solve your problem? : need... Careful, though... I think that this file is wiped and retouched on : boot up (you'd have to update your accounting after booting everyone off but : before the system restarts in your "shutdown" routine) wtmp is not automatically deleted on boot-time (unless some of the newer installations use it as default), it just keeps growing and growing for each login. You should delete and retouch it from time to time yourself. There is some accounting-stuff (acct) as well, haven't tried it myself yet, but you might like to check it out. This is not default and you should find it on anonymous ftp somewhere. Morten Steinvik -- (2b) || (!2b), that is the question... Turn your 486 into an XT--just add Windows! Morten Steinvik mortenst@edb.tih.no Jogger on IRC ------------------------------ From: mortenst@colargol.edb.tih.no (Morten Steinvik) Subject: Re: Everyone's MOUNT - WARNING! Date: 13 Sep 1994 21:17:13 GMT Jay Ashworth (jra@zeus.IntNet.net) wrote: : >>/dev/fd0 /dos-a msdos user,noauto : >> : >>This means that any user can type 'mount /dos-a' and get the floppy all : >>for herself, without nasty suid root problems. : : The option you _meant_ to give him was "nosuid", not noauto. :-) nosuid *AND* noauto, if you ask me. noauto causes the system to not attempt mounting the floppy at boot-time (or whenever else the /etc/fstab is read), it has to be mounted seperately. nosuid is good, too, as it stops attempts to run setuid files on the mounted fs. Why choose one? Morten Steinvik -- (2b) || (!2b), that is the question... Turn your 486 into an XT--just add Windows! Morten Steinvik mortenst@edb.tih.no Jogger on IRC ------------------------------ From: essenber@dutiws.twi.tudelft.nl (Eelco H. Essenberg) Subject: Re: Two Hostnames: Possible? Date: Tue, 13 Sep 1994 20:04:06 GMT In article <353tu2$otn@sun2.ruf.uni-freiburg.de>, Peter Suetterlin wrote: >Bill C. Riemers (bcr@k9.via.term.none) wrote: >: >>>>> "Peter" == Peter Suetterlin writes: >: Sure you already have 2 names: >: localhost >: lupo > >: If you want more, just list them in /etc/hosts. > > >: Bill > >Yes, that was what I hoped. But if I keep lupo (and, of course, >localhost), and add the next name, smail refuses to use it (read: it >doesn't recognize it as 'myself'). >I get error messages of the mail beeing refused some 10 or twenty times >until I change the hostname. >Think I should take an evening and re-read the Manuals.... sigh Someone else has already posted this, but as far as e-mail is concerned the solution lies in your mailer-daemon's configuration. Assuming you use Smail, edit the line in /usr/lib/smail/config that says "other_names" or something similar. Add ":second.host.name" at the end of the line to solve your problem. > > Peter > Regards, Eelco. >------------------ Peter 'PIT' Suetterlin ----------------- >| Kiepenheuer Institut | Sternfreunde Breisgau e.V | >| fuer Sonnenphysik | | >| 0761/3198-210 | 0761/71571 | >---- -- ==========================< Eelco Essenberg >=============================== E.Essenberg@TWI.TUDelft.NL ftp@ftp.twi.tudelft.nl FTP Manager: ftp.twi.tudelft.nl Click me! ------------------------------ From: andreas@orion.mgen.uni-heidelberg.de (Andreas Helke) Crossposted-To: comp.sys.ibm.pc.chips,comp.sys.intel Subject: What would be faster Dx-50 or DX2-66? Bus and memory performance. Date: 13 Sep 1994 21:02:31 GMT kieferal.asa@asa.org wrote: : In article <34oqfo$2l0@nyx10.cs.du.edu>, writes: : > In addition, the 33 MHz motherboard will allow for VESA local bus, which : > I would highly recommend getting. Actually, every motherboard I've seen : > lately that'll go to 50 MHz is a 33/40/50 MHz (sometimes even 20 & 25) : > with local bus. The VLB slots won't work at 50 MHz, though. : This is not true. The definition supports 2 VLB slots at 50MHZ and 3 at 33MHZ. : I have seen 486DX50 VLB motherboards on the market. Vesa 1 spec says slots are not allowed for 50 MHz operation. Vesa 2 says 1 slot for 50 MHz, 2 Slots for 40 MHz, 3 Slots for 33 MHz. Mainboard vendors say I dont care about standards and reliability and will offer 50 MHz boards with 3 slots. And then to reduce the cost of manufacturing they will use a 3222 burst for 50 MHz cache reads reducing the transfer rate below that of a 33 MHz board with 2111 burst. What I dont know is if a 50 MHz board has an atvantage for other memory or bus operations. : > : > Don't cheat yourself. Get a DX/33, DX/40, or a DX2/66. : > VLB has the disatvantage that it needs at least 2 cycles for every transfer. Even follow up cycles in burst mode where PCI only needs one cycle. Therefore if you care about the speed of your bus you have to use PCI and be careful to pick the right chipset and board manufacturer. Good PCI boards operate at 65 MByte/second for STOSD, the best ever seen on a VLB board (actually a Asus PCI/VLB hybrid with Intel Aries chipset where the VLB hanging on the PCI bus over a bridge) was 42 MByte/second. The data come from various c't magazines from 10.93 to 8.94. c't seems to have tested almost every PCI mainboard. BTW 65 MB/second is the maximum transfer rate which can be achieved when using a 486DX chip. 32 Bit PCI has a theoetical limit of 132 MB/second. And now if someone could explain why the Pentium chipsets in general provide less PCI and possibly memory performance than the good 486 chipsets, I would be glad to hear it. Andreas -- Andreas Helke Institut fuer molekulare Genetik, Universitaet Heidelberg Im Neuenheimer Feld 230 69122 Heidelberg, Germany ------------------------------ From: r7980@hopi.dtcc.edu (Joe Rach) Subject: LINUX Security Patch FTP Site??? Date: 13 Sep 1994 17:27:38 -0400 Hello, I'm looking for an FTP site for LINUX security patches and need some help. Does anyone know of a site that has the bug fixes? Thanks in advance. ------------------------------ From: brad@chopin.udel.edu (Brad Cain) Subject: Rstat patch (for xmeter) Date: 13 Sep 1994 17:31:01 -0400 I would like to use xmeter, and check other things beside the cpu and load options. I have an old rstatd with a patch for the kernel, but the patch is for .99pl14 or something. Does anyone have a rstat patch for kernel 1.0 or 1.1? Or do I just need a new rstatd?? I'm using kernel 1.1.49 with an original slackware 2.0 install. thanks -- ****************************************************************************** brad@bach.udel.edu * Brad Cain N3NAF cain@ee.udel.edu * University of Delaware Electrical Engineering PGP key available via finger * -Comp. Sci/Signals/Communications/Networking- ------------------------------ From: vassili@cs.sunysb.edu (Vassili Leonov) Crossposted-To: comp.os.linux.misc,comp.os.linux.development,alt.uu.comp.os.linux.questions Subject: Re: PPP/IP Forwarding Problem Date: 13 Sep 1994 20:29:07 GMT jbarrett@onramp.net wrote: :I'm running Linux 1.1.49 and attempting to configure it as a router between my :local Ethernet and my Internet provider. I've posted on this before but maybe I if you just have an ethernet network and one box on it is connected to an Internet provider (you used some bad language like M#%&@$@ft there - so I'm not quite sure) then you need two things to do it proper: - make sure your kernel on SLIP/PPP machine is configured as ROUTING packets (make config) - make sure that other machines have route add default gw your_slip_hostname metric 1 somewhere in their rc.inet1 Sooo.... It's basic networking - why blame Linux on that... Read some good books on that though.... Has nothing to do with Linux development... Vassili. ------------------------------ From: dns@essnj3.essnjay.com (David Sears) Subject: Re: Remote Shell from linux to Solaris? Date: Sat, 10 Sep 1994 00:19:38 GMT In iialan@iifeak.swan.ac.uk (Alan Cox) writes: >In article <33t80d$mn4@hermes.acs.ryerson.ca> jnaughto@ee.ryerson.ca (JASON NAUGHTON) writes: >> Has anyone successfully xon'ed, rsh'ed, or rlogin'ed to a >>solaris work station from a linux station? Every time I try I happen >>to receive this lovely message: >> >>jnaughto@crystal: ~$ rsh tesla >>No utmpx entry. You must exec "login" from the lowest level "shell". >>rlogin: connection closed. >Get someone to fix the Solaris machine 8) >Alan >-- > ..-----------,,----------------------------,,----------------------------,, > // Alan Cox // iialan@www.linux.org.uk // GW4PTS@GB7SWN.#45.GBR.EU // > ``----------'`----------------------------'`----------------------------'' The solution is a workaround. In your xterm window, you can type `stty' and you will note that the speed of your pty is set to zero. This seems to be because the pty state is copied from the `controlling tty'. I think you are running xdm in daemon mode where there is no controlling terminal, thus the speed is zero. Start xdm in nodaemon mode, or start X with xinit, or type `stty 9600' in your xterm window. With any of the above, you will be able to rlogin to Solaris. I spent two days tracking this one down. Cneers ... -- David Sears | EMAIL: dns@essnj3.essnjay.com (preferred) Huntington, CT USA-06484 | or 76474.3113@compuserve.com ------------------------------ ** 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 ******************************