From: Digestifier To: Linux-Admin@senator-bedfellow.mit.edu Reply-To: Linux-Admin@senator-bedfellow.mit.edu Date: Fri, 19 Nov 93 07:14:04 EST Subject: Linux-Admin Digest #174 Linux-Admin Digest #174, Volume #1 Fri, 19 Nov 93 07:14:04 EST Contents: Re: authd/pidentd (Peter Eriksson) Disk Geometries and performance (Keith Smith) Re: SLIP FAQ (Matt Welsh) Re: Network reliability? (Jeffrey Wescott) Re: Getting Net Stats (Robert Glamm) Re: Major/Minor numbers for Ethernet (Tero Laakkonen) Re: Can't run VI and other programs!!! (Steve VanDevender) Re: SCSI timeouts: worse on a different machine (Kai Makisara) How to reach network? (Maxim Matveev) Slackware 1.1.0 & Adaptec 1542 extended mapping (Joe Real) Help with slack 1.1.0 and named (Vincent R Koser) ^? and Vi (hkennedy@mercury.ncat.edu) [Q]: /dev permissions (Steven M. Gallo) Re: TCL 7.1/TK 3.4 on Linux ? (Joerg Lenneis) Re: Default shell on login [ANSWER] (Zeyd M. Ben-Halim) Re: Getting Net Stats (Alan Cox) Re: SCSI timeouts: worse on a different machine (Eric Youngdale) ---------------------------------------------------------------------------- From: pen@lysator.liu.se (Peter Eriksson) Subject: Re: authd/pidentd Date: Wed, 17 Nov 1993 19:47:45 GMT stevenh@sci.kun.nl (Steven Hessing) writes: >In <2c2dah$g00@genesis.ait.psu.edu> fee@cxf111.rh.psu.edu (Chuck Fee) writes: >>Has anyone successfully compiled authd or identd for Linux? >Yes, Pauline Middelink (middelin@calvin.iaf.nl) has. You can get the >authd package at sunsite.unc.edu:/pub/Linux/system/Daemons/authd.tar.gz >You don't need to apply the kernel patches if you're running Linux >0.99.13q. I have now included the Linux-specific stuff from Pauline's authd/identd (which was based on an old Pidentd (version 1)) into the latest version of Pidentd (version 2.2beta, soon to be non-beta). You can find it on ftp.lysator.liu.se:pub/ident/servers as the file "pidentd-2.2beta.2.tar.gz", or on ftp.uu.net at the address below as soon as it has mirrored ftp.lysator.liu.se (which is the master distribution point). >>btw, the source for both of these is available at >>ftp.uu.net:/networking/ident/servers /Peter -- Peter Eriksson pen@lysator.liu.se Lysator Academic Computer Society ...!uunet!lysator.liu.se!pen University of Linköping, Sweden ------------------------------ From: keith@ksmith.com (Keith Smith) Subject: Disk Geometries and performance Date: Thu, 18 Nov 93 03:21:03 GMT In article <56975@sdcc12.ucsd.edu>, Ian McCloghrie wrote: >keith@ksmith.com (Keith Smith) writes: > >>Optimizing code _at the computer_ for disk geometry was an interesing >>idea that should now be killed. One should now leave the low level data >>movement optimizations to the DRIVE. Then a given drive can use it's >>cache & knowledge of the drive makeup to optimize on a per drive basis, >>probably much more efficiently than you could with the CPU. The obvious >>benifits are simplifying filsystem code to block clustering. > > The problem with this idea is that while the disk controller >knows everything about the disk, it doesn't know anything at all about >the filesystem structure. Said fs structure information becomes >necessary to do cylinder grouping, etc. If the OS can do >cylinder/head/sector positioning, it can make much more intelligent >choices about what to put where than the controller hardware can. >Take as an example a filesystem doing striping in software, if the >disk controller remaps your request to somewhere else on one drive, >but not on the other, you've lose a lot of the benifit. You do? Most of the modern 1G drives have like 256K of cache. I expect that number to hit in the 1M range before very long. Most of the modern drives are also REALLY fast, and have a WHOLE LOT of geometry wierdness, bad blocks & re-maps. To try and build all this into your filesystem code is plainly ludicrous. SURE, it might net you another 3-5% in performance, assuming you get it right and tweak the hell out of it for each particular drive in question (NOT drive type, each INDIVIDUAL drive). Secondly you STILL need to optimize to keep your filesystem clustering reads/writes on back to back block numbers, letting the drive figure the best way to actually retrieve/scribble the blocks. It doesn't take a mac genious to figure that the closer the block numbers the closer the data is on disk. I'd be willing to bet that there is some pretty nifty ROM code on some of the new superfast SCSI-II drives, to really optimize all this ON the drive. I think you'd be hard pressed to actually DEMONSTRATE any appreciable performance improvement with today's disk drives by rolling your own cylinder, head, sector optimizations. -- Keith Smith keith@ksmith.com 5719 Archer Rd. Digital Designs BBS 1-919-423-4216 Hope Mills, NC 28348-2201 Somewhere in the Styx of North Carolina ... ------------------------------ From: mdw@cs.cornell.edu (Matt Welsh) Subject: Re: SLIP FAQ Date: Thu, 18 Nov 1993 03:30:18 GMT In article <1993Nov12.063421.2413@srzts100.alcatel.ch> norbert.kuemin@alcatel.ch writes: >Debi Reid (dreid@mailer.cc.fsu.edu) wrote: > >: I looked several places (sunsite, tsx-11, etc) but could not find >: and FAQ's on SLIP via dialup. I remember seeing them port up in >: serveral inet news linux areas. Could someone mail this FAQ >: or point me in the right direction? > >Take a look to the NET-2-FAQ and to the SERIAL-FAQ The NET-2-FAQ has been superceded by the NET-2-HOWTO posted to c.o.l.announce just yesterday. I know of many people who are reading and relying on the VERY outdated NET-2-FAQ although I have tried to make it quite clear that this document is OBSOLETE. Apparently people are still picking it up from a few (ahem) FTP sites where it still lingers. Please, if you see a copy of the NET-2-FAQ with a release date before November 1993, please nuke it and replace it with the NET-2-HOWTO from sunsite.unc.edu:/pub/Linux/docs/HOWTO. Thanks. -- "Do you want to be Finnish? Sure, we all do!" ------------------------------ From: wescott@spectrum.cs.bucknell.edu (Jeffrey Wescott) Subject: Re: Network reliability? Date: 18 Nov 1993 07:40:28 GMT Reply-To: wescott@spectrum.cs.bucknell.edu >>>>> In article <163149@netnews.upenn.edu>, sjt@enlil.museum.upenn.edu (Steve Tinney) writes: > Nntp-Posting-Host: enlil.museum.upenn.edu > Alan Cox (iiitac@swan.pyr) wrote: > : Try pl13r > : alpha and see how well it goes. If its giving you hiccups let me know, > : Alan > I can't even get /etc/ifconfig sl0 to report a known interface with pl13r. I sat > and watched slip.o get added to net.a: do I need to recompile ifconfig? The > rest of the setup, except the kernel, is from Slackware 1.1.0 and is pl13. I am not sure whether or not slip.o being compiled in means anything or not (you are probably right in assuming that it does), but I had the same problems until I added a line in /usr/src/linux/drivers/net/CONFIG that said: CARDS = -DSLIP ... and of course the normal: SLIP_OPTS = -DSL_DUMP -DSL_COMPRESSED Hope this helps. -- Jeffrey Wescott ------------------------------ Crossposted-To: comp.os.linux.help From: glamm@rushmore.ee.umn.edu (Robert Glamm) Subject: Re: Getting Net Stats Date: Thu, 18 Nov 1993 16:37:49 GMT >>My SLIP connection is now working reliablym with no problems not caused by >>my own confusion but...its slower than I hoped (eg 12-14kbps effective on a >>nominal 38.4kps connection, on GIFS which should compress fairly well). >>Watching the lights yields what I'd expect (receive on continuously, an >>occassional flash on transmit) so... >you will never see 12-14kbps (i am assuming a 14.4K modem). i get about >2.5Kbytes/sec for text files and 1.5Kbytes/sec for binary and compressed >files. from the people i have talked to, this seems to be typical Ahh, but what you quote above is MORE than 12-14kbps. It takes about 10 bits to send a character; if you get 1.5kBytes/s, your connection is 15 kpbs. If you send 2.5kBytes/s, your connection is 'effectively' 25 kbps. Bob Glamm glamm@everest.ee.umn.edu -- Bob Glamm | There once was a programmer in C, 614 Delaware St. SE | Who had a very hard time with struct, did he; Minneapolis, MN 55455 | "My data I will scatter, glam0001@student.tc.umn.edu | throughout my program's grey matter," ------------------------------ From: laakkone@klaava.Helsinki.FI (Tero Laakkonen) Subject: Re: Major/Minor numbers for Ethernet Date: 18 Nov 1993 10:36:10 +0200 In <1993Nov17.182326.5857@swan.pyr> iiitac@swan.pyr (Alan Cox) writes: >In article mmatveev@boi.hp.com (Maxim Matveev) writes: >>Hi, >> >>Can anyone say me the correct major/minor pairs for LANCE and DE-600 >>pocket Ethernet adapter. I run SLS 1.03 with 0.99pl13. ANd only >>devices for WD and 3Com cards are in /dev/. I read NET-2-HOWTO - no >>information. >This comes up again and again. You don't need no /dev entries or anything >of the sort unless you are playing with Fred Van Kempens NET2E code. Neither >2D (in Slackware/SLS/...) nor NET2Debugged (in pl13f-r [ALPHA pl14]) need >the dev entries. hey dudes, real cool. could somebody explain *why* they're not needed? if you're some kind of kernel or inet guru the above question might look like a rather stupid question to you. but i bet you once asked dumb questions too ;-). -tero -- "i abhor you pretentious insight. i respect conscious guessing because it consists of two good qualities: courage and modesty." -imre lakatos ------------------------------ From: stevev@miser.uoregon.edu (Steve VanDevender) Subject: Re: Can't run VI and other programs!!! Date: 18 Nov 93 01:21:31 In article ez025807@othello.ucdavis.edu ( ) writes: When I want to run programs such as CU or VI I get an error message saying: "No space left on device" Does anyone recognize this problem????? You have filled up the partition (or partitions) where /tmp and /usr reside. Deleting unnecessary files may fix your problem. Try removing the stuff in /tmp first; /tmp can accumulate things. -- Steve VanDevender stevev@greylady.uoregon.edu "Bipedalism--an unrecognized disease affecting over 99% of the population. Symptoms include lack of traffic sense, slow rate of travel, and the classic, easily recognized behavior known as walking." ------------------------------ From: makisara@vtinsx.ins.vtt.fi (Kai Makisara) Subject: Re: SCSI timeouts: worse on a different machine Date: 18 Nov 1993 10:38:43 GMT In article rnichols@cbnewsg.cb.att.com (robert.k.nichols) writes: ... The only time I see this is when the tape drive is performing a lengthy operation that does not involve an ongoing data transfer. Rewind is the big culprit, but the first read operation on a newly loaded tape also takes a while. I haven't yet examined the SCSI operation with a logic analyzer, but judging by the blinkenlights, it would appear that the drive is not disconnecting, but hangs onto the bus during the entire operation. Perhaps the firmware on the host adapter is not giving the target the disconnect privilege. I'll have to check into it. (Gimme a chance! I haven't even got this thing booting reliably yet.) -- Bob Nichols AT&T Bell Laboratories rnichols@ihlpm.ih.att.com My adapter/drive combination (Adaptec 1542B and Wangtek 5525ES) seems to disconnect during lengthy operations. There is a longer timeout for the things expected to take long. If I remember correctly, I at some time found the location where the short timeout was incorrectly used (it was not an obvious place) and corrected it. The correction must then be among some other patches which I will submit to Eric Youngdale for inclusion in the official kernel very soon. I am just waiting for a second opinion for one of the patches. If someone else wants to try the patched driver, just drop me a note. Kai -- DISCLAIMER: The views expressed here are mine and may not always coincide with the views of my employer. * Kai Makisara * email Kai.Makisara@vtt.fi * ------------------------------ From: mmatveev@boi.hp.com (Maxim Matveev) Subject: How to reach network? Date: Thu, 18 Nov 1993 22:16:02 GMT Hi, I try to configure NET-2 stuff and any time then I try to run route or smth other I get "SIOCADDRT: Network unreacheble". I run SLS-1.03 with kernel 0.99pl13. Ethernet card - WD8003EP. I can ping my Linux box from other WS on the NET, Linux receive some info about ARP, but no outgoing connection. I've tried to solve this problem by myself, but neiver NET-2-HOWTO, neiver Network Adm. Guide don't help. -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Max Matveev (208) 396-7900 (work) mmatveev@hpbs669.boi.hp.com (208) 385-9103 (home) ------------------------------ From: LARS@istda.zer.de (Joe Real) Subject: Slackware 1.1.0 & Adaptec 1542 extended mapping Date: Wed, 17 Nov 1993 11:00:00 +0100 Hallo Eric ! e> >Fine, say, does this version run with the extended mapping feature of the e> >Adaptec 1542C or 1542B with new Bios and Mcode? e> e> I doubt it. You should not be trying to run Linux with the extended e> mapping in the first place - Adaptec only recommends that this be used with e> DOS-only systems. It is possible to use all of a >1Gb disk with linux e> without enabling this creature. How can I use more than one Gig ? You mean just with the new Bios and new Mcode it runs ? cu, Lars ## CrossPoint v2.92 R ## ------------------------------ From: vkoser@niblick.ecn.purdue.edu (Vincent R Koser) Subject: Help with slack 1.1.0 and named Date: Thu, 18 Nov 1993 21:35:13 GMT I have gotten to the point where my machine is on the net and can find anything that is in the etc/hosts file. If I try to telnet to anywhere other than machines listed in here I get unknown host. I got the named running but to no avail. I put my nameserver in the resolv.conf file but it didn't help either. Any helpful hints or suggestions vince -- Vincent R Koser Krannert Technical Services vkoser@ecn.purdue.edu ------------------------------ From: hkennedy@mercury.ncat.edu Subject: ^? and Vi Date: Thu, 18 Nov 1993 13:19:07 GMT Hi, After recompling the kernal and upgrading the lib (4.4.1), whenever I use vi in command mode the backspace outputs the characters ^? now. Has anyone seen this problem before and what is the solution? This also happens with less. Thanks, hkennedy@ncat.edu ------------------------------ From: smgallo@cs.buffalo.edu (Steven M. Gallo) Subject: [Q]: /dev permissions Date: Thu, 18 Nov 1993 14:26:21 GMT Could someone post a listing of the permissions for their /dev directory? I think mine got screwed up somewhere along the line. Thanks, Steve -- Steven M. Gallo "If you don't expect too much from SUNY at Buffalo, Dept. of Comp. Sci. me you might not be let down.." smgallo@cs.buffalo.edu - Gin Blossoms ------------------------------ From: lenneis@wu-wien.ac.at (Joerg Lenneis) Crossposted-To: comp.lang.tcl Subject: Re: TCL 7.1/TK 3.4 on Linux ? Date: 18 Nov 1993 15:36:53 GMT Nils Rennebarth (nils@hertha.exp-math.uni-essen.de) wrote: : Joerg Schaefer (joscha@atab.lahn.de) wrote: : : Philip C. Cox (pcc@s1.gov) schrieb: : : > Has anyone gotten tcl 7.1 & tk 3.4 to run on linux? I am : : > working on it, but do not want to duplicat effort :-) : : Yes. : : I compiled it with only two changes. In tclMain.c i inserted : : #include (for the definition of R_OK). The same in tkMain.c. : : Annyone here made dll-Libraries for tcl7.1 and tk3.4 ???? : I once made them for tcl 7.0 and tk3.3 Patchlevel 2 each, just : to learn how to make shared libraries. Worked ok, but some of the : tests failed (due to bugs that had nothing to do with the sharedness : of the library), and I had no idea about a possible compatibility : with the earlier shared library versions, so I didn't upload : anything. If someone made the SHARED libraries already, please : upload and announce them. I tried the same. What I have not worked out so far are which imported symbols tcl/tk needs. I guess the ones from libc, but any others as well? -- Joerg Lenneis University of Economics and Business Adminstration Department for Applied Statistics and Data Processing Augasse 2-6, 1090 Vienna, Austria Tel. *43/222/31336 4758 email: lenneis@wu-wien.ac.at ------------------------------ From: zmbenhal@netcom.com (Zeyd M. Ben-Halim) Subject: Re: Default shell on login [ANSWER] Date: Thu, 18 Nov 1993 21:07:33 GMT In article koch@rtsg.mot.com (Clifton Koch) writes: > > Thanks to all who responded. > > The problem boiled down to bash not reading the .bashrc on login. I had >assumed that sh was being run (which I know is also bash, but it >behaves differently when run that way) from some sort of default. Bad >assumption. The answer is that on a login shell, bash reads .bash_profile, >and not .bashrc. On anything else it appears that it runs .bashrc and not >.bash_profile. > > I don't feel too bad on not figuring this out, because I can't find this >documented *anywhere* (except maybe the source :). It IS documented! Look under the INVOCATION section. Its behavior is explicitly mentioned. Zeyd >Cliff >-- >----------------------------------------------------------------------------- > Cliff Koch > Motorola Cellular Infrastructure Division > koch@rtsg.mot.com -- --- Zeyd M. Ben-Halim zmbenhal@netcom.com 10479 1/4 Santa Monica Blvd, LA, CA, 90025 (310) 470-0281 ------------------------------ Crossposted-To: comp.os.linux.help From: iiitac@swan.pyr (Alan Cox) Subject: Re: Getting Net Stats Date: Thu, 18 Nov 1993 13:55:09 GMT In article <1993Nov15.184615.9318@wabbit.demon.co.uk> root@wabbit.demon.co.uk (root) writes: >In article <2c8348$3re@terminator.rs.itd.umich.edu> sarr@citi.umich.edu writes: >>Is there any way existing tool that will dump error counts, >>retransmissions/duplicates, etc. Netstat does that on many systems, but >>apparently not this one. >Not that I know of. I think ther is some debug code in the slip source you could >enable (or if not add some) but as per above your setup seems to be working fine. For the basics of the story. cat /proc/net/dev Alan iiitac@pyr.swan.ac.uk ------------------------------ From: eric@tantalus.nrl.navy.mil (Eric Youngdale) Subject: Re: SCSI timeouts: worse on a different machine Date: Thu, 18 Nov 1993 15:19:25 GMT In article rnichols@cbnewsg.cb.att.com (robert.k.nichols) writes: >The only time I see this is when the tape drive is performing a lengthy >operation that does not involve an ongoing data transfer. Rewind is the >big culprit, but the first read operation on a newly loaded tape also >takes a while. I haven't yet examined the SCSI operation with a logic >analyzer, but judging by the blinkenlights, it would appear that the >drive is not disconnecting, but hangs onto the bus during the entire >operation. Perhaps the firmware on the host adapter is not giving the >target the disconnect privilege. I'll have to check into it. >(Gimme a chance! I haven't even got this thing booting reliably yet.) It is also possible that the tape drive itself is not smart enough to disconnect. I now have a utility available that can be used to query all scsi devices and find out information from their mode pages - the disconnect/reconnect page would be of particular interest here. If the drive does not have a disconnect/reconnect page, this would imply that it would hold onto the bus during long operations. The latest version of the utility is called scsiinfo-1.4.tar.gz on tsx-11 in pub/linux/ALPHA/scsi, and it has a really cool X-based interface that lets you modify parameters and send them back to the drive. You need to have wish (part of Tcl/Tk) to use the X-interface, but it is still possible to get a display of the information from the drive without having wish or X on your system. -Eric -- "The woods are lovely, dark and deep. But I have promises to keep, And lines to code before I sleep, And lines to code before I sleep." ------------------------------ ** 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 ******************************