Re: Apache dies silently?
You're probably using unstable, are using PHP4 packages, and do not include PHP4 support with apache-ssl. Your system probably chugged along happily because you haven't done any apt-get update apt-get upgrade for a while. At least, that's how I got burned. There's some kind of conflict with PHP4 and apache in Unstable. The following worked for me: more /etc/apt/preferences Package: php4 Pin: release a=stable Pin-Priority: 1001 Package: php4-mysql Pin: release a=stable Pin-Priority: 1001 Package: php4-imap Pin: release a=stable Pin-Priority: 1001 Package: php4-cgi Pin: release a=stable Pin-Priority: 1001 Package: php4-pear Pin: release a=stable Pin-Priority: 1001 Package: php.* Pin: release a=stable Pin-Priority: 1001 Package: * Pin: release a=unstable Pin-Priority: 500 (Pins the PHP4 stuff to stable until this mess gets straightened out!). Thanks to Fabio Massimo Di Nitto from the debian-apache team for helping me figure out what what up. Ah well... that's why it's named unstable! --Rich j2 wrote: I have the following cookiemonster:~# dpkg -l | grep -ir apache ii apache 1.3.27.1-3 Versatile, high-performance HTTP server ii apache-common 1.3.27.1-3 Support files for all Apache webservers ii apache-ssl 1.3.27.1-3 Versatile, high-performance HTTP server with All of a sudden apache dies silently, nothing in any logs, it just never starts up. apache-ssl runs fine tho. I did a strace apache which didnt really reveal anything that i could see. Quite frankly, i am not even sure how to start to debug this one? The server has been running fine for ages? -- _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Apache dies silently?
John M Flinchbaugh wrote: On Tue, Sep 23, 2003 at 07:00:55PM +0200, j2 wrote: Sounds familiar, but i cant recollect php or paache being upgrade lately.. more /etc/apt/preferences slightly deeper, i think it was a conflict in the libs used by 2 php modules: pg and imap. i dropped the imap module and i was fine. i needed the postgres module. the debian bugs database should have more details. I didn't look into the PHP bugs... just found similar Apache bugs, including some duplicates. In my case, it wasn't the pgsql.so and imap.so Kerberos library symbol conflict. I didn't need pgsql so I never installed that one, and still had trouble (in my case, apache also exited cleanly from gdb, IIRC it segfaults with the pgsql/imap module conflict). The actual problem appears to be with php4 and the new apache. See the most recent notes on: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=200255 ...for some more info on php4 deaths unrelated to pgsql/imap. --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Anti-Spam ideas for usenet/list harvested email addresses
should be able to help in some way, either by adding to the EHLO command set or something on the users web site. There have to be better and still simple ways of doing this that make it cost much more to find our email addresses than it costs us to filter the junk. True. But you still don't solve the problem of having someone easily contact you off list. In the case of this email, I've decided I have something worthwhile to say on the topic at hand (or I'm bored, and want to babble about email filters...) so I hit reply to all. If I had to break my train of thought to sift through your website to find your email address, I'm probably not going to bother. Also consider the fact that some people do have to read email offline, and rely on the assumption that all necessary contact info is contained in the email itself. Enhancing EHLO would probably not be realistic, given that virtually all email clients would have to implement it. It's like saying oh, just turn on SMTP authentication, and we can be sure that the sender isn't a spammer, or at least can track them down. Images with pictures of your email address is fine, but again, it's just a slightly more difficult form of jacob at cachevalley dot com... eventually wouldn't the spammers just create OCR software that looks for email addresses in images on websites linked from your website? The sad part is that I've already squandered my username at this email address by putting it where it can be harvested in mass by worm/virus and UCE/UBE collection scripts, and I had already read an article cautioning me against this. Oh well live and learn (someday I'll learn anyway.) I'm going to look into setting up a new email address with mail server rules for delivery driven by a user supplied whitelist after waiting a few days for comments and flames on this idea. If you know of links to pages already discussing how to do this with postfix, please share them. Look to SpamAssassin. That will make a huge dent in your spam problem. Tack on Amavis for the latest in MS malware, and you're in business. I believe both integrate fairly well with Postfix. Amavis is also able to reject viruses during the SMTP transaction. This I would agree with, if your configuration allows it. Some good thoughts there... but I wonder just how many mailing lists would need to apply such a solution to make an impact, and how difficult it would be to apply. OTOH, you might find better results with simpler methods... --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Anti-Spam ideas for usenet/list harvested email addresses
Jeronimo Pellegrini wrote: Right, I forgot about that. Anyway... Blocking servers wouldn't help in the case of viruses, I think. Ordinary people get viruses, and the mail is sent through their (probably correctly configured) smarthost. Maybe something like Postfix header_checks? But that would also require some work :-( J. Unfortunately (or fortunately, depending on how the blocking works), some of the newer worms do not use the smarthost, but instead they use their own little SMTP engine. IIRC, Sobig.F did this. --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: weak perfomance on woody
[loonyx | rolf joho] wrote: hi there on my new woody-box i experience some major performance problems. thez appeared a day a go and i can not link this to any of my activities... what i observe: * there is constantly a hell lot of disk-activity. * starting X/gdm at boot takes around 2 minutes (restarting goes much faster). * applications sometimes take ages to start, eventhough right now things run quite well. * multiple instances of sendmail (currently 3) are continiously running. What are the continuous instances doing? `ps awux | grep sendmail` will provide much more info than just ps. I would guess that you have one daemon on port 25, one MSP daemon, and possibly a queue runner hanging around... * there might be a problem in delivering root-mail although i have set up an alias to myself. /var/mail is empty (no mailboxes there) - the mailqueue is empty right now. * memory- swap- and disk-usage are in the usual range, hardware seems to be fine. what i've done so far: * i did some monitoring using top and ps (see output below) but their output seemed pretty 'standard' to me. Hit M to sort by memory usage, also hit P to sort by CPU usage. Check the top offenders in each catagory to see if anything stands out. Looks like something's chewing RAM... * installed exim instead of sendmail but went back as this did not have any effect. If sendmail was appearing to be the culprit, I'd suggest turning off host status caching (unless the cache was on a reiserfs filesystem) what i would like to know: * has anyone any idea how to effectively track those performance-problems? * how do i monitor disk-activity at all (there is not even a disk-led on my box)? * any idea why there are no mailbox-files in /var/mail ? Have you set things up to deliver differently? Has the machine actually received mail to be delivered locally (the mbox files will not be created until you receive mail, IIRC). i know these are a lot of questions and i am not even sure if there is any relationsship between them. below is the output of top and ps. thanks for any advice rolf I also suggest that you uninstall identd. Also turn off identd checking in sendmail.mc (define(`confTO_IDENT',`0')dnl). It's not needed, not useful, and will slow down your mail run. Turn off the other stuff you don't need (do you really need named running? lwresd? gdm? lpt? XFree86? oafd?) You're running a lot of stuff on this box, some of which may be contending for resources. Looks like you're running SCSI... what's your partitioning scheme look like? What are you mounting remotely via NFS? what are you exporting via NFS? Run `ps awux | grep D` a couple of times, look for processes that are getting blocked during disk accesses. What kernel are you running? What SCSI controller? What processor are you running? Good luck! --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Intel 10/100 VE and eepro100
Andrew Perrin wrote: Greetings- Setting up a new computer (an IBM NetVista) that has an Intel 10/100 VE adapter built in, I'm unable to use a net install. What I did was to burn a CD with the vanilla kernel setup, and boot to it. I can boot fine, but when I try to load the eepro100 module, I get init_module: no such device or init_module: device or resource busy. Without that, I can't start the network, so I can't do the net install. Any advice on this? I thought eepro100 would work, even if e100 is probably the ultimate better choice. e100 is available only as source, so I can't use it to do the net install (right?). I thought that the 2.4bf image had both? If so, you should be able to do the net install that way. Might want to give that a try... --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SMTP over SSH
How about setting up your laptop to have SSH forward local port 25 to mail server port 25? No changes to exim, change your MUA to use localhost as your SMTP server, change your SSH client to forward localhost:25 to (your_smtp_server):25 and life is good. No need to 'bind' ports 22 and 25, SSH can do this on its own. If you're running linux on the laptop, look at ssh -L. If you're running another OS, look at the docs for the SSH client. I believe putty can do this, and I know that Tera Term SSH can do it. --Rich Alex Malinovich wrote: I have my laptop set up to work fine with my home mail server from just about anywhere. The only problem is that I have a couple of classes that I use my laptop for which block certain ports for the network. One of those ports is 25. So I'm able to read my mail just fine over IMAP, I just can't send any outgoing mail. I need some way to get around this. I can set up my MTA to listen to connections on, say 2525, but that is going to be up all the time. I could also set up iptables to forward any connections from port 2525 to port 25, but again, this will stay up all the time, which I don't want. What I'd like is to have a way to send traffic from my end on 2525 (or something else that isn't blocked) and have it go to port 25 on my mail server. What I had in mind, though I admit that I don't know if it's possible, is to establish a regular old SSH connection on port 22, then somehow 'bind' ports 22 and 25 on both ends. So that as far as the school firewall is concerned, I'm using SSH, but I'm actually sending to and from port 25 on each of the machines. Anyone have any suggestions? Or any other alternatives? Preferably, it won't require any modifications to my MUA or MTA setups. (i.e. I won't have to set Evolution to send on port 2525 or some such, nor force exim to listen for connections on port 2525) -- _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Converting ext3-XFS
Tom Badran wrote: On Friday 22 Aug 2003 12:37, Johann Koenig wrote: Which is exactly what i was saying, except i dont know how to make a tarball that retains all file permissions/attributes etc.. Those should be retained by default. Make a directory with a few files with odd permissions, tar it, untar it, and see what happens. Ahh, i had tried this already, but i just did the untar as root and it works. If you extract a tarball as user it seems to set all file ownerships to user.user instead of owner.group This is a Good Thing. Imagine if you could set permissions, attributes, and ownership of untarred files as an arbitrary user. Now imagine you have an unprivileged account on a machine somewhere, and root on your box at home. At home you tar up an archive which includes a suid root file which spawns a (root) shell (or just does something really bad to the system). Now you scp the .tar to the unpriv. machine, untar it, and run your little suid root program. Nice, xfs here i come Tom -- _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Code rights for employees (was Re: SCO identifies code?)
Steve Lamb wrote: So do you think that any artist would sign on and say that x type of work is owned, whole or in part, by the company regardless of whether or not the work was done for that copy or on their own? If they did, that would be their own fault, wouldn't it? It also wouldn't be the first time a starving artist took a deal that turned out to be not so beneficial (think of most of the VH1 Behind the Music specials, or the Prince fiasco of the early 90's). Actually, when was the last time you saw a rock star tell a record label you know, I wrote this song on my own time, so I don't think you really own the rights to it? Granted DC Comics owns the copyright for Sandman and Warner Brothers owns the copyright for Babylon 5 so one can argue that if Neil Gaiman or JMS write stories in those universes they can be considered property of the companies that own them. But the larger argument of th genre cannot be made. IE, JMS still can (and does) write science fiction even though that may compete with the WB property Babylon 5 which is science fiction. He can't call it Babylon 5 or relate it to Babylon 5 but he can write it, sell it, whatever even if it were during the time he was working for WB's project provided it was on his own time. Then apparently, JMS is not forbidden to do so by contract. If he signed a contract that stated he would not do ANY Sci-Fi writing for any other bodies, then he would not be able to write, sell, or whatever, even on his own time. This is an assumption, however, based on your observation that he continues to write science fiction. We don't know for sure if he's contractually forbidden to write science fiction unless we have a copy of his contract :-) It is the same principle here. If an author working on eBay (to keep with the example) goes on to found a competing on-line auction site as long as there is no code share and none of the work was done on eBay's time then eBay has no claim as long as he doesn't call it eBay. IE eBay = title, auction site = genre. Again, this assumes that the author working on eBay did not sign a noncompete contract. If he was a key developer, I would not be surprised if eBay's contract stipulated that should the author leave eBay, that he not work for any other online auction sites (at least for a period of time, or within a geographic area). It's fairly common practice in different industries (heck, I can't talk to my old insurance agent for another couple of months, since he had a noncompete signed with his old insurance company). The arguments made so far are putting words into employment contracts, or assuming that no contract exists. IANAL, but it seems that there are several questions here: 1) Did the employee sign a contract? no: It's likely that the employee retains ownership of everything done on their own time, with possible IP exceptions, depending on copyright law, and IP laws of their state. Yes: 1a) did the contract contain provisions regarding ownership of intellectual property, inventions, copyrights, etc? If the contract did, it is likely that in signing the contract, the employee VOLUNTARILY AGREED that any inventions, code, etc. are the property of their employer, even if done on their own time. 2) What do state and federal laws dictate in the absence of a contract? 3) Can a contract impose ownership restrictions even if state and federal law do not? 4) Can/do state and/or federal law place limits upon what IP, code, copyright, etc. rights may be contractually transferred to a company as part of an employment contract? --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: monitoring bind 9 using mrtg
Gabriel Granger wrote: Hi All, Quick question, has anyone here set up mrtg to monitor dns servers? I'm monitoring sendmail stats and devices satts from the network just would be nice to know what my name servers and doing. I did a quick google but did find anything really helpful. I guess what I'm after is, named.conf options to turn on stats script to collect this data for mrtg and anything else I might need that i might not be aware of - Regards - Gabe I've got a bind-stats setup. I'm pushing the info up to a web page now. The scripts and info will be located at: http://users.2z.net/rpuhek/scripts_public/bind-stats/ --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Spamassassin and procmail
Antony Gelberg wrote: Hi all, I currently use fetchmail and procmail to get and sort my mail. I'd like to use spamassassin as well, however when I add :0fw: spamassassin.lock | /usr/bin/spamassassin to my .procmailrc, it works ok, but then the mail gets delivered to /var/mail/username, rather than following the rest of my procmail recipes. Any ideas on how to alter this behaviour? A :0fw # Or, use spamd and spamd, which is a lot nicer to your machine # under heavy load: #|/usr/bin/spamc -s 50 -d host -p 783 | /usr/bin/spamassassin -P :0: * ^X-Spam-Status: Yes .spam/ ...might work better. The pipe to spamassassin just tags the mail, the second rule is what actually delivers the message. Depending on the version of spamassassin, the -P flag is not needed any more (my 2.55 install's man page says that pipe to stdout is default behavior). You could also sort based on the level of spamminess: #Sort into spam folders based on spamminess. Folder names # have leading number so that Netscape sorts them correctly. :0: * ^X-Spam-Status: Yes #.spam/ { :0: * ^X-Spam-Level: \*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\* .spam.4-very-spammy/ :0: * ^X-Spam-Level: \*\*\*\*\*\*\*\*\*\* .spam.3-pretty-spammy/ :0: * ^X-Spam-Level: \*\*\*\*\*\* .spam.2-somewhat-spammy/ :0: .spam/ } #Drop potential FN here: :0: * ^X-Spam-Level: \*\*\*\* .spam.1-slightly-spammy/ --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: kernel-source-2.6.0-test1?
Noah L. Meyerhans wrote: On Mon, Aug 04, 2003 at 03:29:51PM -0500, Marino Fernandez wrote: I cannot help you with the debian way... that I found more complex than the regular way. Download a pristine kernel (get 2.6.0-test2) cd /usr/src Untar kernel in /usr/src mv linux linux~ ln -s linux-2.6.0-test2 linux make menuconfig or make xconfig make make modules_install Put bzimage in /boot, rename it vmlinuz-2.6.0-test2 Update lilo or grub reboot Please tell me how that is less complex than $ cd kernel source dir $ fakeroot make-kpkg --config menu kernel_image $ sudo dpkg -i ../kernel-image-whatever the version was.deb $ sudo reboot I just don't get it. noah Plus, the Debian way is much easier, IMHO, for compiling on one machine, and installing on another. This is useful in cases where you: 1) Don't want to bog down a machine with compiling a kernel (I like to compile during the day when I'm awake enough to be sure I caught everything... people get mad at me if all the RAM disappears on my spamassassin server, or if I steal all the CPU cycles on my MySQL server). 2) Don't want to put gcc et. al. on a machine (for security or space concerns). 3) Have more than one machine with the same hardware, etc. If only one kernel config is required, why recompile 10 times (or compile once, copy bzImage to all the machines, copy the modules ot all the machines, reboot, discover you forgot system.map on one, and forgot to run LILO on three machines). Very handy to compile on one machine, scp kernal_image.servername.1 servername:/usr/local/src and dpkg -i the image... --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bridged wireless AP and samba
James Goldwater wrote: I have a debian box acting as an AP, with a bog-standard bridge between the wlan0 and eth0. I can browse the internet and see machines in my internal network just fine, but am now unable to connect to samba shares from my windows xp box. I can try \\mysambabox or \\192.168.0.64 and both report the network location cannot be reached. If I connect the windows xp box through the wired interface instead of wireless, all is fine. Through wireless I am able to ping the samba box, ssh to it, connect my imap client to the imap server on it fiine etc, it's just samba not working. I've run nmap on a udp scan of the samba box from the AP and it just seems to hang; running the nmap scan from the samba box on itself comes up with netbios ports pretty quickly. So is there some complexity with briding and UDP, or the hostap driver and UDP? Does anyone have any pointers as to where the problem may lie? I've installed Debian testing with a custom 2.4.20 kernel and the hostap driver; I have a realtek8139 wired card and a Netgear MA311 wireless card. I have no iptables, no wep etc. Thanks for any light you can shed, James. PS I don't know if this is related, but in a previous config I had two subnets (one wireless, one wired) with straightforward nat'ing between them - no other iptables lines. I had identical symptoms. Run a WINS server. IIRC, the problem is that the so called browse master ends up on the other end of the bridge, so you can only see what's on your side. You can either configure the samba box as a WINS server, or just use the XP box (probably does WINS by default). Dig through the samba docs and the comments in the config file for more info. I had a similar problem develop here. All I had to do to fix it was have my DHCP server return a WINS server attribute (and wait around 15 min for windows to figure out what the hell was going on). As for nmap not seeing the ports, check that you're allowing connections from the AP's IP address (in smb.conf, hosts_allow = something?). --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Log rotation
Matthew Daubenspeck wrote: On Tue, Apr 01, 2003 at 01:44:24PM -0500, Derrick 'dman' Hudson wrote: It can. Just be sure to anchor the glob. For example, using /var/log/samba/smb* is really bad because the first rotated file (smb_foo.1) will match as well. The above anchor with *.log prevents ...log.1 from matching. This is where I run into the problem. The files do not have a .log extension. They are in the format of MM.DD. Then anchor appropriatelly: /var/log/samba/smb\.[0-9][0-9]\.[0-9][0-9]\.[0-9][0-9][0-9][0-9] --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Sendmail, TLS, and incorrect hostname
I recently added SASL and TLS support to my SMTP server. Looks to be working great, with one exception. My certificate was generated with the local hostname, instead of the proper FQDN (hostname was the common name of the machine, I need a cert for smtp, since that's how my users connect). Result is that users configured for TLS get a warning that hotnames don't match. I've started digging through the OpenSSL docs. Looks like all I need to do is regenerate the certificate interactively, and specify the desired hostname, overriding the default. Problem is, I'm not sure about any other switches I'll need. Is there a Debian-way to do this nice and cleanly? Has anyone else run into (and solved) the issue? Thanks for any ideas! --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: sync root passwords?
sean finney wrote: um, under whose implementation of passwd? certainly not debian's. please correct me if i'm wrong, but passwd doesn't take as an argument the new password, as it rightly shouldn't. think about what that would mean. anyone with enough sense to run top -c (or it's ps counterpart) could learn your root password. sean My bad... Instead of passwd, you'll need to do something like `echo root:$1 | chpasswd -e`. If having other users see the passwd is a problem (good concern if other people can log in locally) do a crypt on the passwd first before passing to chpasswd, and eliminate the -e. --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] Fried motherboard
probably. I've never had a power supply die on me. maybe I'm lucky. I've even plugged a power supply that was set to 110V into a 220V circut, tripped a breaker, but the power supply itself was fine(haven't tried going the other way around). I've done that (220V setting on a supply, plugged into a 110V outlet). Just acted like a dead supply. Had the darn thing swapped out of the PC and a new one in place before I thought to look at the switch on the back. I believe the supply was alright. Hard to believe you didn't fry more in your case... the transformer should have dumped 2X the normal voltage into the regulator/switching portion of the supply, which can be very bad. --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: spamassassin installation help
Stephen Gran wrote: It is. Dman's page is excellent, and if you follow the instructions, you'll have a very nicely working setup very quickly. The only thing I would recommend doing is (if you're using either woody or sarge) is backporting the version from unstable - I have gotten much better results, with far lower false positives/negatives than before. This is easily done by downloading the source, cd into the toplevel source directory and typing `fakeroot debian/rules binary`. You may need to first install the build-depends, which will be listed in debian/control. Or, add the following to /etc/apt/apt.conf: APT::Default-Release woody; and add a mirror for unstable to sources.list then do apt-get update apt-get install unstable/spamassassin should get you the latest packaged SA. Someone is also building .debs of the CVS snapshots, if you really want to be bleeding edge. See the SA archives for details. --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: spamassassin installation help
This is doable, but the problem is with spamc - it depends on a newer version of libc6 than what is in Woody. This means that if you want to install the spamc from unstable (which you must to satisfy the Depends: for spamassassin) you have to upgrade libc6 as well. I run Sid on one of my boxes, without any real problems, but I wouldn't want it running on a mail server. I found it much easier to backport for Woody than to mix distros like this, but YMMV. Ahh, that's right. Forgot about that. I'm running spamd on a dedicated box. Didn't have any trouble with the libc6 upgrade, but wouldn't have done that on the mailserver itself. --Rich _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT question sorry but i need salution fast ....
Kirk Strauser wrote: At 2002-06-11T18:14:24Z, faisal gillani [EMAIL PROTECTED] writes: well sorry fro this but i couldent find any other place to ask this question i need the salution very badly this will gratly benift me well i have 2 ethernet networks running which i want to connect but the distance between them is above 400 meters .. so this is way beyond the normal lan hardware .. I'm not saying that this is an ideal solution, but could you install a cheap 4-port switch (not hub!) every 100 meters? Since switches generates signals themselves, rather than just passing the original electrical signal, that might extend your range sufficiently. Anyone care to tell me if or why this is a bad idea? You are correct about switches regenerating the signal, but why not just connect the two networks with fiber. Fiber will cost the same (or possibly less) to run between the two locations, will eliminate the need for multiple switches, and as a bonus will eliminate issues like grounding concerns (I'm guessing the two locations are in seperate buildings, if so, grounding and lightening protection become a hassle with copper). Toss a media converter on each end (or a fiber card for a router or switch), and you're set. As a further bonus, you'll easily be able to upgrade to higher speed networking between the ethernets just by replacing the equipment on each end. --Rich _ Rich Puhek ETN Systems Inc. _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT question sorry but i need salution fast ....
Hall Stevenson wrote: I'm not saying that this is an ideal solution, but could you install a cheap 4-port switch (not hub!) every 100 meters? Anyone care to tell me if or why this is a bad idea? A device is a device, right ?? Whether it's a network card, hub, or switch, keep the distance between them under 100 meters. Sounds logical ;-) Wrong... The following: (PC#1)-HUB-HUB--(PC#2) is NOT the same as: (PC#1)-switch--switch-(PC#2) The problem is, a hub just electrically connects the signal (handling crossover, etc). A Switch will completely regenerate the signal on the other port(s). The overall length applies to each segment. The top diagram is one segment, which can have an *overall* length of 100 meters, not 400 meters as your analysis might imply. The bottom diagram can have 100m between the first PC and the first switch, 100m between the two switches, and 100m between the 2nd switch and the 2nd PC (assuming 10 base-T Ethernet). The issue, IIRC, has to do with the minimum size of an Ethernet frame, the speed at which the frame propagates down the physical wire, and the need for all devices to be properly able to sense a collision. Basicly, if the overall length of the network is too long, PC#1 could begin transmitting a frame while listening for a collision. PC#2, if far enough away, could start transmitting a frame at nearly the same time, also listening for a collision. If the network is long enough to introduce enough delay that the start of the frame transmitted by PC#1 did not reach PC#2 before PC#2 was done transmitting (and if the frame from PC#2 did not reach PC#1 before PC#1 was done transmitting) neither side would detect a collision. The result would be a corrupted packet, which would hopefully be handled at a higher layer. I'm a test engineer by trade, not a communications engineer, so my explanation may be a bit off. Anyone have a copy of IEEE 802.3 laying around? I assume that deals with this problem in more detail. --Rich _ Rich Puhek ETN Systems Inc. _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Problem: wall-clock jumping like Mexican bean
Jonathan Matthews wrote: I think it's possibly a dodgy RTC on the motherboard. I saw this exact complaint come up on linux-kernel a while ago, and someone mentioned that it was specific to a certain brand of mobo. No URLs for you, I'm afraid, but if you google a bit, I'm sure it'll appear. But I thought that the RTC was ignored after boot? If so, a bad RTC would explain a clock that comes up bad on boot every now and then, but not why a running system would suddenly shift time. --Rich _ Rich Puhek ETN Systems Inc. _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: debian-beer (was Re: wrapping [was: Re: disable paragraphflows in mozilla?])
Gary Turner wrote: Guiness absolutely sets the standard, especially if you can find it on tap and nitrogen charged. I've been favorably impressed by some other Irish beers (but not recently nor often enough to name names). ...But with Guiness you need to use a fork to drink your beer! Kinda thick stuff... German imports to the US are good to very good for the most part. Haven't found an outstanding brew (yet). German beer doesn't really count if it leaves the country. Gotta try one of the local brews, non-pasteurized, local flavor. Very good stuff. Sam Adams Boston Lager is my favorite for day to day sipping. Sam Adams has some good stuff. Leinenkeugel (sp?) from Wisconsin is pretty good too. Beer trivia: According to a rumor I once heard (so I have no idea of the truth to it) Rice-based beers like Budweiser do not naturally produce any head, so soap is added to produce the familiar bubbles. -- Rich _ Rich Puhek ETN Systems Inc. _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sendmail: to recompile or not recompile
Jamin Cleveland wrote: I am currently attempting to set up a virtual user table in Sendmail 8.12 on my Woody machine. According to sendmail.org the standard way to do this is to build the virtual user table by running the following: makemap dbm /etc/mail/virtusertable sourcefile When I do that I get the this: makemap: Need to recompile with -DNDBM for dbm support So my question is: How do you all do it? Do you recompile? Or is there a Debian workaround for building Sendmail virtuser db's? All advice is much appreciated. Give a shot at running 'sendmailconfig' now that you have a virtusertable. The Debian way will create a makefile (assuming you have the relevant entries in your sendmail.mc file) allowing you to just make in your sendmail dir when updating files... --Rich _ Rich Puhek ETN Systems Inc. _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [Question] Harddisk Error!!! - damage
ben wrote: so that e2fsck on an unmounted partition is okay? but to check a mounted partition it should be efsck? or what? sorry, i haven't been following the thread. i'm not sure if it matters but since using the 2.4.17 kernel on my machine , crashes nessecitating the check haven't happened, at all. Umm no... First off, fsck just calls the appropriate filesystem-specific program, which happens to be e2fsck for most of us. I've never heard of efsck... possible typo? Second, You always want to fsck unmounted partitions only. If nothing else, trying to fsck a mounted partition produces the following message: WARNING!!! Running e2fsck on a mounted filesystem may cause SEVERE filesystem damage. Mr. Ts'o (the guy who wrote a large portion of the ext2 stuff in Linux) knows much more about filesystems in general, and ext2 in particular, than I do, so I listen to that error message. Finally, It's not just crashes that will cause the system to run an fsck at boot time. There's a mount counter that will automatically fsck after so many reboots (or was it mounts?). I recall the number being set to 20. You can also force checking upon reboot by supplying the -F option to the shutdown command. --Rich _ Rich Puhek ETN Systems Inc. _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Maildir performance mutt (was Re: Someone tell me the secret of mutt)
dman wrote: | With 1000+ messages (and 5000+ is pretty easily attainable), performance | on opening a folder sucks. I assume it's because mutt has to do an | fopen() on each file, scan headers, group output, and sort it. The | result is a several-seconds (sometimes 10-20) on opening large folders. | | Is there any way to speed this up? More memory for disk cache? I've found that the first time I open a large maildir folder the disk will crank for a while. I expect it would do just as bad if it was mbox, probably worse. The next time it is real fast because the disk is cached in memory. At any rate, updating the folder (ie deleting one message) is much faster. I'm guessing it's the sheer number of messages in the directory... ext2 starts to get kinda slow when you go over 1000 messages. This was the first reason I switched to Maildirs BTW... had a server with about 3000 users. Access to /var/spool/mail was very slow as a result... I'm hoping ReiserFS starts to solidify as a file system... That's not supposed to have the performance hit with 1000s of files in one directory, but I still don't trust it for a mail spool. I'm eventually going to try reiser for the host status directory... that's another area that gets a large number of files, and I don't care if it gets corrupted (worst case I'll just wipe it out and let it re-populate). --Rich _ Rich Puhek ETN Systems Inc. _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Potato vs. Woody
Alan Poulton wrote: I was wondering.. should I be sticking with Potato or switch to Woody? (snip) If I upgrade to Woody, how do I do it? Would I just edit /etc/apt/sources.list, replace all stable with woody, then apt-get update; apt-get upgrade ? A friend of mine, who knows way more about Debian and Linux than I do, told me that when she upgraded to Woody some time ago, there were many security holes in her system. Do I need to worry about this still? Depending on your comfort level, you can do either one. I think you'll find that woody is quite stable, especially if you're on the i386 architecture, so you can comfortably upgrade. Otherwise, you can upgrade specific packages fairly easily (see informaion on pinning releases in the archives of this list). NO! you cannot do apt-get upgrade. You are correct in your upgrade procedure except that you must edit sources.list, apt-get update, then apt-get dist-upgrade. The dist-upgrade takes care of juggling the complicated dependancies and conflicts that happen with upgrading the entire ball of wax. The upgrade option seems intuitive, but won't quite work... I've done this on about eight different systems, with few problems (had one box with an old-format /etc/lilo.conf file that caused me some difficulty when I rebuilt the kernel, but that's about it). Security holes shouldn't be an issue... you're dealing with the newest software. What issues did your friend have? Good luck! --Rich _ Rich Puhek ETN Systems Inc. _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [HELP] RAID5 IN DEBIAN
axacheng wrote: Hello List : I wanna migrate my OS (SuSE7.3 To Debian) and wanna use raid5 to accese date Everybody knows that where would i find VERY useful document or HOWTO about raid5 in Debian? I had already known some URL as follow: http://www.linuxdoc.org/HOWTO/Boot+Root+Raid+LILO-3.html http://www.linuxgazette.com/issue17/raid.html Very Thanks -- Do you want to do software RAID 5 or hardware RAID 5? that will make a big difference in the necessary approach. --Rich _ Rich Puhek ETN Systems Inc. _ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: mistake: apt-get and debian 2.2 - UNSUBSCRIBE
Heh... This should lead to an even-tempered discussion. --Rich Jonathan Tabaco wrote: UNSUBSCRIBE Sorry to all those who are inconvenienced by this but I have tried multiple times to get removed from the mailing list. I even sent in the confirm CONFIRM u03180819391769 PLEASE REMOVE ME MANUALLY, as your automated system doesn't work. I will unfortunately have to resort to spamming the Debian-user mailing list until I am off the list, as all the e-mails I am STILL receiving are now spam to me. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: High powered Debian advocacy?
Noah Sombrero wrote: On Fri, 22 Mar 2002 23:09:07 +1100, you wrote: If I was you, I would concentrate on ease of upgrade (as mentioned before by someone, apt-get), Apt-get is a great tool, however it insists on installing everything in /usr. Which means that it is difficult to make use of extra hard drives. Apt installs there because that's where the FHS (http://www.pathname.com/fhs/) says things should go. That's actually an advantage, knowing that a typical installation will use X MB of space under /usr, and Y MB under /var, and so on. You're also able to make consistent decisions about what directories can be NFS-mounted, and which ones must be local (to anyone who accidentally messed up /lib, this is much appreciated). Rather than making it difficult to use extra hard drives, this easily allows you to dedicate a partition to /usr, possibly one to /usr/local (if you install a lot of stuff that's not under apt's control). --Rich _ Rich Puhek ETN Systems Inc. _
Re: ext3 not mounting (ghost mounts created)
Thedore Knab wrote: What did I do wrong to prevent the proper mounting of multiple /var/* directories ? SNIP, fstab trimmed for readability... # /etc/fstab: static file system information. # # file system mount point type options dump pass /dev/hda2 / ext3defaults,errors=remount-ro 0 1 /dev/hda3 noneswapsw 0 0 proc/proc procdefaults0 0 #/dev/fd0 /floppy autodefaults,user,noauto0 0 #/dev/cdrom /cdrom iso9660 defaults,ro,user,noauto 0 0 /dev/hda1 /boot ext3 rw 0 2 /dev/hda5 /home ext3 rw 0 2 /dev/hda6 /var/log ext3 rw0 2 #source for idea #http://www.stahl.bau.tu-bs.de/~hildeb/postfix/ext3.shtml /dev/hda8 /var/spool ext3 rw,data=journal,noatime 0 2 /dev/hda7 /var ext3 rw0 2 /dev/hda9 /usr ext3 rw0 2 There's your problem... you can't mount /var/spool or /var/log until /var is mounted, so you need to change your fstab so that you mount /var first. This is done through the pass field (the number in the last column of /etc/fstab). See how swap is pass 0, then / is pass 1, then the other partitions are on pass 2. Change so that /var/spool and /var/log are mounted in pass 3, and life is good. The man pages for mount and fstab aren't very clear about this gotcha, but it makes sense if you think about it... let's say you didn't mount /var, and tried to mount /var/local/some/path/name. Where would/should that end up? Should your machine magically create the /var/local/some/path/name? What if you later mounted /var/local/some which had directories path/ foo/ and bar/? etc. etc. when you manually remounted your spool and log partitions, you had a mounted and running /var partition, that's why remounting manually cleared that up... --Rich _ Rich Puhek ETN Systems Inc. _
Re: File permissions on nfs mounted directory (Woody)
Dave Sherohman wrote: On Wed, Mar 13, 2002 at 05:20:29PM +0100, Alexis Kotte wrote: movemail: Permission denied for /var/mail/kotte What does `ls -l /var/mail/kotte` show as the file's owner? File permissions across NFS only work if your UID is the same on both machines or you run an additional daemon (whose name escapes me) to handle mapping the client's UID to the server's. ugidd is the daemon IIRC. I'm surpried it works as root... must have no-root-squash set. --Rich _ Rich Puhek ETN Systems Inc. _
Re: Kernel question
Charles Parker wrote: I don't see a 2.4 kernel in the stable Debian packages. I've been told it's usually NOT a good idea to take a kernel directly from kernel.org because it won't contain the customizations provided by your distribution, and things will likely break. I've also been told to use a 2.4 kernel for better USB support (which started this whole thing). What's the story, and what do the cognoscenti recommend? Thanx - Charlie I don't believe Debian customizes the kernel at all. There are pre-compiled versions available with different options set (see things like kernel-image-2.2.19-compact and kernel-image-2.2.19-ide I'd suggest running 2.4.17 or 2.4.18 if you can... seems to be working fine. One warning (might have been discussed earlier on the thread): don't use a 2.4.x kernel with Debian Potato (a.k.a stable a.k.a 2.2.r5). There were some issues there. --Rich _ Rich Puhek ETN Systems Inc. _
Re: /tmp size
Richard Otte wrote: I recently discovered that /tmp on my machine is rather small, around 50mb. I was trying to use xcdroast, but was unable to extract an audio cd to /tmp because it wasn't big enough. This is strange, because I probably have 50gb empty on my hard drive. I'm wondering if /tmp is a separate partition (is this my swap partition?) or why it won't use up the empty disk space. I don't know how to find out the exact limits on /tmp, except by what xcdroast told me. Can anyone explain to me what is going on, and what to do. Thanks, Ric How did you determine that /tmp was 50mb? It may be a seperate partition, use mount, or df -h to find out. See if xcdroast will allow you to specify where to put the temp files. That may do it for you. No, it is not your swap partition, your swap partition won't show up in your file lists. --Rich _ Rich Puhek ETN Systems Inc. _
Re: 3c509 setup disk?
Ron Johnson wrote: On Sun, 24 Feb 2002 22:09:33 -0800 (PST) Paul 'Baloo' Johnson [EMAIL PROTECTED] wrote: Anybody got the 3c509 etherdisk 2 on a bootable floppy that someone can send me the dd image for? And how can I find out what IRQs and IO addresses are taken, ala Microsoft's MSD or the device manager? $ cat /proc/interrupts CPU0 0: 119855853 XT-PIC timer 1: 369901 XT-PIC keyboard 2: 0 XT-PIC cascade 4:7484798 XT-PIC eth0 10: 154430 XT-PIC ide2 12:3236750 XT-PIC PS/2 Mouse 14:1456522 XT-PIC ide0 15:2366439 XT-PIC ide1 NMI: 0 ERR: 0 $ cat /proc/ioports -001f : dma1 0020-003f : pic1 There's also the 3c5x9utils package for the card itself. That will allow you to change the settings if you have to, similar to the config program that used to ship with those cards. Might be necessary if you've got a '509 sitting around, and have no idea if it's set to the typical IRQ 10/IO 0x300 (well, typical in the DOS 6.22/Win3.11 setups we were doing way back when) or not. --Rich _ Rich Puhek ETN Systems Inc. _
Re: (OT) chicago
Mark S. Reglewski wrote: On Thu, Feb 21, 2002 at 08:25:02PM -0600, Rich Puhek wrote: I do love driving the Chicago area freeways though. Watch Road Warrior before you go and you'll be ready for the evening traffic. Nothing like following someone doing 85 on 355 (speed limit supposedly 55 MPH) and seeing someone even crazier rocket by on the right while burning wreckage sits on the right. Just make sure you've got a lot of quarters (or get Ipass). Heh, heh. You drive the limit? Speed limits on the expressways (never called freeways by the natives) and tollways are *only* recommendations. Except around holiday weekends, the roads are only lightly patrolled by state troopers. And their speeder-catching hideouts are well known to regular travelers. Haven't done 55 in twenty years. Zero tickets in twenty years, without using a radar detector. That should say something about speed enforcement around here. Ummm, sure, I always drive the speed limit. Especially if any members of the Illinois law enforcement community are on the list :-) I agree with the assesment of the speed limits. Just add about 15 MPH and you've got the true limit, aside from the occational traffic tangles at stuff like hwy 53. Now if I could only learn the local nicknames. I turn on the TV in the hotel or the radio in the car thinking that I'll get info on how traffic is moving on 90, 94, 55, etc. I never hear that, but they will tell you how fast the Ike is moving inbound, or that there's an accident on the tristate. And I've still never found the post office. Everyone in Chicago seems to go to the same post office, since they'll tell you how long it takes to get there! Of course, I get the fun of heading back home through Wisconsin. Been lucky so far, but they've got a reputation when it comes to speeding, and pricey tickets to boot (over $100 for 66 MPH in a 65 zone, and some areas they really mean it). --Rich -- _ Rich Puhek ETN Systems Inc. _
Re: (OT) chicago
Mark S. Reglewski wrote: On Thu, Feb 21, 2002 at 02:38:21PM -0800, Karsten M. Self wrote: on Thu, Feb 21, 2002, Mark S. Reglewski ([EMAIL PROTECTED]) wrote: On Wed, 20 Feb 2002, dman wrote: The details aren't finished yet, but it looks like I'll be working in Chicago (Elgin, actually) next quarter. Is anyone here in the Chicago area? [snip] Elgin is about forty miles NW of the Loop. . . . If you don't have a car out here, you won't actually be *in* Chicago very much, though it's an easy drive on I-90 if you do. I'd recommend Metra Rail over driving -- far less hassle, no parking problems, drops you straight into the Loop, and you can hack for an hour on the way in and out ;-) ...just beware the mid-day service blackout -- 11 am - 1pm or so. Otherwise, coverage is great. Don't know if there is service from Elgin per se, but a co-worker commutes into town via Metra from Crystal Lake, which is just spitting distance from Elgin. Well, easy driving distance. If you were mad enough to try to walk to Crystal Lake from Elgin, you'd show up in the Loop picking corn stalks out of your hair. I've been out that area. I believe Metra has an Elgin stop (right along the river, about 2 miles past the riverboat casino). There's a pretty good bus network to get you to the station, or there's a park and ride. Got a real damn long red light at one of the intersections there though. If you rail into the city, there are many money-losing bus and train routes to get around, but bus and train schedules get sparser and sparser with each passing year as the annual losses erode service. You could use cabs to get around the city, but you might have to speak languages other than English, C, or Python to make yourself understood. And you *may* find out that you have a better idea of where you are going than the cabbie. Also understand that it is next to *impossible* to get a cab pickup in some Chicago neighborhoods (mine is an example). So you might take a cab to a destination and find out that you can't get a pickup to go back where you came from. So a Chicago-newbie should plan any excursions from Elgin carefully. My opinion: the flexibility of having a car makes driving a clear win for the Chicago novice, even with the cost of parking and the hassle of the drive figured in. Depends on time of year. In the summer, you can get to almost anything interesting on the free shuttles. Biggest hassle with commuting in is the hike across the river from the train station to get to the red line if you're heading to a Cubbies game. If the weather's decent, you can really do some hiking downtown. We've hoofed it both ways from the train station to Navy Pier. Everyone else was looking in the stores, at ESPN, at the N*SYNC buses that blew by us. I was trying to figure out where the Ameritech NAP was and trying to decide if I've seen an uglier building than IBM. Occupational hazard I guess. Biggest hassle is the late train schedule back out of the city. If I remember correctly, we could hit an 6:00 PM or a 10:30 PM train or something like that to head back to Palatine (one stop before Elgin). Gotta decide to head back to the suburbs early or stay pretty late, not much of a happy medium on the schedule. I do love driving the Chigaco area freeways though. Watch Road Warrior before you go and you'll be ready for the evening traffic. Nothing like following someone doing 85 on 355 (speed limit supposedly 55 MPH) and seeing someone even crazier rocket by on the right while burning wreakage sits on the right. Just make sure you've got a lot of quarters (or get Ipass). --Rich _ Rich Puhek ETN Systems Inc. _
Re: strange vi
Ron Johnson wrote: On Mon, 18 Feb 2002 22:44:03 -0600 Rich Puhek [EMAIL PROTECTED] wrote: Ron Johnson wrote: Is is console or xterm? What's the value of $TERM? either console or SSH session. $TERM is vt100. Looked at my termcap, which seems to match what I had before. try setting TERM to linux That worked. Thanks. Strange though, on my potato boxes, TERM=vt100, and the arrows work in nvi. Oh well... I'm happy I can arrow around now. I can handle the traditional vi keys alright, just mentally I tend to forget what mode I'm in if I'm not switching between arrows and home row. Thanks! --Rich _ Rich Puhek ETN Systems Inc. _
Re: strange vi
Alex Malinovich wrote: Oh, I'm this is just begging to start a flame war, but I have to say it... :) Get a REAL editor like Emacs! :) *runs and hides* heh... was wondering when that would start. I tried using Emacs. works best if you put a rock on your control key. I think Emacs guys must have their left thumb on the wrong side of their hand or something... hitting that darn key all the time :-) --Rich _ Rich Puhek ETN Systems Inc. _
Re: Latest *stable* 2.4.x kernel ?
Adam Bogacki wrote: Hi, I apt-downloaded the 2.4.14 kernel and all seemed to go well for two sessions Try 2.4.17. See prior discussions about 2.4.14 stability issues. before it crashed and I cannot login beyond a low runlevel. Ker 2.2.x would not recognise the Promise controller on which my removable HD was mounted so a Are you sure you just didn't need to cook your own 2.2.x kernel? I thought I recalled seeing Promise cards in the 2.2.x kernels. 2.4.x ker was needed. I have worked around the problem by physically moving the empty drive to a PCI slot on the first controller and installed 2.2r4 ... which leaves the problem of rescuing the corrupted drive. I'd like to install asafe and stable 2.4.x kernel onto /dev/hdb so that I could have access to it. Any ideas ? Adam Bogacki, [EMAIL PROTECTED] (current ISP) [EMAIL PROTECTED] (forwarding) If by 2.2r4 you mean the Debian version, there's a portion of your problem. Without the bunk packages, potato (Debian release 2.2rx) doesn't play well with 2.4.x kernels. Upgrade to Debian 2.3 (woody) if you need 2.4.x kernels. Installing a 2.4.x kernel on your hdb could work, as long as you can boot from hdb. I'd suggest investigating if your particular Promise card can be compiled into the 2.2.x kernel. If not, might be best to apt-get dist-upgrade to woody, then install 2.4.17 kernel. --Rich _ Rich Puhek ETN Systems Inc. _
Re: strange vi
dman wrote: On Mon, Feb 18, 2002 at 06:59:07PM -0800, justin cunningham wrote: | Michel Loos wrote: | | On Mon, Feb 18, 2002, justin cunningham wrote: | | dpkg -l | grep vi shows nvi 1.79-20. the potato's show 1.79-16a.1. | | | | ok, so if I use the letter keys instead how to I get to the end | | of a line of data then press return to enter a new line without | | taking the last character with me? | | | | Like usual A | | Sigh... ok, you've made it clear it's time pick up that vi book behind | me. Yep :-). Alternatively, you can simply press 'o' to open a new line below the current line. Then the cursor will be placed on the first column of the new, blank, line. The 'O' (note the case difference) command does the same thing, except the blank line is opened above the current line. -D ...ok, all that's cool, but nothing so far has answered the original question: how does one get nvi to recognize the arrow keys? Things worked fine until I upgraded to woody, now nvi isn't playing nice. All the man page has to say is The cursor arrow keys should work, too. anyone with any ideas? --Rich _ Rich Puhek ETN Systems Inc. _
Re: strange vi
Ron Johnson wrote: Is is console or xterm? What's the value of $TERM? either console or SSH session. $TERM is vt100. Looked at my termcap, which seems to match what I had before. --Rich _ Rich Puhek ETN Systems Inc. _
Re: Debian TVIO like PVR
nate wrote: Dave Carrigan said: Paul McHale [EMAIL PROTECTED] writes: I know the Tivo doesn't have much horse power. They are case in point for a design which is just fast enough. They appeared to have spared every expense. It is an awesome unit. Just saying, I don't think they have a 600MHz processor ... Could be completely wrong. As you can see, it is a 54MHz PowerPC chip with 16MB of memory; not a powerhouse by any means. it is a powerhorse when it comes to video capture though. i have a tivo and i also do a ton of video capture from it(i haven't hacked it, not going to touch the insides of it until tivo goes under and i dont have any other choice). i have a 1.3ghz athlon with wintv pci and 768MB ram with a dual 20GB raid0 array. tivo can easily encode in ~640x480(approx what NTSC is). wintv on my athlon struggles to encode at 320x240. infact i make it a point NOT to move my mouse cursor when capturing otherwise it will drop frames. CPU isn't really a factor as the cpu hovers between 15-45%. just the slightest movement of the mouse and it drops a frameor 2. its a very fragile setup, i restart X frequently and i disable opengl in my nvidia drivers when i want to capture, and re enable it when i want to play unreal tournament. otherwise it drops more frames. I/O is not a limitation either, the video is encoded at a compression level that drops the bitrate to ~90kbyte/second. so adding a 50-disk raid0 array wont do anything. One thing to remember is that the Tivo has a few minor advantages: 1) Your serial port/USB port probably has a fairly high-priority interrupt, this is why your mouse movement will cause your processor to stop working on the encoding. 2) The Tivo probably has hardware dedicated to the specific purpose of video processing (a FPGA/CPLD type device, or maybe even an ASIC), where as your PC has to make do with a general-purpose microprocessor. 3) Your microprocessor is a CISC device (i386 class, etc are all CISC processors). I would guess that it is quite likely that the different execution times for the different instructions would be a challange in a realtime (or near realtime) environment. The PowerPC is a RISC chip, which (among many other things...) means that every opcode on the chip will average about the same (around 1.5 to 1.7 cycles/instruction IIRC). It may not be a real big deal, but then again, might be interesting to see how performance is with Debian on PPC. --Rich _ Rich Puhek ETN Systems Inc. _
Re: Advice on aacraid, Dell 1U server
Rory Campbell-Lange wrote: I'm thinking of ordering a Dell PowerEdge 1550 1U rackmount server for the office to act as a backup server. The 1U form factor is important to us. For 2200 pounds one gets a Pentium III/133MHz 256K cache, 512Mb RAM, 108Gb 10,000rpm Ultra3 SCSI box with 3 years on-site service. For another approx 140 pounds we can get hardware raid, using aacraid v1.0.6. 3 questions: How much of the 108Gb are we likely to lose to RAIDing? (Presumably RAID5?) Basicly, you will lose one hard drive to the RAID 5 parity data, so if you have 10 drives of 10.8 GB each, you'll lose 10.8 GB. If you have 3 drives of 36GB each, you'll lose 36GB. Is aacraid reliable? We are using Mylex on the VALinux boxes. I've been happy with the PERC3 Dell puts in the 2450 series boxes. Others also appear to have been happy with the PERC3 and PERC2 cards. Any recommendations for this box, apart from the price? Strongly consider getting the RAID card... It should increase performance (depending on your application), and will definately increase reliability. On the other hand if backup server means that it's normally just sitting there waiting for a main server to die, you may not be that concerned. Thanks for any help. Rory -- Rich _ Rich Puhek ETN Systems Inc. _
Re: managing multiple machines
How about setting up a custom /etc/ for each client... doing something like: On NFS server: == export/etc/ client1/ client2/ client3/ ... On clients: === mount server:export/etc/clientn /etc/ (where n is the client number...) /etc isn't huge, so having multiple copies isn't a huge deal. You could even keep common files in export/etc and have the individual client directories contain symlinks (er... or would it need to be hard links... getting ahead of myself) to them, easing the burden of common configuration. --Rich David Wright wrote: Several posters have pointed out that I can get DHCP to assign IPs based on MAC, which goes a long way toward solving my problem. I guess if I share hosts using ldap, I can still give each machine a unique name, too. Now the only problem I can think of with sharing /usr and /etc is that my file server's /etc will contain stuff like /etc/exports, which will end up telling every machine to be a file server. H. I'll go and re-read the Diskless-HOWTO, but if anyone knows a workaround I'd still appreciate hearing from you. _ Rich Puhek ETN Systems Inc. _
Re: proper format for /etc/networks
nate, I believe you just add an alias for the given network (so to speak). Let's say you have a local ethernet on 10.0.0.0 and your accounting dept. has their own subnet at 192.168.50.0. You might want to do: localnet 10.0.0.0 accounting 192.168.50.0 Then, instead of seeing 10.0.0.0 in your routing table, you'll see the word localnet. Snort shouldn't care about this... look at the DEBIAN_SNORT_HOME_NET option in /etc/snort/snort.conf for defining your home network. Other snort options are similar (and more flexible, since snort knows about subnet masks). --Rich nate wrote: whats the proper format for /etc/networks in debian? i've never had to use this file before but it looks like i do now in order to use certain options with snort. theres no manpage on it, i looked at the solaris version of networks and tried to do the file that way but it didn't seem to work. network_name 192.168.50.0 from the solaris manpage: Network numbers may be specified in the conventional dot (`.') notation using the inet_network routine from the Internet address manipulation library, inet(7P). Network names may contain any printable character other than a field delimiter, NEWLINE, or comment character. [..] The network database does not support subnet masks in gen- eral, so getnetbyaddr(3N) cannot differentiate between net- works of 11.128.0.0/255.192.0.0 and 11.128.0.0/255.240.0.0. nate _ Rich Puhek ETN Systems Inc. _
Re: NFS tries following remote symlinks as if local!
Adam, Perhaps a hard link would work in the manner you're looking for? It does appear that there is no way to disable this behavior (as applied to symlinks). Note that NFS servers will return the pathname of a symbolic link, regardless of whether it is absolute or relative (the server really doesn't care... just happens that one answer starts with a /). You might want to pick up the O'Reilly NFS book (Managing NFS and NIS). It's kinda handy for the little NFS oddities like this. Don't be too hard on people who seem to underestimate your knowledge of a subject... often it's easiest to start from a certain point in an explanation to be certain everyone's on the same page... That's also very handy for others following along in the list or in the archives of a list. --Rich Adam Warner wrote: On Tue, 2001-11-20 at 14:35, nate wrote: Adam Warner said: Hi all, I've come across this crazy problem and I hope someone knows what's going on. I am using kernel-level NFS. Debian unstable. 2.4.14. I have exported / I can mount the remote filesystem on my client machine no problem. But if I try to change to a remote symlinked directory I get, for example: this is normal and expected behavior. all a symlink is is a pointer. it points to a file. even if a file doesn't exist. Nate as you can gather from my follow up post I have a higher level of understanding than that. The remote symlink points to the correct location on the remote system. Having always used Samba for remote file accesses (so I could also access the files using a Windows machine) I expected NFS to follow remote symlinks on the remote filesystem (just like when I navigate a file system when ssh'ed into a remote box). I've learned an important lesson. Since I primarily use symlinks to hide the underlying file system layout I just expected the symlinks to continue to act transparently. Though I can see the point of Dave's Good Thing response. If I am backing up the files on my remote computer I will have to be careful to use the -d option in cp. Otherwise I will also back up parts of my local computer as the remote symlinks translate to local paths on my computer. I just tried converting my absolute symlinks to relative ones. But it still doesn't work (because it appears I don't have the ability to descend into mounts). Here's the scenario: I'm in the NFS mount point on my local computer. I ls -l and see this particular link. home - mnt/data/home So if I cd to home I should continue down the remote filesystem. Note that there is no slash (/) before mnt. This is a relative path. But it doesn't happen. Instead the mnt/data directory is empty. It appears NFS doesn't descend into mount points either. I've tried adding the nohide option to disable this but it made no difference. Regards, Adam _ Rich Puhek ETN Systems Inc. _
Re: Making a TCP or UDP or Unux Socket Server Listen on a port
From man services: services is a plain ASCII file providing a mapping between friendly textual names for internet services, and their underlying assigned port numbers and protocol types. So... that file is just a way to translate between port names and numbers, not a magic way to make a program listen on that port. You are correct in that your entry would associate a.out with port 9889, in that once you get the program running, a netstat would show a.out listening on port a.out instead of port 9889. When you run your server program, try running a netstat. Do you see something on port 9889? If so, look at how the client is contacting it. If not, your server program isn't binding to a socket. see man 2 socket, man 7 socket, and a bunch of other man pages for further information. --Rich [EMAIL PROTECTED] wrote: I am having Server Programs (for TCP,UDP,Unix Sockets) , that listen on a port number say , 9888 at 10.10.1.4 and they are being contacted by a Client (for TCP,UDP,Unix Sockets) Linux machine from 10.10.1.1 . Both macines are on the LAN , nothing more nothing less . We have not yet gone for any higher stuff like DNS , etc . So , basically 10.10.1.4 and 10.10.1.1 are not on any Linux network , but simply they identify and respond on the LAN. My Clients and Server (socket programs) do not communicate at all on this network . Please guide me on what I should do . My line of thinking is : Add some entry to the /etc/services specifying the port , but this is what I tried and failed :- # At the bottom of /etc/services :- tcp 9888/tcp a.out 9889/tcp Please help me get these programs to communicate . Warm regards, ([EMAIL PROTECTED]) Shyam _ Rich Puhek ETN Systems Inc. _
Re: Unable to remove mysql-server.
I've had similar problems before. I got around them by creating the following /etc/init.d/mysql (or whatever...) file: #!/bin/sh exit 0; That way, the removal script is happy. The same trick works for upgrades as well. --Rich Chrys wrote: Hi, I tried to uninstall mysql-server to compile the original binary file, but it looks like I have a problem :-( Here is what apt-get shows me: mobi:~# apt-get remove mysql-server Reading Package Lists... Done Building Dependency Tree... Done The following packages will be REMOVED: mysql-server 0 packages upgraded, 0 newly installed, 1 to remove and 0 not upgraded. 1 packages not fully installed or removed. Need to get 0B of archives. After unpacking 1987kB will be freed. Do you want to continue? [Y/n] y (Reading database ... 23902 files and directories currently installed.) Removing mysql-server ... /var/lib/dpkg/info/mysql-server.prerm: /etc/init.d/mysql: No such file or directory dpkg: error processing mysql-server (--remove): subprocess pre-removal script returned error exit status 1 chown: /usr/share/mysql: No such file or directory dpkg: error while cleaning up: subprocess post-installation script returned error exit status 1 Errors were encountered while processing: mysql-server E: Sub-process /usr/bin/dpkg returned an error code (1) so I tried with the -f option, but: mobi:~# apt-get remove -f mysql-server Reading Package Lists... Done Building Dependency Tree... Done The following packages will be REMOVED: mysql-server 0 packages upgraded, 0 newly installed, 1 to remove and 0 not upgraded. 1 packages not fully installed or removed. Need to get 0B of archives. After unpacking 1987kB will be freed. Do you want to continue? [Y/n] y (Reading database ... 23902 files and directories currently installed.) Removing mysql-server ... /var/lib/dpkg/info/mysql-server.prerm: /etc/init.d/mysql: No such file or directory dpkg: error processing mysql-server (--remove): subprocess pre-removal script returned error exit status 1 chown: /usr/share/mysql: No such file or directory dpkg: error while cleaning up: subprocess post-installation script returned error exit status 1 Errors were encountered while processing: mysql-server E: Sub-process /usr/bin/dpkg returned an error code (1) What I realy need is version 3.23, but it's still in the testing section. Can you tell tell me how to install version 3.23 ? Chrys [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: Raid adaptec 2100S
Not out of the box. You'll need to create a custom boot floppy for the install. If I can ever find mine, I'll post an image, but I lost the darn thing in the blizzard of paper on my desk. Once you get it running, things work great from what I've seen. --Rich Hi im not on this list but im currently wondering are Adaptecs 2100S raid controller supported ? if you know please let me know Sincerely Anders Johansson [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: [debian-user] install with 3 partitions
Um... You can't mount /home, /tmp, /usr, and /var on hda5. You can only mount a partition at one place at a time. You _can_ make one partition a piece for home, tmp, usr, and var and do something like: mount /dev/hda5 /home mount /dev/hda6 /usr mount /dev/hda7 /var ...I'm not sure you want to mount /tmp seperately. To be on the safe side... might want to leave that where it is. [EMAIL PROTECTED] wrote: Yall, Following Installing Debian GNU/Linux 2.2 For Intel x86 Chapter 4 Partitioning Your Hard Drive As recommended, I have layed down three partitions: hda1 Primary, bootable 50MB ext2 hda5 Logical, 1GB ext2 hda6 Logical, 233MB swap I want: / - hda1 /home, /tmp, /usr, /var, - dha5 From the Potato 2.2rev2 installer, I choose Mount a Previously-Initialized Partion after Initialize a Linux Partition, but it identifies that there are no unmounted partions. I tried to use Initialize a Linux Partition with 'other', multiple ',' seperated (guess) arguments, but the installer really did not like me then. Somebody must have a good way of doing what I am trying to do. Best regards, Lloyd D Budd _ Rich Puhek ETN Systems Inc. _
Re: exporting /usr NFS for small network
Richard Cobbe wrote: Lo, on Wednesday, October 24, joe golden did write: I am getting tired of updating 7 machines. I have home directories exported NFS for our network (with minimal security concerns) and this seems to work fine. My question is how do I export usr NFS. What are the configuration issues. How must disks be partitioned? Are all the X/card and monitor specifics guaranteed to be all in /etc? If any of these specs are in usr I'll have hashed spaghetti flying everywhere in no time. So long as the distributions in question follow the FHS (see http://www.pathname.com/ for details), then sharing /usr like this is fine. According to the FHS, /usr is for shareable, read-only data; config files belong in /etc. I'm pretty sure Debian qualifies here. I don't think partitioning is relevant to this situation; NFS exports files and directories based on the directory structure, not the physical devices. I was looking for advice on pitfalls to avoid. The only thing I'm not sure how to do is keep /usr/local local to each machine, even though /usr is mounted across the network. This may or may not be a requirement in your situation, however. I don't think you need to worry about that. From FHS 2.2, section 4.9.1: The /usr/local hierarchy is for use by the system administrator when installing software locally.It needs to be safe from being overwritten when the system software is updated. It may be used for programs and data that are shareable amongst a group of hosts, but not found in /usr. So you shouldn't need to keep /usr/local seperate on each machine. If your particular installation required such an arrangement, you could mount /usr from an NFS server, and then you should be able to mount /usr/local from a local partition if required. If a need for different /usr/local is present in your environment, you will have a concern here. A possibility is to have your NFS server export something like: /export/usr, /export/usr/local/flavor1, and /export/usr/local/flavor2. Then, on your machines you would do the following: on machines that need to have flavor 1 of local: mount nfsserver:/export/usr /usr mount nfsserver:/export/local/flavor1 /usr/local on machines that need to have flavor 2 of local: mount nfsserver:/export/usr /usr mount nfsserver:/export/local/flavor2 /usr/local (bad example, since I believe you'd have to export HTH, Richard -- Rich _ Rich Puhek ETN Systems Inc. _
Re: Trouble installing Potato on Dell Poweredge 2450
Stuart, Do you have the Perc 2 or the Perc 3? I believe you are probably running the Perc 3 if you have a 2450. Very nice card, but it's a little bit of a challange to get it working. See the following for info: http://lists.debian.org/debian-user/2001/debian-user-200101/msg00453.html Also see Kevin's page at: http://www.merilus.com/~kevin/aacraid.html --Rich Stuart Allen wrote: I am having trouble installing Potato on a Dell Poweredge 2450. The problem appears to be the Perc 2/DC RAID controller. When booting off the rescue floppy, the system hangs after the following three lines of output: megaraid: v1.11 (Aug 23, 2000) megaraid: found 0x8086 : 0x1960:idx 0:bus 0:slot 2:func 1 scsi2: Found MegaRAID controller at 0xe0004008, IRQ: 11 From what I have found on the net, support for this controller is built into the standard kernel. Perhaps I need some special boot parameters? One I have tried without success is aic7xxx=no_probe. Regards, Stuart -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: network number
Hmmm, Your network number should be correct, and you're right about what the netmask *should* be. Apparantly an installation script got horribly confused about netblocks. Try hand-editing your network file with the correct information. --Rich Stan Brown wrote: I upgraded a Debian machien this weeknd, and now it wants a bit more information in the /etc/network files. It wants network number I'm confused by this. Heres what I have IP 170.85.109.24 netmask 255.255.255.128 I put in 170.85.109.0 for a netwokr number, but this must be wrong based upon what the broadcast adress of the interface becomes. It should be 170.85.109.127, but instead it's 170.85.255.255 -- Stan Brown [EMAIL PROTECTED] 843-745-3154 Charleston SC. -- Windows 98: n. useless extension to a minor patch release for 32-bit extensions and a graphical shell for a 16-bit patch to an 8-bit operating system originally coded for a 4-bit microprocessor, written by a 2-bit company that can't stand for 1 bit of competition. - (c) 2000 Stan Brown. Redistribution via the Microsoft Network is prohibited. _ Rich Puhek ETN Systems Inc. _
Re: network number
Dave Sherohman wrote: On Mon, Oct 08, 2001 at 10:44:45AM -0400, Stan Brown wrote: Heres what I have IP 170.85.109.24 netmask 255.255.255.128 I put in 170.85.109.0 for a netwokr number, but this must be wrong based upon what the broadcast adress of the interface becomes. It should be 170.85.109.127, but instead it's 170.85.255.255 Add the following lines to /etc/network/interfaces: netmask 255.255.255.128 broadcast 170.85.109.127 What's happening is that 170.x.x.x falls within the address range originally allocated for Class B netblocks (networks with a netmask of 255.255.0.0), so ifconfig is assuming that your network is a Class B. If you explicitly specify the correct information, it will be used instead. Dave, True, that would be the correct netmask if he was in the old 170.85.0.0 class B, but doesn't the network address take precedence in determining the netmask (as far as the configuration scripts go, not as far as IP addressing goes)? Sounds like Stan did enter in his network address correctly, so it's strange that his broadcast got set to the default Class B mask. --Rich _ Rich Puhek ETN Systems Inc. _
Re: Needing a random number generator for scripting
Dave gets my vote for the best answer. I'm a bit biased in that I like a meat grinder approach with sed, awk, cat, and a lot of pipesigns, but you've got to admit, his little oneliner script is much tidier than the here's how it's done in my favorite language answers. Dave Thayer wrote: awk 'BEGIN {srand()} {print rand() \t $0 }' unshuffled.m3u \ | sort | cut -f2 shuffled.m3u _ Rich Puhek ETN Systems Inc. _
Re: OT: Looking for 10/100 ISA NIC's for Linux project
Gack! $70? That seems a bit pricy for an ISA card. 10/100 cards in an ISA slot? That's a bit odd... On the off chance you meant to type PCI, not ISA, I'd use the 3c905 card from 3com. They're supported, reliable, and they're $45/each if you buy them in a 25-pack from Datacomm. If it's a school, you can probably get an education discount from a reseller (not sure if a charter school will qualify for 3com's GEP program, or whatever it's called now). Search around out there, and you can probably find someone selling a lot of 100 used cards for much less per card. Or, if you really want ISA and 10 meg, go with the trusty old 3c509. Again, you should be able to find plenty of used ones there for well under $70 each. --Rich John Purser wrote: Hello, I'm looking for a source for nearly 100 10/100 ISA Ethernet cards for a linux network I'm helping with at a local Charter School. I've found some for around $70 but was hoping to cut that price in half if I could. Anybody got some leads or recommend a particular card to buy and/or stay away from? Thanks, John Purser _ Rich Puhek ETN Systems Inc. _
Re: Something fishy is going on
Warning: New Distributed Denial of Service attack on the loose! Synopsis: In a dastardly clever (yet simple) scheme, a new DDOS is attaching Linux newsgroups at an increasing rate. Artfully designed to capitalize on user paranoia following the massive hype surounding the Code Red family of worms, this program simply startles the user by having a fish swim across their desktop at some unpredictable time. Upon receiving this signal, the PC user will respond in one of three modes, depending on the time of day: Sleep mode: If the victim is infected late at night, the user will attribute the apparition to too much caffeene and not enough sleep. Result: user sleeps indefinately. Propagation mode: If the user is infected during the workday, the user will attempt to reproduce the phenomanon, possibly on neighboring systems. Attack mode: If inected during the late afternoon or evening, the user will transfer a SMTP message to a mailing list. The result is to trigger a small transfer of data on said list as other clients attempt to handle the data. Although the attack mode is of low traffic, we anticipate that the cumulative result of many thousands of clients will eventually bring the Internet to a halt. The client behavior after the attack is currently unresearched. A group is studying the possibilty of constructing a fishbowl, so that more detailed analysis may be conducted. Suggested Snort rules: alert tcp any any - $HOME_NET 25 (msg:Wanda Infection detected!; content:fish;) alert tcp any any - $HOME_NET 25 (msg:Wanda DDOS response detected!; content:Gnome Easter Egg;) Remedy: Applying procmail rules to filter the initiating email may help limit the response to the email probe message. Unfortunately, this will not be effective unless adopted on a wide scale. /funny --Rich [EMAIL PROTECTED] wrote: Jeez, this has popped up on the list A LOT lately ... check the archives. It's an apparently harmless Gnome Easter Egg. Poor Wanda has come in for a lot of paranoia the last month or so! :) Glenn Becker -- _ Rich Puhek ETN Systems Inc. _
Re: forgot root password on head- and keyboardless machine *blush*
Heh... got screwed by the AT/PS2 keyboard problem huh? Happens here at work all the time. Got a few trusty old AT-style machines and a few newer ATX cases with PS2 keyboards and mice. I'd either bite the bullet and buy a keyboard (they're not that expensive...and yours is probably looking kinda beat up now isn't it?) or I'd buy an adaptor for the AT to PS2 style. You should be able to find an adaptor at Radio Shack, Circuit City, etc. Or, raid a PC boneyard at a computer store or large business around you. They should be overflowing with junky keyboards. You probably won't care if the spacebar needs to be hit a few times, just as long as it's somewhat functional. --Rich Stig Brautaset wrote: * dman [EMAIL PROTECTED] spake thus: On Tue, Aug 21, 2001 at 06:02:07PM +0100, Stig Brautaset wrote: | I have a head- and keyboardless machine running debian potato that I | used to log into with ssh. Now I have forgotten the password. *blush* The easiest way is to borrow a head and keyboard from somewhere and boot into single user mode. Hmm, now if you had a way to reboot ... I know that this is a solution but I don't have a keyboard. I have a screen I could use, but I really don't want to buy a new keyboard just to do this... (my friends all have ps/2 keyboards, whilst my machine uses the old din-style). Regards, Stig -- www.brautaset.org _ Rich Puhek ETN Systems Inc. _
Re: Killing your keyb.controller... was: Re: forgot root password on head- and keyboardless machine *blush*
Emil Pedersen wrote: Just to add some more noice to the list ;-) [statement] Hot-plugging keyboards works _MOST_ of the time. It is true for at least PS/2-keyboard, since the only machine I've managed to destroy this way is an Digital Celebris 590. My other machines with PS/2 have survived, so for ps2 types the statement is true. When it comes to DIN-keyboards, I have NOT been able to kill any machine this way. Finaly, since I have one more Celebris 590 I _could_ verify that these machines DO die when keyboard is hot-swapped, but I think it might be a waste of computers if I succeed... ;-) Regards, Emil Btw. Have anyone mantioned those adapters for some dollar or two? I'm sure it's possible to physically damage a machine if: a) you managed to short some wires together (Mac types have been warned not to plug/unplug ADB devices since circa 1984 for this reason). b) in the process of fumbling around, you managed to send a stout jolt of electrostatic discharge into the keyboard port. Physical damage should not be confused with an OS (or BIOS?) that gets confused by the sudden absense/presence of a keyboard (back to Mac, ADB addresses the devices semi-dynamically at startup IIRC, hence a potential ID conflict if you add a device later). BTW: yes, a couple of people (including myself) have mentioned the adapters. --Rich _ Rich Puhek ETN Systems Inc. _
Re: META QUESTION: how to read a bulk list (and stay happy :-)
dman wrote: On Tue, Aug 14, 2001 at 10:24:31PM -0500, Rich Puhek wrote: | Personally, I'm stuck with an NT box at work, so I end up using Netscape Not to fear -- mutt works great with cygwin (just patch attachment.c to use binary mode for opening files or else M$ will screw up the streams)! I haven't tried fetchmail or procmail though (I ssh to the school's system like I'm doing now which is Solaris). -D I suppose I could try that. Problem is, I'm really an elm guy for text mail reader (not too surprising to see a :x at the bottom of my Windows emails). Of course, now I switched to Maildir/ format, so I'm back to pine. --Rich _ Rich Puhek ETN Systems Inc. _
Re: debian-user-digest Digest V101 #1132
Kevin C. Smith wrote: I recently saw a list become almost useless because of name calling, snipes, and such. Just let it pass, please. Just wanted to plead for the turning of the cheak strategy. Thanks. Or, in modern Internet terms, a turn on the bozo filter strategy. I find those kinds of emails amusing at times. When I tire of it, it's killfile time. The user can yell and scream to be removed from the list, argue that the Internet won't be truly free until everyone runs their own root server with their own TLDs, can have a long-running debate about the personal hygene habits of Bill Gates, or any other offence that's landed them there, and it won't bother me at all. --Rich _ Rich Puhek ETN Systems Inc. _
Re: META QUESTION: how to read a bulk list (and stay happy :-)
John Galt wrote: .procmailrc recipie: :0: * ^Resent-From: debian-user@lists.debian.org debian-user About 99.9995% effective. On Wed, 15 Aug 2001, Gaelle T. Morin wrote: Um, how about using: X-Mailing-List: debian-user@lists.debian.org instead of the Resent-From? I believe that's what that header is intended for. Of course, not every mailing list has that feature, but for those that do, you're going to have better luck than with Resent-From:, To:, CC:, and similar headers. Personally, I'm stuck with an NT box at work, so I end up using Netscape for email reading. It's a PITA, but better than anything else I've found (Eudora bugs me for some reason, The Bat! costs money, and don't even talk to me about Outlook). I parse all mailing lists (Some of which are even more active than deb-user!) into seperate folders. Then, I can scan through the folders quickly, either by date or by thread. Some things are handy in curses-based environments, but a quick scroll bar in Windows (or, X, I suppose) works pretty well for scanning mailing lists, expecially once you get the feel for the list. --Rich -- _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _
Re: Sendmail directories... /var/state/sendmail...
Ian, Those directories are used to store state information about hosts your sendmail has tried to contact. By storing the info here, sendmail can check to see what the last state was of a host it may have recently tried to connect. The idea is that if you have multiple queue runners delivering mail, each runner knows if a host is already known to be down, and won't waste its time trying to deliver there. You are correct in that it's kind of a reverse hash, with top level domains under /var/state/sendmail (or whatever directory you specified with HostStatusDirectory) and that subdomains exist under there. The bat book has more info on the guts of the files you find there, if it's any concern to you. It also suggests that you try enabling and disabling persistent host status to see which is best for performance on your system. Anyone have any comments on running that on a filesystem other than ext2? I've noticed that a df of that directory (or even a simple ls of the com subdirectory) can take a long time on ext2. Perhaps reiserfs or something like that might be better suited for such a directory? --Rich Ian Perry wrote: Hi, I have noticed some wierd directories on one of our mail servers, and so I checked the others. hey are on all of them in some varying degree or another. /var/state/sendmail then ae. at. au. etc (SNIP) I can see that these are in reverse order... such as acay.com.au Is this a quick DNS lookup or what purpose do they serve ? Many Thanks Ian _ Rich Puhek ETN Systems Inc. _
Re: centrally managed bookmarks for multiple users, accessible from everywhere?
The easiest way to have a shared bookmark is to just have a links page... just raw HTML. Your users could connect right to it by typing it in, could make it their home page, etc. A more advanced step would be to write something like a PERL CGI or a PHP program to allow the users to update the page. Even more advanced (Perhaps overkill) is to store the info in a MySQL database, and have the page generate automagically. MySQL would not actually be the program listening, your central machine would have a webserver running for the users to connect to. If you want to integrate the bookmarks into a browser, just have the clients synchronize their bookmarks file with the master file, either manually or through a cron job. The actual transfer could be ftp, wget, rsync, etc. If you've got NFS mounting, you could have some neat tricks with an NFS mount and symlinks, etc... --Rich Walter Tautz wrote: I was wondering if it would be possible to keep bookmarks on a central machine where one could access and change them via any browser than can ``connect'' to the bookmark server. To elaborate: It seems to me that one could adapt a standard database program like mysql that listens to connections from the network to centrally manage bookmarks from a multitude of users in such a way they could add and delete entries from a browser running anywhere. Perhaps the newer opensource browsers will have a module like this? -- _ Rich Puhek ETN Systems Inc. _
Re: Adaptec Raid
Michael, You need to patch your kernel to add the necessary drivers that support the Adaptec card. Adaptec hasn't gotten the message that Linux is not RedHat, so it's not a trivial process. See: http://people.FreeBSD.org/~gibbs/linux/ for the necessary patches for the kernel. If you're trying to boot off of the array you'll need to create a custom boot floppy. If your boot partition (and probably your root partition) are running off of the other SCSI controller, you can start the install, then patch the kernel. That part (custom disk) gets tricky, but it's doable. Basicly, you'll compile the kernel on another box, make a copy of your boot disk, and replace the kernel image on the boot disk with your new image. I've got the 2100S running fine for the last few months on a box I did this too. Once I get off my lazy butt I'll post images of my boot disk somewhere to save the headache of creating your own. Good luck! --Rich Michael Blood wrote: Hello, I am attempting to install Debian 2.2 on a box with an Adaptec 2100s Raid 5 Controller. After creating the disk array using the utilities that come with the adaptec card. All that being done I run the debian installation. When using the default installation the process reports that it can not find any hard drives to partition. when I view the script it shows that it can find another scsi adapter that I also have in the box but it does not find the raid controller. I have already tried to run at the boot prompt boot: linux aha152x=iobase but I get the exact some results. Does any one have suggestions. Thanks, Michael Blood -- _ Rich Puhek ETN Systems Inc. _
Re: Adaptec Raid
Err... Sorry, wrong Adaptec-related web site. The correct site is: http://www.aurore.net/source/ The dpt patches are the ones you need. --Rich Rich Puhek wrote: Michael, You need to patch your kernel to add the necessary drivers that support the Adaptec card. Adaptec hasn't gotten the message that Linux is not RedHat, so it's not a trivial process. See: http://people.FreeBSD.org/~gibbs/linux/ for the necessary patches for the kernel. -- _ Rich Puhek ETN Systems Inc. _
Re: HELP! Can't boot no more - When Kernel Upgrades Go Bad
John, You probably forgot to include support for ext2 filesystems in your new kernel. To try to get your machine running for now, boot your machine, and hold down the shift key (IIRC...). You should see the machine stop at boot: Hitting TAB should give you a list of kernels you can chose (again, by memory is foggy here, so it may be something else...). Hopefully you have your old kernel image laying around (if you rebuilt your kernel the Debian way (tm), you will have one). Chose the old kernel, and see if that boots ok. If that all fails, go ahead and re-install, just make sure that you have ext2 enabled. --Rich John Griffiths wrote: Hi everyone, hoping someone can help me here I followed the instructions on http://www.fs.tum.de/~bunk/kernel-24.html for upgrading 2.2r3 to the 2.4 kernel. I thought i'd followed the isntructions to the letter, but when i rebooted all seemed well until halfway through the reboot when it stopped. the last good line in the bootup says Linux IP multicast router 0.06 plus PIM-SM then we get to the bad stuff request_module[block-major-8: Root fs not mounted VFS: Cannot open root device 801 or 08:01 Please append a correct root= bot option Kernel panic: VFS: Unable to mount root fs on 08:01 and that's where it ends. any help to either a) get the system as-is to boot or b) re-install and get 2.4 pon the right way would be appreciated. I'd be very grateful for any help John -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: [very very OT] noisy power supply
Man... keep that fan blowing, and hard. If the fan is making machanical noise, replace it. If you're not brave enough to slice the warranty void if removed... no user servicable parts inside... sticker on the PS, replace the entire power supply. If the noice is coming from the air movement, make sure there's no restrictions (like extra cables hanging behind your case) that are restricting airflow. If the noice really bugs you, heck, just clip the leads to the fan, let your system cook. You may or may not see premature death of your CPU and hard drives as a result of the increased heat. If you're running the machine at home, and plan to upgrade again in a year, you may be fine, and may not care. If the machine is important to you, you should be happy with the noise (assiming it's wind noise, not bearing noise), since that means your machine is running happily along. As for PWM modulation of a fan... PWM is more often used for things like small DC motors in electromechanical systems... the print head in your desktop inkjet is probably driven by a PWM-contolled motor, since it needs precise controll over positioning, and needs to start, stop, and change position rapidly. It doesn't sound very appropriate for a fan, which is either on, off, or possibly running at a reduced speed. There's probably a few EEs with motor-controller experience out there who could say more... --Rich Joost Kooij wrote: On Thu, Jul 12, 2001 at 09:18:21AM +0200, thomas wrote: So my idea is: I open the power supply, flip the fan so that it blows cool air from outside into the case, voila, much less noise. Is this a good idea or rather stupid? bad idea. you will move all the hot air in the case. if your man enough take you PSU apart and mod your fans to 7V, that will make it almost unhearable! IIRC you need to switch the fan on and off in rapid succession, it's called pulse width modulation. Presumably, it's better for the electronics inside the fan. Cheers, -- _ Rich Puhek ETN Systems Inc. _
Re: Reiserfs and disk spindown: separate /var partition?
heh heh heh... my butt still hurts from when I tried the very same thing. That was when I upgraded my first Debian Box from hamm or slink. I figured the box was so hosed... I might as well do a clean install. Lucky it only took you an hour to repair the damage :-) Amazing what a dynamically-linked program can do to you! --Rich Matthew Sackman wrote: In the interests of maximising hard-disc usage, I once moved the contents of /lib to another partition (non-root) and created a sym-link. An hour later I had repaired the damage - it really don't like it!!! :-( -- _ Rich Puhek ETN Systems Inc. _
Re: Reiserfs and disk spindown: separate /var partition?
Craig and Christian, The Filesystem Hierarchy Standard, or FHS (http://www.pathname.com/fhs/) helps decide what can be mounted on a different filesystem. Some excerpts from the 2.2 standard: /bin contains commands that ...are required when no other filesystems are mounted (e.g. in single user mode). /sbin contains binaries essential for booting, restoring, recovering, and/or repairing the system in addition to the binaries in /bin. The FHS also encourages small root partitions, which would seem to imply that it's a good idea to give /usr a seperate partition. IMHO, in virtually all cases, /var should definately be in its own partition, since otherwise you risk filling the root partition with log messages, etc. --Rich Craig Dickson wrote: (snip) - Getting it right for the boot process (should I make / reiser or ext2?): I assume I need to have /etc, /bin and /sbin on the root partition. What about /usr? It's never occurred to me to put /etc, /bin, or /sbin on any filesystem other than the root. I would guess that that would not be a good idea, since the boot process might (?) need access to them before it gets around to mounting all the default filesystems (for one thing, the mount command itself lives in /bin), but I'm not really sure. /usr can certainly be a separate filesystem, though I've never done so except when I wanted it to be on a separate hard disk from the root. Typically, I leave /usr in the root filesystem, though /usr/local or /usr/share might be separate. Craig -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: [users] Re: Time to fight for our beloved DEB format!
Dave Sherohman wrote: SNIP... And, to me at least, `xdm stop` obviously means shut xdm down, while `init 3` has no readily apparent relationship to X or xdm unless you're bringing outside knowledge with you. Bingo. That's what the arguements boils down to. What is most important IMHO, is the ability to customize runlevels to a local condition. You may be running RH 6.x on a laptop that needs to have X running with one network configuration in runlevel 5, and console only with a different configuration in runlevel 3, with levels 2 and 4 having various other combinations. I may have Debian 2.2.r3 running a webserver in runlevel 3, and I may have the machine setup as a backup MX server and a web server in runlevel 5, without X in either. --Rich _ Rich Puhek ETN Systems Inc. _
Re: V = I * R and the rest (Re: OT: C++ Newbie and KDE/QT)
And ignore the abacus and the slide rule? For shame! we must remember to study our roots! Remember heck week from one of the later Revenge of the Nerds movies? --Rich D-Man wrote: That's a good idea, but we really ought to start with vacuum tubes, now shouldn't we .5 wink? -D -- _ Rich Puhek ETN Systems Inc. _
Re: [users] Re: a quickie
Also, if you're running, oh, say, and email or web server on you server rack, you might be concerned if the server were rebooted, since the service would be unavaliable for a while. On a heavily-loaded email server with a large (ext2) mail partition with quota support enabled, the checkquota proces alone will be intolerably long for the middle of the day. My suggestion: purchase a KVM. In my case, I've got a low-end 4-port KVM on my racks. There are about 12 machines there, but most are running Debian, so I rarely need a console connection on those. I leave windows machines and our voice mail server attached to ports 1-3. Port 4 I have as a roamer and attach to whichever Debian box I need at the moment (had a machine that tended to lock up and segfault for instance). By adding a KVM, not only have I eliminated the possibility of rebooting a Linux machine when I intended to log into an NT server, I have also largely emiminated having to rummage around the back side of the rack swapping cables. I hadn't realized that was a problem until one of our techs went through like a bull in a china shop and knocked the power cord loose from my email server. Now, since I've made the NT boxes all a pushbutton away, I'm the only one who ever needs to swap cables. Since I had to clean up the mess whenever the mail server got abruptly booted, I am much more careful than the people who caused the crashes. As for pride: SNMP station: $ uptime 5:38pm up 272 days, 19:12, 1 user, load average: 0.22, 0.59, 0.60 $ Utility web server/general use server: $ uptime 5:41pm up 205 days, 18:26, 1 user, load average: 1.03, 1.03, 1.00 $ mail server: $ uptime 5:42pm up 285 days, 23:40, 1 user, load average: 0.01, 0.02, 0.00 $ Web server: $ uptime 5:43pm up 285 days, 23:38, 2 users, load average: 0.00, 0.00, 0.00 $ ... and people ask why we run Debian :-) --Rich Paul Wright wrote: so what does 114 days of uptime buy you? A sense of pride. does it matter that much??? To me, no. To others, maybe. -- Paul T. Wright [EMAIL PROTECTED] -currently seeking employment- -- -- _ Rich Puhek ETN Systems Inc. _
Re: turning off exim on port 25
Ummm, maybe it's just too late at night and I'm missing something, but I think you can do what you want by editing /etc/inetd.conf, and removing or commenting out the following line: smtp stream tcp nowait mail/usr/sbin/exim exim -bs --Rich Bryan Walton wrote: This may be a better question for another list. I am building a firewall for my home LAN. I have exim configured for local delivery only (the only thing I want it to do is move email from root to another userid). Even though I have configured exim for only local delivery, the exim daemon is still listening on port 25. Is there a flag I can use when starting up Exim so that it won't listen on port 25? Thanks, Bryan -- Bryan K. WaltonNetwork Operations Center Analyst Berbee...putting the E in businesshttp://www.berbee.com/ GPG fingerprint: BF68 340D A650 E2D7 86B9 FED5 DDFF 3EEE 3229 7B5D -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: turning off exim on port 25
Yes, Exim will still deliver from the queue (there's a cron job to run every 30 minutes), and exim will still send outgoing email if needed. I use exim on any of my machines that will not be receiving mail for a domain. By eliminating the smtp line from inetd.conf, I don't show up with an active port 25 to tempt spammers. Output of cron jobs and the like will still be passed on to my smarthost. --Rich Jim Breton wrote: On Wed, May 23, 2001 at 11:33:40PM -0700, Eric N. Valor wrote: That pretty much turns off exim altogether. Actually the script in /etc/init.d/ will start exim in stand-alone mode if you disable the listener in inetd.conf. So you will still have it listening on 25/tcp. While effective for disabling the Port 25 listen, it doesn't allow Bryan to use exim for his purposes. I think he's also using it in daemon mode rather than being run from inetd. I'm not sure whether exim will still do deliveries from the queue if you disable the tcp listener (I don't use exim), but if it does, I'd suggest shutting it off altogether. Just put an exit 0 at the top of the script. (Again I'm not sure if exim will still work correctly after that, and I don't have a box handy with exim on it to test... so try it out.) -- Jim B. [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: login problem!
Log in as root. Look to see if you have a /etc/nologin file or an /etc/nologin.boot file. If you're seeing the Sytem bootup in progress - please wait message, that means that the nologin files are there, and only root is allowed to login. If those files are there, something went kinda wrong with your last boot process. You didn't jump the gun and try logging in before your system had completed starting up did you? --Rich J. Ramón Fdez wrote: Hi all, When I try login in my debian 2.4 as normal user, system say: login: jramon Sytem bootup in progress - please wait Password: *** Login incorrect I put the correct password, but it doesn't woork. However, I can login as root successfully. Where is the mistake? Thanks -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: GERMAN INVASION!!!
cat oldmessage | tr :German: :English: newmessage Isn't working for me... is my tr broken? Just a little Monday humor! --Rich MaD dUCK wrote: what's going on??? martin; (greetings from the heart of the sun.) \ echo mailto: !#^.*|tr * mailto:; [EMAIL PROTECTED] -- de gustibus non est disputandum. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: smbfs crashes after inactivity
One question/suggestion: Is samba running from inetd or as a standalone server? That's probably not directly applicable to your situation (now that I fully read your email...), but may not hurt. My W2K memory is fuzzy, but I recall that W2K is setup to time out and essentially disconnect from the server after 30 min or so, then transparantly reconnect if you use the share again. I have a hunch that this is related to your problem. You might want to check the server setup to see if there's a switch to correct/change that. --Rich Robert Hawkey wrote: I know this probably isn't the best list for this question but... We've got several machines set up in our office running Debian, RedHat and Slackware Linux that all mount a share on our Windows 2000 server. If any of the machines don't use the share for longer than a half hour or so Samba crashes and you can no longer access the share. If you do an ls on the share it will say ls: Input/Output error. When the crash occurs you cannot umount the mounted folder, all that happens when you type umount /sambashare is umount runs, and quits with no messages, but when you try to remount the share it tells you that the share is already mounted. The only way to get the share back is to reboot the machine. Does anyone know why this is happening? Here's the command we use to mount the share: mount -t smbfs -o username=user,password=password,rw,fmask=0777,dmask=0777 //192.168.0.2/SHARE NAME folder on our machines Thanks, Robert Hawkey _ Rich Puhek ETN Systems Inc. _
Re: 3c905C Drivers
Was is still broken that recently? I thought that potato boot disks worked ok with the 905c, and that it was broken with slink. Of course, I get confused, since I still have a bunch of 3c905b cards floating around. As for a new kernel, It's not too bad to do a boot disk with a custom kernel... of course, that can be kind of a chicken and egg problem for someone without an existing installation around! Thanks for the clarification on the version. --Rich Simon Law wrote: The 905C support was broken in 2.2.18pre21, which ships with potato. The best way to fix that is via a new kernel, but that may be messy to get on the new machine without the network. Simon On Mon, 14 May 2001, Rich Puhek wrote: Also note that if you're installing an older version of Debian, the 3c509c won't work. The older driver only worked up to the 3c509b. I'm not sure when exactly things changed, but if you're using the latest disk images you're ok. --Rich Jason Majors wrote: The 3c59x kernel module covers that card. On Sun, May 13, 2001 at 12:14:33PM -0700, The Reutzels scribbled... I want to get my NIC working and the only drivers that I found for the 3c905C are not for Debian. Could anybody help point me to the right place to get these drivers so I can get my network up and running. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _ -- _ Rich Puhek ETN Systems Inc. _
Re: CPAN, woody and perl?
Try setting your CPAN to ask for dependancies instead of automatically installing them. That way, you can skip the perl-5.6.1 portion... maybe anyhow. --Rich :x Robert L. Harris wrote: I just ran a perl -MCPAN and did the install Bundle::CPAN and it's trying to install and compile perl-5.6.1... I'd rather not do this as I'd like to keep the perl package nice and clean in the .deb format.. Thoughts? :wq! --- Robert L. Harris| Micros~1 : Senior System Engineer |For when quality, reliability at RnD Consulting | and security just aren't \_ that important! DISCLAIMER: These are MY OPINIONS ALONE. I speak for no-one else. FYI: perl -e 'print $i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10);' -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: dual NICs
Adaptec makes a 4-port PCI card, Intel has some two port cards. I'm sure there are others out there as well. If I remember correctly, the Adaptec unfortunately took an IRQ for each port, which was a bit of a pain. --Rich Matthew Sackman wrote: Hay all. Does anyone have any knowledge of a network card that has two independant eth ports on it? The reason I ask is that I've gotta get 4 eth ports into a server squashed into a 2U rack which means I only have 3 expansion cards available... I look forward to hearing from you! Matthew -- Matthew Sackman Nottingham, ENGLAND Using Debian/GNU Linux Enjoying computing -- _ Rich Puhek ETN Systems Inc. _
Re: Sendmail and local addresses
Jason, Try: MASQUERADE_AS(whizzird.net) in your sendmail.mc file. That should rewrite your outgoing email as if it all came from whizzird.net instead of the FQDN of the machine. --Rich Jason Majors wrote: I have several machines, one acts as a mailserver, with an MX entry and all. The others are clients that know to use the server as a smarthost. When I send mail to a local account, I get an error from the smarthost server saying that the address [EMAIL PROTECTED] can't be found. But when I add the @whizzird.net to the address it works fine. These boxes are running the sendmail from testing (the one from stable stopped working on me). I tried to use the sendmail address rewriting mini-HOWTO's advice to allow it to send local mail, but the sendmail I have doesn't like the format of the file for some reason. Is there an easy way to send local mail with sendmail? Thanks, Jason -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: 3c905C Drivers
Also note that if you're installing an older version of Debian, the 3c509c won't work. The older driver only worked up to the 3c509b. I'm not sure when exactly things changed, but if you're using the latest disk images you're ok. --Rich Jason Majors wrote: The 3c59x kernel module covers that card. On Sun, May 13, 2001 at 12:14:33PM -0700, The Reutzels scribbled... I want to get my NIC working and the only drivers that I found for the 3c905C are not for Debian. Could anybody help point me to the right place to get these drivers so I can get my network up and running. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- _ Rich Puhek ETN Systems Inc. _
Re: Linksys EtherFast NIC, full duplex?
Karsten M. Self wrote: on Sun, May 06, 2001 at 08:58:33AM -0500, Jack ([EMAIL PROTECTED]) wrote: Hi, I have two machine connected using crossover cable and LinkSys EtherFast 10/100 Cards. (same card on both machine). I only get 4-5Mbyte/sec when ftp or nfs between each other. Both machine are running Woody. I am sure it's not the best it can get with those NICs. I can get 9-10Mbytes/sec if I boot one machine into freebsd. here is ifconfig from freebsd: dc0: flags=8843UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST mtu 1500 inet 192.168.0.2 netmask 0xff00 broadcast 192.168.0.255 inet6 fe80::203:6dff:fe1b:b60e%dc0 prefixlen 64 scopeid 0x2 ether 00:03:6d:1b:b6:0e media: autoselect (100baseTX full-duplex) status: active supported media: autoselect 100baseTX full-duplex 100baseTX 10baseT/UTP full-duplex 10baseT/UTP none (I do not know how to get duplex status on debian, help me?) I just tested it again (using 130M big file): ncftp reports: (debian) -- (freebsd) 8.24 MB/s (debian) -- (freebsd) 9.54 MB/s I can only get 60% of that when have both to be Debian. Both debian are using the driver compiled from the source comes in floppy(4.1 version) Haven't seen a response. My understanding is that running NICs through a hub (non-switched) network results in half-duplex operation. Only switched networks are full-duplex. But I don't know what I'm talking about. -- Karsten M. Self kmself@ix.netcom.comhttp://kmself.home.netcom.com/ What part of Gestalt don't you understand? There is no K5 cabal http://gestalt-system.sourceforge.net/ http://www.kuro5hin.org Part 1.2Type: application/pgp-signature You are on the right track. In half-duplex, the card basicly listens on its receive line while it's talking on the transmit line. Normally, in a non-switched, half-duplex, environment, only one card may transmit at any one time. If the card hears anyone else transmit during the time it's transmitting, it treats that as a collision. It will transmit a signal indicating a collision (which the hub will relay to all ports), and sleeps for a random time interval (the holdoff time is based on an interesting algorithm that makes the possible interval wider, depending on the number of collisions the card has seen). In a simple case of two network cards talking directly to each other (or a network card talking to a properly configured switch), both cards may transmit at once. Since the transmit and recieve pairs are seperate physical media, the transmission of one card won't interfere with the transmission of the other. To allow this, we put the card (and the ports on a switch) into full-duplex mode, which basicly means just go on talking if the card at the other end is talking too or ignore collisions. What can become an issue is whether the card and the switch properly autoconfigure, or recognise that they're attached in a configuration that will allow them to safely operate in full-duplex mode. As another consideration for the performace issues the original post referenced, many higher end cards have the ability to change how some transmit and receive buffers are allocated. They can chose how full the buffer can get before triggering an interrupt, some can shift the buffer sizes to favor transmission over reception, etc... --Rich -- _ Rich Puhek ETN Systems Inc. _
Re: Browser preferences/options (was Re: Strong encryption for mozilla (woody))
Karsten M. Self wrote: Biggest browser beefs: - Stability. Quit with the fucking crashing already. Don't lose my stuff (this includes state). ***STABILITY IS NOT OPTIONAL*** - Speed. Render. Quickly. Load. Quickly. Stop. Quickly. Ties strongly to lightweight code as well. - Standards compliant. Support standards. Don't promote proprietarysms.. - Dependencies. Codependency sucks. - Bloat. A browser. Not a fucking kitchen sink, thank you very much. - Privacy. Allow me to control cookies, Java, Javascript. Support SSL. Default mode should maximise privacy. Don't do shit behind my back. - Security. Strong overlap with above. - Plugins. Suck. Third party apps should launch as same. They should *NOT* launch within my browser. Flash sucks. Period. - Unobtrusive. Stay out of my face. Do what I say. - Configurability. Allow my to configure: fonts, scaling steps, colors, animation handling (none), proxies, cookie/java/javascript handling (none, by default), preferred mail/news/telnet/ftp apps. SNIP -- Karsten M. Self kmself@ix.netcom.comhttp://kmself.home.netcom.com/ What part of Gestalt don't you understand? There is no K5 cabal http://gestalt-system.sourceforge.net/ http://www.kuro5hin.org Part 1.2Type: application/pgp-signature Whew! I thought I was the only one who had a hard time understanding why a web browser, really a HTML --that's Hyper TEXT Markup Language, yes kids, that means a way to make boring old text look kinda snazzy-- decides to consume well over 10MB just to get off the ground! More of us need to scream about the above to Netscape et al. To your list I'd add: - Cache. It should work properly... if I've got a few MB set aside on my HD, why redownload an entire page just 'cause I resized my window? - Configurability/Privacy: Let's figure out a cleaner way to say I want to save cookies for /. or similar sites and not for every other site under the sun. - Cookies: Let me know if A site is offering cookies (if I so chose), but don't stop the whole show with a damn dialog box a million times just because the frickin site wants to give me the equivilent of a big bag of Double-Stuff(tm) - Banners. I hate when a simple little web page that should take 5 seconds to load over a modem connection takes over a minute on a T1... just cause ads1.joesannoyingdiscountbanneradshere.com/cgi-bin/ads1234/crappyscript.cgi?makelotsofmoney=123456789oopsbadform doesn't happen to be responding at the moment. - Old OSes. My grandpa used to surf on a Macintosh Performa 675. Ever try to get one of those surfing reliably with Netscape??? Why the hell do we need over 500MHz and 64+MB RAM to surf the web and do email? Gotta agree most with Stability and Standards. A single page of bad HTML shouldn't sink all of H.M.S Netscape. That and bad HTML really shouldn't exist. Yes, I was about to argue that that's really the designer's fault, but browser-specific tags is what got us into this mess. --Rich -- _ Rich Puhek ETN Systems Inc. _
Re: (OT) Storage (8*IDE HDs) any experiences? - diodes
http://www.dell.com/us/en/bsd/products/model_pedge_1_pedge_2550.htm ...Has a system with dual power supplies. If anyone is interested in putting together a system like that, I suggest they go ahead and buy one. Otherwise, let's leave the design stuff to the power supply engineers at Power One and the other big power supply companies. They've got loads of (trained) people to take care of the design debates (including the use of Schottky Diodes). If further discussion on this toppic is desired, perhaps news:alt.engineering.electrical would be a more appropriate forum (or any of a number of other resources that show up in a google search). The debate was off-topic to begin with, strayed even further, yet stayed interesting with the concerns over load-ballancing supplies. Debating design features of power supplies is really wandering off of the beaten path... especially given the lack of experience most of us (the members of debian-user) have with the subject. Sorry for adding to the noise, but geez, this list is chatty enough already to be almost unmanagable. Bottom line: let's let the discussion die or else move it over to another list or NG. --Rich (loads of chatter deleted) -- _ Rich Puhek ETN Systems Inc. _
Re: FW: Sendmail
How are you trying to start sendmail? If you are just trying sendmail as root, you're not starting the daemon up to listen for incoming mail, you're invoking the sendmail program as if you want to create a message. Try (as root) /etc/init.d/sendmail start and see what you get. Sendmail is a bear to set up at first. It's a great MTA, very capable, but can be a bit intimidating to configure. If your needs on the box are fairly simple, you may want to consider using exim instead. As something else to consider: Do you need to receive mail on that box? For most of my machines, I have absolutely no need to receive email directly on the machine. The only machines I run sendmail on are those that are actual mail servers (those that are either receiving email for my domains or those that are relaying mail for my clients). The rest of my servers need to be able to send email (so that daemons can send error messages to me, and stuff like that). For all those servers, I run Exim setup to use a mail server as a smarthost. Most importantly, I remove exim from inetd.conf, so that they're never used to relay mail. --Rich Peter Donaldson wrote: I tried sendmailconfig but still i am having the same problem. I also can not receive mail through that box :( -Original Message- From: ktb [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 02, 2001 11:40 AM To: Debian Users Subject: Re: Sendmail On Wed, May 02, 2001 at 11:26:26AM +1000, Peter Donaldson wrote: I have been having a problem with sendmail for a while but because i am just playing with linux it hasn't really bothered me that much. But could someone please healp me out. When i go to start sendmail it is giving me this message-- peter... Recipient names must be specified -- Help would b greatly appreciated :) Have you tried setting it up with sendmailconfig? kent -- _ Rich Puhek ETN Systems Inc. _
Re: (OT) Storage (8*IDE HDs) any experiences?
Dell Poweredge 2450 style servers is what you're looking for. They have two power supplies, each with its own power cord. Yes, it can run on one PS... the last one I set up ran that way on my desk since I only had one cord handy. Of course, you'll want to make damn sure the grounds are at the same potential, so doing funny tricks with where you plug them in could be a bad idea. Another nice thing about this (and probably all the machines Matthew's referring to) is that the motherboard doesn't need to support multiple supplies, nor do the hard drives, fans, tape drives, etc. I agree with Matthew in that there _is_ a reason to share the load, actually a few that I can think of. Let's say you have a pair of 300W supplies on a box that draws 250W at rest. Rather than let one supply crank along at 250W, let's let both supplies run at about 125W. That way, both supplies will run cooler (Depending on the supply design, the supply may actually have slightly lower efficiency at the lower load factor, but that's a trade off we can live with). Also consider what happens if the load was near the capacity of a single supply, and spiked over the capacity. If we were using the second supply as a backup to only be switched in if the primary failed, how would that be handled? --Rich Matthew Sackman wrote: On Fri, Apr 27, 2001 at 12:05:11PM -0500, Dimitri Maziuk wrote: On Thu, Apr 26, 2001 at 05:32:31PM -0700, Alvin Oga wrote: ... - even if you had 2 power supplies... - most motherboards only has one atx power connector True. And if you went for redundant PS's and a mobo that supports them, the cost would go way up. - are the two power supplies properly doing load sharing... Usually not. I imagine that's too hard and not worth the trouble anyway: what you usually want is redundancy, not load sharing. (I mean, if one PS dies, it will overload and kill the other one pretty fast. Not a good idea. And if each PS can handle the load alone, there's little point in sharing the load.) Every production server that I've seen that has 2 PSUs has both continuously running. At hopefully 50% capacity. There is no switch-over - if one goes then the other has to cope with both. Of course, the irony is that as they are both routed to the same power inlet, if the fuse in the plug goes then you're buggered anyway! :-) Matthew -- Matthew Sackman Nottingham, ENGLAND -- _ Rich Puhek ETN Systems Inc. _
Re: (OT) Storage (8*IDE HDs) any experiences?
-- _ Rich Puhek ETN Systems Inc. 2125 1st Ave East Hibbing MN 55746 tel: 218.262.1130 email: [EMAIL PROTECTED] _
Re: (OT) Storage (8*IDE HDs) any experiences?
(Sorry about the blank email... too much caffeene got me a twitchy trigger finger). Dimitri Maziuk wrote: In fact, there will be some point at which each individual PSU will run just as hot as if it handled all the load on its own (you can be sure your box will draw exactly that much load, thank you Mr Murphy). Hmmm, I'm doubtful that that's the case if our load is around 25% to 50% or something like that. That's based on almost a pure guess, but I recall UPS efficiency numbers of something like high 90% numbers for almost full load versus 85% or so at 50% load. Obviously, a very lightly loaded power supply will put out about as much heat regardless of whether it's running at 1% or 2% capacity, but I'm thinking that will diverge fairly quickly. Also, in most PS and case designs, the PS should have a negligable effect on the heat transfer of the box, given that its fan exhausts directly to the exterior of the case. If the case is ventilated through the power supply (as common in desktop PCs and low end servers), there will be a small change, depending on the temperature within the power supply case (been about 4-years since I decided 'Thermo and an ME degree weren't for me, so the ideas are fuzzy for me too). I'm not arguing that this is the case, I'm saying that this kind of argument can be twisted and turned any which way you like. Very true, which is why redundancy will be the main factor, not heat production. I had mid-80s AT PSUs. They'd still be working if I didn't have to move house and throw all that junk out. If a PSU lasts for 15 years, will 2 load-sharing PSUs last 30 years? Do I care? (Will I last 30 more years?) I know that CPU will maybe last 1/5th of that, disks maybe 1/3rd. So what is it I'm going to achieve by setting up load-balancing PSUs? Well, those mid-80s AT PSUs (in general, I mean) seem to have either been A) oversized for the AT PCs or B) just better quality than normal desktop-grade power supplies of today. (BTW: I recall hearing that Intel speced Pentium CPUs for a lifetime of 10 or 20 years in normal usage... used as a rationale for overclocking and the reduced lifespan it causes). No, I don't anticipate a linear relationship between load and lifespan, nor would I anticipate a linear relationship between load and heat disappated, heat dissapated and lifespan, etc... I would however, anticipate that keeping a power supply running somewhere under its full rated capacity will increase its lifespan to some extent. Also, in a load-balancing configuration, you eliminate the ... Also consider what happens if the load was near the capacity of a single supply, and spiked over the capacity. If we were using the second supply as a backup to only be switched in if the primary failed, how would that be handled? Well if you mean some piece of hardware suddently decides to draw $BIGNUM times its normal current, the PSU will die. Depending on the design, there's a circuit somewhere (eg. on the backplane) that does the appropriate magic and switches the second PSU on. Of course it'll die very soon too, unless the FPOH in question magically fixed itself in the meantime. I was thinking more like the combined load in the box was something like 95% of the rated capacity of the power supply, then spiked to 110% (like having a bunch of SCSI drives spin up). A decent power supply probably won't let the smoke out, but it probably won't give the best power either. A redundant load-sharing arrangement would have both supplies running at something like 42.5%, then spike to something like 55%. Granted, this is a bit of a stretch, but I've seen too many cases recently of servers in a simple consumer PC box that gradually got stuffed full of SCSI drives until a PS failed. Sometimes the magic fails -- I remember the look on my boss's face when he pulled a hot-swappable PSU out of a live swerver, and the box went down. Oops. (Only happened once; we later tried to reproduce the problem, quite unsuccessfully: PSUs switched over like a charm, every bloody time. Surprise, surprise.) Dima Like you said, Mr. Murphy pays a visit every now and then :-) (whew... ok, I'm done.. this topic has wandered far enough!) --Rich -- _ Rich Puhek ETN Systems Inc. _
Re:
Hmmm... Try looking near the top of the case in one of the 5.25 bays.. That's often where the CD ROM can be found. The d drive is usually found in MS Windows or in MS DOS, not Linux. :-) (Sorry.. it's Friday) Seriously though, I'm guessing you probably have your CD ROM and your second hard drive D on the secondary IDE controller. You probably had one of the two die on you, taking the entire IDE chain down with it. Reboot your machine, and see which drives are recognized by the BIOS. Maybe try re autodetecting your drives if your BIOS supports it (most BIOSes after '95 or so will). Try unplugging the hard drive and see if the CD ROM reappears... Try plugging in just the hard drive and unplugging the CD ROM and see if the hard drive reappears. Of course, before doing all this you should check to see if anything obvious (like an accidentally unplugged cable) has happened. --Rich cris wrote: my cd-rom is missing, also my d drive where do i look ? -- _ Rich Puhek ETN Systems Inc. _
Re: Linux Books
Number one is to get what you find interesting. I'd find a Barnes and Noble or similar bookstore to do some of your selection. I like to wander through the computer books section and page through some of what catches my eye. Often I find that a book that looked real interesting is actually very shallow in its explanation. Other times I find that the book dives much further into the subject than I really care about. By loooking at the book itself, you'll see the scope of its coverage and its applicability to your needs. As for some other suggestions... You might want Programming PERL from O'Reilly and Assoc. There are a load of other books out there on PERL, but you'll probably find that the camel book will get you up and running in PERL rather quickly (assuming that's a desireable goal for you). If you're planning on administering your own machines at some point, and need to provide web or email services, the Apache and Sendmail books from O'Reilly are almost mandatory. Be forwarned that the Sendmail book will likely require $400 worth of reinforcing to your bookshelf. It has also been rumored to distort gravaty in the near vicinity, so you might confuse some of your collegues in the Physics department. By your current books and by your job, I assume you'll be doing some C++ programming. You might want a C++ programmer's reference to keep handy in addition to the manuals you already have. The reference will be easier to search through once you've got the basics down. --Rich George M. Butler wrote: Hi all, I have been asking questions on this list and received lots of help. I am new to Linux but have some limited experiece with Unix in the past. I have just discovered that my employer will let me have $400 to buy books related to my job. I am a member of a mathematics faculty so naturally Linux is job related. I would be interested to hear from the contributors of this list what are their favorite Linux, Unix, Networking, Programming Language, or related books. Titles I already own are: Unix, System V, Release 4, Rosen et al, McGraw Hill, 1990, Common Lisp, Steele Jr, Digital Press, 1984, A Programmer's Guide to Common Lisp, Digital Press, 1987, Linux in a Nutshell, Siever, O'Rielly, 2nEd, 1999, Learning Debain/Gnu Linux, Mc Carty, O'Reilly, 1999, C++, How to Program, Dietel Dietel, Prentice Hall, 2nd ed, 1998, Running Linux, Welsh et al, O'Reilly, 1999, C++ Primer, Lippman Lajoie, Addison Wesley, 1998. The TeX Book, Knunth, Addison-Wesley, 1984, I have to spend the money by May 1. I hope this message is not off topic but I feel lots would be interested to hear the responses of the readers of this list. Thanks for your help. George -- -- _ Rich Puhek ETN Systems Inc. _
Re: named/bind vs. /etc/hosts.deny -- can't verify hostname
Will, A few questions, mostly to ask yourself, that may help you find what's going on. Why mess with bind on the internal machines? Why not just populate /etc/hosts and be done with it? Regardless, which machines are entered into /etc/hosts on duo? Does an nslookup or a dig against the DNS server jive with the /etc/bind files? Shouldn't you have a $ORIGIN lan. in your first file (after the @ sections)? How does your machine show up in the logfiles? (something like telnetd ... connect from mac (192.168.1.100) or ...connect from mac.lan. (208...? --Rich will trillich wrote: Apr 17 14:58:33 duo xinetd[325]: warning: /etc/hosts.deny, line 15: can't verify hostname: gethostbyname(kat.lan) failed aaugh! my wife's machine is windo~1 98 at 192.168.1.200; my machine is a mac os 8.1 at 192.168.1.100. i have no trouble connecting via ftp (or ssh or http) but she's bounced out with the xinetd[325]: warning: /etc/hosts.deny, line 15: can't verify hostname: gethostbyname(kat.lan) failed we both have the same nameserver setup (name server is debian potato at 192.168.1.1) ... what do i need to look for? here are the /etc/bind/lan* files that pertain: ; ; *.LAN bind/named/dns ; $TTL 2W @ IN SOA lan. root.lan. ( 200104171 ; Serial 8H ; Refresh 2H ; Retry 1W ; Expire 1D ); Default TTL ; @ NS ns A 192.168.1.1 ns A 192.168.1.1 duo A 192.168.1.2 mac A 192.168.1.100 kat A 192.168.1.200 and here's the reverse-lookup file to match: ; ; *.LAN reverse lookup bind/named/dns ; (1.168.192.in-addr.arpa) ; $TTL 2W @ IN SOA lan. root.lan. ( 200104173 ; Serial 8H ; Refresh 2H ; Retry 1W ; Expire 1D ); Default TTL @ NS ns.lan. @ PTR lan. ; 1 IN PTR ns.lan. ; 2 IN PTR duo.lan. 100 IN PTR mac.lan. 200 IN PTR kat.lan. duo.lan is a secondary debian server, and she can't get in from 192.168.1.200 because of a gripe against /etc/hosts.deny, which contains ALL: PARANOID but i can get in from 192.168.1.100 with no trouble. what gives? -- don't visit this page. it's bad for you. take my expert word for it. http://www.salon.com/people/col/pagl/2001/03/21/spring/index1.html [EMAIL PROTECTED] http://sourceforge.net/projects/newbiedoc -- we need your brain! http://www.dontUthink.com/ -- your brain needs us! -- _ Rich Puhek ETN Systems Inc. _
Re: Diagnostic advice: How to find; what filled up a 27% of 13.3 Gb drive overnite!
John, Start by doing a cd to / (root directory). Then (as superuser, just to eliminate annoying permission denied messages) do: du -sh * | more Which will give you the disk usage of all the subdirectories. One of them will likely be much bigger than the others (probably /var). If so, cd to that directory, and repeat the above command. By repeating that sequence, you'll find the biggest problems. Also, you'll probably want to look at a better partitioning scheme to limit the damage when/if this happens again. For a system like yours, probably a scheme like the following: /dev/hda1 200MB swap /dev/hda2 5GB/ /dev/hda3 2GB/usr /dev/hda4 (everything else) /var Push more onto the / partition if your /home gets real big, more onto /usr if /usr/local/src gets big, and more onto /var if your logs are killing you for some reason. --Rich John Foster wrote: I am running a mixed testing/unstable system and after my last upgrade I have 2 new problems. My test system is on a 13.3 Gb drive and I have been at 63% full for about a year. I was downloading some mail and got a disk full error. I had NEVER seen this before on any of my Linux systems. I checked and sure enough the disk is showing 100% full. I moved a bunch of old archives (about 7%) off it and left it sit for a day. When I came back it was completely full again. Any ideas on how to locate the problem. I have NO Clue. Thanks! John -- _ Rich Puhek ETN Systems Inc. _
Re: Funny Story
http://spaceflight.nasa.gov/station/crew/exp1/ex1logs.html Karsten M. Self wrote: SNIP Yes, you can read all about it in the journal of one of the astronauts. Speaking of which, I went hunting last night but couldn't find them. Link, anyone? -- _ Rich Puhek ETN Systems Inc. _
Re: NFS mount at startup
The Automounter will help you. The documentation wasn't real clear to me at first, but I managed to get it up and running. See the Automounter mini-HOWTO at: http://www.kernel.org/LDP/HOWTO/mini/Automount.html or your favorite LDP mirror for more details. --Rich Stephen E. Hargrove wrote: How do I prevent NFS filesystems from mounting when I start up my computer. As the system is booting, it attempts to mount them, and I only want them mounted when I want them. Here's one example from my fstab: papa:/music/music nfs rsize=1024,wsize=1024 0 0 -- steve -- _ Rich Puhek ETN Systems Inc. _
Re: backing up and rebuilding the kernel
Look into doing something like: tar -clzvf /home/somedir/My_backup.tar.gz /* The -l (el not one) option will keep tar from trying to move off to another filesystem. You might want to leave off the -z from the options to skip compressing the archive (name it My_backup.tar then). You've got a 2Gb root and 18Gb /home, so the space savings might not be important. As for safely fiddling with your kernel... you'll want to get familiar with how LILO works and where your kernel goodies are. The Debian kernel install tools help a lot, since they try to keep your last successful kernel around in case the new one fails to load. Basicly, you'll end up with two files in root (actually, they're usually symlinks to files located in /boot): vmlinuz and vmlinuz.old. Lilo will try to load up the vmlinuz image at boot. If that image fails, you can hold down shift (or was it spacebar?) as your machine restarts. You will be presented with a boot: prompt, which will allow you to chose a kernel image to load. It's been a long time since I've had to do this (been lucky with my kernel builds lately), but I believe all one needs to do is enter vmlinuz.old to load the old image. Don't worry about problems enabling SMP support... that's fairly simple (a checkbox in the kernel config). The other little details like device drivers will get you. If you're a constant fiddler you're probably able to figure it out though :-) --Rich JACKSON, DEAN wrote: Right I have just spent the last 3 evenings building a debian server (much to my fiancés disgust) it has 2 hard drives a 2gb and a 18gb all the system files are mounted on the 2 gb(sda1) /home is mounted on the 18gb(sdb2) (swap is sda1) what is the easiest way to back up the system files (2gb sda1) to a single file on the 18gb/home (so it can be backed up onto a tape drive one day) and also if I need to restore what is the best way? a bulk restore would be required to get the system back to how it is now/when I last backed it up I need to do this as I am a constant fiddler and newbie so I cock up regularly and do not know how to undo what ive done sometimes. and the next task involve a kernel re-build to enable dual processor support (smb) and im scared just thinking about it so any advice here would be nice. all I want to do is add smb can I do this without deleting anything I already have?(newbie not sure exactly what ive got but it works!) Dean Jackson TeleWare email [EMAIL PROTECTED] Telephone 01908 251474 Dean Jackson (E-mail).vcf This message has been checked for all known viruses, by Star Internet, delivered through the MessageLabs Virus Control Centre. For further information visit: http://www.star.net.uk/stats.asp -- _ Rich Puhek ETN Systems Inc. _
Re: Daemon mgmt was Re: turning off /sbin/portmap
Also look at the update-rc.d command (see man update-rc.d for details). That will allow you to do things like: update-rc.d postres start 3 (start postres in runlevel 3) update-rc.d postres stop 50 6 (stop postres at sequence 50 in runlevel 6) --Rich Alan Chen wrote: Just as an excercise to my own sys admin knowledge, I'll summarize my general knowledge and just ask if anyone has suggestions or differences in my understanding. Daemons (or services) can be manually manipulated in debian using /etc/init.d/daemon with the command start, stop, restart, etc.. This will only change what is currently running. If you reboot, whatever was configured for your runlevel will be started again. rcS.d/ is stuff started for every runlevel rcn.d/ lists runlevel specific daemons that are started at boot update-alternatives (or was it alternatives-update) is a admin tool for adding, removing daemons from various runlevels. To remove a daemon from starting at a given runlevel, i generally just delete the entry in the /etc/rcn.d directory. Are there any reasons for not doing that? I wish the update-alternatives would accept syntax like update-alternatives +3 postres to add postgres to runlevel 3. Maybe it does, it's been a while since I last used it. Any other general notes? -- _ Rich Puhek ETN Systems Inc. _
Re: Linux Virus
Well... remember that most of the recent Melissa style worms are slapped together with Visual Basic... Not a great risk that ext2 support will show up :-) --Rich ...and the paperclip winked at me and said: It looks like you're writing a macro virus... Would you like help? (another stolen .sig) Ethan Benson wrote: something more nefarious would be for the virus when run from windows to find linux partitions and use internal ext2 support to modify binaries on the linux filesystems. -- _ Rich Puhek ETN Systems Inc. _