Send tech-support questions, answers and article ideas to The Answer Gang <firstname.lastname@example.org>. Other mail (including questions or comments about the Gazette itself) should go to <email@example.com>. All material sent to either of these addresses will be considered for publication in the next issue. Please send answers to the original querent too, so that s/he can get the answer without waiting for the next issue.
Unanswered questions might appear here. Questions with answers--or answers only--appear in The Answer Gang, 2-Cent Tips, or here, depending on their content. There is no guarantee that questions will ever be answered, especially if not related to Linux.
Before asking a question, please check the Linux Gazette FAQ to see if it has been answered there.
Dear Answer Guy,
I really hope you can help me in my quest to change from Windows to Linux. Here is what I have done so far:
I've got a 2 year old standard PC: PII 333MHz, 128MB RAM, 4.3GB, CD-ROM, Modem, ESS soundcard.
I've downloaded the ISO file from SuSE and it is called live-evaluation-i386-70.iso
I've put this onto a CD. I can see the file on the CD under Windows. When I try to boot up using the CD ROM it ignores it. My BIOS is seto to Boot From CD and it will boot from a Windows OS CD so I know the capability is there.
I'm starting to feel that there is something I need to do to the ISO file when writing it to the CD. I've seen mention of doing this by "Burning the Image" - whatever that is. I've read the HOW-TO for CD and ISO and it doesn't explain how to do this under Windows only. It assumes you have a working Linux environment.
I came across your page, and dozens of others on this subject, and I thought I had found the holy grail as the question is exactly what I was going to ask. The thing is your answer is for a working Linux system - the guy had this as well as Windows - there is no mention of how to do this under Windows. Then the person who asked the original question writes back saying:
Thanks for your information. And ironically, shortly (next day) after I wrote you the email, I did find out what was going wrong and how to fix it.
WinOnCD did have the capability, but it was somewhat a "hidden" feature of sorts.
I do appreciate your response, though.
He didn't even mention how he did it in WinOnCD!!!! ARRggghhh! I can find this software on the Web but I still need the knowledge to find the "hidden feature". Oh this is so frustrating., so close yet so far.
Please please please can you help me to make my ISO file, which took hours to download, into a bootable CD so that I can install my first ever Linux OS?
To trim it down considerably, he has tried Nero Burning software but couldn't make sense of it, and the image that seems to result doesn't work. YaST2 booted from floppy gets as far as reaching for the CD, then continues to complain that the CD isn't valid.
Usually the Gang would suggest a rescue disk - our perennial favorites wouldn't help, they don't have cdrecord aboard, but you could try muLinux (http://mulinux.nevalabs.org/) boot that from floppy, and use the Linux software on it to mount up your DOS filesystem, then cdrecord according to the normal HOWTO to burn the SuSE CD. I have no idea if it would work though it seems worth a shot.
Would some kind soul out there who lives in both worlds point us to a reliable CD burning app for Windows, along with some fairly simple instructions? We'd be glad to let you put the article in the Gazette if it will help enough potential Linux'ers out there. -- Heather
I read Ben Okopnik's article in the February issue of Linux Gazette titled
No More Spam! (a "procmail"-based solution with tips on "fetchmail" and "mutt")
and I tried implementing some of his suggestions. My problem seems to be this: While fetchmail will get my e-mail, and pass it off to procmail, which delivers it to a designated file, I can't seem to read the e-mail with kmail. I've set my designated file as a local mailbox for kmail to read, but when I try to read it using kmail I get 1 of 2 behaviors:
I tried setting up kmail to read from the local mailbox with the following
lock file options:
mutt dotlocked priveledged
Any of the above will cause either behavior 1) or 2).
Anyway, I'm hoping you might be able to help me figure out how to read my local mailbox with kmail.
P.s. Included is a snapshot of my kmail configuration for reading the local mailbox into my inbox. Hopefully, it will be of some use to you for diagnosing my problem.
P.p.s. Thanks in advance for any help you can give me on this.
P.p.p.s. Here are the configuration files for the various utilities: (I blanked my password in the .fetchmailrc file in this e-mail, for security reasons )
Cari amici della gazette
mi chiamo Francesco Barilà
Insieme ad alcuni amici vorremmo creare uno shell-wordprocessor emacs
based: qualcuno di voi è interessato?
Se lo siete mandate un email a firstname.lastname@example.org oppure visitate linux.interpuntonet.it/angolinux al tread "nuovo progetto"
Dear gazette- friends
I am Francesco Barilà
I and my friends want create a shell- woprdprocessor emacs-based: is anyone
mail to email@example.com or visit linux.interpuntonet.it/angolinux and see the tread "nuovo progetto"
Let him know what you think, folks. For my own part, we could still use a wordprocessor that actually works... -- Heather
Hi, I'm fighting with syslog-ng, trying to centralize the logs of all a network to a log server. Till now, I'm justing testing it between two machines ... I've already been on it a long time and still no result, simply nothing appears in the final log on the logserver ... I've attached the syslog-ng.conf of the client and the server's one too.
You must know that for the server's source, I've already tried to enter
parameters (I mean an ip and a port) but it doesn't work either so at
the moment, I'm trying to use the defaults. The same thing for the
Can you help me on this?
I hope you can help me with this problem.
We've upgraded our kernel from 2.0.0 to 2.2.14-5.0 (Red Hat 6.2). We have a master node that is connected to a total of 14 nodes. In the past, when we rsh'd into a node and wanted to reboot we would just type at the prompt 'reboot'. We would immediatly be kicked back to the master node prompt while the other node was rebooting. This worked out fine.
However, when we rsh into a node and type 'reboot' the terminal hangs, for several minutes, until the node has completed rebooted and then it drops us back to the master node. We want it to work the way it has in the past. Is there some file or configuration I'm missing, some flag to turn off or on?
Any of you gentle readers with experience in this kind of clustering, if you have any idea what it might be, then feel free to lend a hand. We'll publish your answer here. -- Heather
Hii james, this is mehul from india...
need to know if i can control bandwidth on a particular interface using ipchains or any other utilities in linux... basically, ill have a multi-homed linux workstation and i want to limit bandwidth on one interface to 512kbps... is that possible?
I would like to define keyboard shortcuts not specific to a Window Manager. I have read documentation on xmodmap, but the information I have found only talks about and show examples of changing key actions to another pre-defined action. That is, being able to change the caps lock key to function as an escape key. How would I go about assigning a command to a lctrl + F1 key code?
The ultimate use for this is to help a fellow LUG member. He wants to be able to assign text strings to "hot keys" whereby the text string would be copied to the window he is working in. He wants to be able to save the text strings as he is working so that they can be used in other session; and have the ability to delete or change these text strings. I can proably figure out a script to do this, but I am stuck on defining the keyboard shortcut.
Daniel S. Washko
is there any kernel upgrade from 2.2.1# to 2.4 how-to 's. I have upgraded the kernel using a general how-to and I have problems mounting the vfat partitions by getting the error "invalid major and minor numbers". I know the way the system deals with special device files has changed at 2.4. Also the "eth0" ethernet adapter is not being recognised. Any help in finding good documentation for this process would be greatly appreciated.
On our project if an established client socket connection on a remote chassis is suddenly terminated (e.g., the chassis is powered off), the socket connection on the local chassis changes from ESTABLISHED to FIN_WAIT1. If we then try to restart our application on the local chassis, it does not work because the socket connection is stuck in a FIN_WAIT1. After 5-10 minutes the socket connection stuck in FIN_WAIT1 clears itself and we can successfully restart our application on the local chassis, but this wait is too long.
Do any of you know how to expedite the process of clearing FIN_WAITs on a Linux/UNIX chassis under these conditions? The only we can get them to clear is to either wait 10 minutes or perform a chassis reboot (sync; shutdown -r now). Is there a system call that can be used to tell the operating system to close/delete/clean-up/remove all socket connections immediately?
The only thing I have run across so far is to possibly make a kernel change (e.g., change #defines in .../include/net/tcp.h), or set a socket option which causes the sockets to close/clean-up faster. Note: these changes may be risky if we shorten the timeouts too much.
Any help would be appreciated.
I've read the articles from issue 57 of the Linux gazette, but am unable to get my cd-rw to work. I don't think I am managing to emulate scsi correctly, although I have followed the things sugested. I have a sony cd-rw (CRX 145E ATAPI) and run mandrake 7.0, I also have a iomega 100Mb zip drive on hdb, a dvd-rom on hdc and the cd-rw is on hdd. This is what I've done and what the comp says:
In /etc/rc.d/rc.local added
In /etc/conf.modules added
alias scd0 srmod alias scsi_hostadapter ide-scsi options ide-cd ignore=hdd
(also tried replacing srmod with sr_mod)
In /etc/lilo.conf added
then in console typed
lilo (and tried /sbin/lilo)
then rebooted, then dmesg gives at the end:
hdb:<3>ide-scsi: hdb: unsupported command in request queue (0) end_request: I/O error, dev 03:40 (hdb), sector 0 unable to read partition table scsi0 : SCSI host adapter emulation for IDE ATAPI devices scsi : 1 host. Vendor: IOMEGA Model: ZIP 100 Rev: 14.A Type: Direct-Access ANSI SCSI revision: 00 Detected scsi removable disk sda at scsi0, channel 0, id 0, lun 0 sda : READ CAPACITY failed. sda : status = 0, message = 00, host = 0, driver = 28 sda : extended sense code = 2 sda : block size assumed to be 512 bytes, disk size 1GB. sda:scsidisk I/O error: dev 08:00, sector 0 unable to read partition table
For some reason I can't get the machine to emulate scsi on anything other than hdb. cdrecord -scanbus only lists the zip drive too. The /sbin/insmod ide-scsi comand also stops the zip drive from working. Have you any ideas as to what might be going wrong?
Many thanks for any advice you can offer
This is in reply to firstname.lastname@example.org's mail about scilinux.freeservers.com not having a e-mail address on the page. In fact at the bottom of the page before the disclaimer we have links to contact us. It appears when i look at it from India On Netscape. I am one of the maintainers and please do contact me for your suggestions, ideas, flames, etc ..
We are just crystallising our thoughts after the fruitful discussion with TAG (Jan 2000 issue) and have decided to first install all the relevant packages on a RedHat partition of our PC, find the library dependencies, make a "library farm" and use ldconfig after adding the libraries in the library farm to the /etc/ld.so.conf. To check what problems this gives we will then make a "zeroth" version of our CDROM and installation script and try installing all on the SLACKWARE (my dream distro) partition of our PC. Test the software, check if anything else goes wrong, etc .. It seemed to work for scilab and octave when I did a manual check. From what I have read till now about libraries (not much at all, maybe I should read up and write an article for the LINUX GAZETTE -> best way to learn something) this does not seem to be a wrong thing to do.
Thanks, Manoj, we'd love to see your article! -- Heather
Another thing is we are not trying to make a platform independent package or installer, but just want to install some selected scientific software packages on any old/new Linux PC. We hope to make a zeroth version, test it at our institute PCs and only after thorough testing keep it for download at some voulenteering site or mail it to anyone who pays the posting and CDROM cost.
PS: We are still looking for some server to host our site. We do thank freeservers, but the it would be nice to host the site at some server that displays Linux related advts only.
My Enviornment for scientific computing on Linux page : http://Scilinux.freeservers.com
...I love what you do but how about putting links to the rest of LinuxGazette (or at least the front page) at the tops and bottoms of each Answer Gang article? Thanks!
-- Jim Coleman
Once upon a time we had the complete "normal" navbar as well as the TAG bnavbar, but people said it was confusing. They kept hitting next article when the meant next TAG item, or vice versa. Having a pointer to the TOC doesn't sound so bad though. So, I've added that into the tag sub-items navbar. Let me know how you like it. -- Heather
I'm glad you're using mailto's that include the TAG address. The problem is the URL construction was incorrect.
It should be an equals sign (=) after the cc parameter. A quick run through sed should fix this before you get too many messages pointing out the error.
If there's any way to add the subject, you could make the URLs look like this:
Note that the subject has to go through a filter that replaces spaces with the hex code "%20" to keep the URL legal. In perl, I'd do this:
$mailto_subject =~ s/ /%20/g;
You can quickly test the functionality of a mailto URL by typing variations directly into the address/URL box of the target browser and see if it calls the mail client as you intended.
Thanks, Tony, you win the AnswerBubble for the month. I like it when folks not only nail us for a bug (that one was me, I'll go fix the script so it'll be right next time ... but also send us Tip grade material about fixing it. I tried it, and I ran it against some browsers we have around the place. My scripts work great ... but the browsers don't accept subject values reliably, and I don't want to inflict any extra problems on our readers, considering that what it did when it misbehaved was to generate a bunch of insane "To:" values. Oh well -- Heather
I just want to say thank you.
Her greeting in The Answer Gang is beautiful. I'm a relative newbie to Linux (It took me a week of reading and playing to compile a kernel that would find my NICs). I understand the open source idea, and love it. I deal in 2 other fields where this idea (in a general sense) applies. Emergency Medical Services where we share ideas, issues, and solutions, not for money, but for the knowledge of helping others. Also in restoring old cars, Things are learned, forgotten, learned again, and most importantly past on. By sharing, we (Linux users) end up with a better app or method.
Again, Thank you
Remember, there are people out there who really appreciate the work that is done.
- Mike Gargiullo
In Issue 64 you invite Ray Taylor to join The Answer Gang. How would one do that? Is it like a mailing list open to everyone? Can anyone help?
TAG is run like a mailing list in reverse. The public sends in questions, and the subscribers are the answerers. To join, send e-mail to email@example.com with "subscribe tag firstname.lastname@example.org" in the message body. Then just jump in whenever you have something to say. At the end of the month, Heather selects some of the messages for publishing. -- Mike
In the article on ssh, scp, and sftp in the March issue, there is an important area that isn't covered: client/server compatibility.
If you're just doing a basic ssh (to get a remote shell), you're using a standard SSH protocol and any program named "ssh" is likely to work with any remote system that offers a service it calls "ssh."
But scp and sftp are not standard protocols. If you run the scp program from openssh against a remote system that's running an original ssh server, it will not work. (And when I learned this the hard way, it was very hard indeed: the error message isn't "this server doesn't implement this scp protocol." It is, for reasons that took a day of debugging to figure out, "invalid file descriptor"!
First off let me apologize to all the developers or others who I have offended with my views on the Linux Router Project (LRP). By no means did I want to start a flame war. The truth is that I wrote about something outdated. Second, that article was entirely my own; not the work or opinions of Linux Gazette.
Since I wrote "Mean Thoughts" I have received a great many meaningful and insightful messages from LRP users and developers. If I wrote any untrue information I want to know about it. One point of contention for example was whether the ip command is 'nonstandard.' This is purely subjective. If ip really is standard, it should replace ifconfig or route like ipchains replaced ipfwadm.
Nevertheless my views on the LRP have changed. I received such an education that I feel obligated to state for the record I have learned uses for each of the three main LRP distributions, EigerStein (http://lrp.steinkuehler.net/DiskImages/Eiger/EigerStein.htm"), Oxygen (http://leaf.sourceforge.net/pub/oxygen), LRP 2.9.8 (http://www.linuxrouter.org) --even in embedded systems. I am not brand-loyal. Advocacy is fine, but fanaticism has got to go. I'll use the best tool for the job, and how I determine what is the 'best tool' is purely subjective. Five years ago I preferred 3Com to any other NIC. Why? Two reasons: The founder of 3Com invented Ethernet, and the cards were recognized by all the OSs that the company used. I knew I would not have to worry about cross platform compatibility. Now I prefer SMC. Why? Mainly because all the OSs recognize them but also because I can jumper-select IRQ & I/O on the models I use.
Would I write another 'anti-Linux' article? Sure. But not one that could potentially insult anyone like when I said, 'developers wasting time'. Linux is merely a product. Windows NT is also a product. Never mind the fact that I despise products from Redmond, Washington: I don't think it's a sin to admit that NT is better than Linux at being a "Domain Controller." It does not change how much I like Linux.
Look at the article and notice its verbosity. It's an opinion, not a review. I did not write it solely for explaining my (i.e., not Linux Gazette == Don't shoot the messenger.) thoughts on the LRP, I also wanted to express and present other information that may be useful to the Linux community, for example the bit on standardization. I did not write it to maliciously annoy anyone. Also to my knowledge there is no technically inaccurate information. I wrote specifically, "I have not done a lot of work/research with LRP incarnation at linuxrouter.org as such but I am familiar with the Materhorn Project." My mistake was that I equated <Linux Router Project> with one flavor, Materhorn.
I may or may not follow with a "Nice Thoughts on the Linux Router Project" article. In any case, I'd like to put all hard feelings aside and hope that anyone who I have offended would do the same.
[Mike] We received several complaints about the article, feeling that it attacked the LRP unfairly. Dave Cinege, the creator of the LRP, was going to write a response addressing the inaccuracies he felt were in the article, but he did not have time to finish his letter by press time. I encourage readers with an interest in routing to follow the links above to the projects' home pages and decide for themselves if the LRP and its offspring are right for them.
Regarding Oxygen, EigerStein and 2.9.8, Dave writes:
They are derivatives of stable releases of LRP, which is currently 2.9.8. I have been creating a new OS similar to LRP for quite some time now. Many things have come out of this are a new multi-packaging system (standard?) that is more powerful then rpm or deb, yet not tied to any specific OS.
Regarding the 'ip' command, he writes:
ip allows you to control the extended routing features of 2.2 and 2.4, IE multiple routing tables. Ifconfig still works for the primary routing table and interface configuration. ip can replace ifconfig, but ifconfig is still the known standard.
A few letters questioned LG's editorial policy in allowing this article to be published. LG's policy is pretty open. If an article is about Linux, contains hard facts or cultural value (e.g., humorous articles, cartoons and articles about Linux VIPs), covers a topic relevant to a significant portion of the readership, is not an advertisement in disguise, and would still be relevant several months from now, we'll probably publish it. There are borderline cases, and this was one of them.
LG does not have a technical review board to screen every article, although I do send a few questionable articles to The Answer Gang for comment. You, our readers, are LG's technical review board, and usually this system works very well. 99% of LG's articles are published without complaint.
In any case, please remember this article describes one person's experience with certain routing programs. It's not meant to be gospel, in spite of the letter I received that said, "But newbies will read it and think it's gospel!" That's not how it works. If you want gospel, read several people's articles and compare them with your own experience.
Another thing this article does is raise the question, just because we can use Linux in a wide variety of routing situations, should we? Are you choosing a Linux router because it's the most appropriate solution for the task, or simply because "we're a Linux-only shop"? Even if the article failed to present LRP in a fair light, these are still questions worth asking.
As always, if you have any comments about an article, whether good or bad, send them to LG and we will forward them to the author.