Re: Catch-all mail server for test environment
On Wed, Oct 22, 2008 at 11:29 PM, Boyd Stephen Smith Jr. < [EMAIL PROTECTED]> wrote: > On Wednesday 22 October 2008, "Ketil Froyn" <[EMAIL PROTECTED]> wrote > about 'Catch-all mail server for test environment': > >Any tips on the best way to achieve this with Debian? And maybe exim, > > which is the installed mta? > > I'm not that expierenced with exim, but it seems entirely capable. My > first attempt at what you are trying would probably look like this: > > Use debconf to setup the system as a normal internet-connected host, making > sure to allow relaying for the entire internet (something like 0.0.0.0/0 > might work). Then change your dnslookup_relay_to_domains router to use a > different transport (e.g. maildir_home OR LOCAL_DELIVERY). > Actually, it was even simpler. I just set up a system filter with a "deliver [EMAIL PROTECTED]" line, and now all mail is sent there instead. Thanks for the help. :-) Ketil Froyn [EMAIL PROTECTED] http://ketil.froyn.name/
Re: making ssh connections persistent
On Wed, Oct 22, 2008 at 9:29 PM, Ron Johnson <[EMAIL PROTECTED]> wrote: > On 10/22/08 14:23, Kamaraju S Kusumanchi wrote: > >> I often use ssh to access other computers at school. However, I can't make >> these connections persistent. If there is no activity (say in 20 minutes), >> the ssh connection just freezes up. As a result, I have to kill all the >> gvim, xterm windows opened from that terminal. This is very annoying if I >> am coding something. >> >> Is it possible to make ssh connections which do not freeze even if there >> is >> no activity for a while? >> > > That sounds like a security feature configured on the host side. This is usually caused by a PAT router/firewall en route that times out and closes the connection when it doesn't see any packets for a while. The trick is to do like Tyler suggested or something similar, to cause ssh or tcp to send keepalive packets regularly over this tcp connection. Other more low level options are /proc/sys/net/ipv4/tcp_keepalive_{intvl,probes,time}. Ketil Froyn [EMAIL PROTECTED] http://ketil.froyn.name/
Catch-all mail server for test environment
Hi, I am setting up a qa/test environment in which to test a system. As part of that, I need to set up a mail server that will accept email to any email address (any local part, any domain part), and either deliver it to a local account (where I can pop it or something), or forward it to a real email address. So something like an open relay that just rewrites the recipient address. Any tips on the best way to achieve this with Debian? And maybe exim, which is the installed mta? Ketil Froyn [EMAIL PROTECTED] http://ketil.froyn.name/
Re: Determine order of network interfaces
On 2/23/06, Florian Kulzer <[EMAIL PROTECTED]> wrote: If you do not want to switch to udev, maybe the package 'ifrename' canensure a deterministic naming of your network interfaces. I always hadproblems when using ethX (as Magnus has also pointed out in his mail) which I avoided by assigning new names to all interfaces, e.g. "lan0","wlan0", "modem" etc. For a while I got lucky by simply determining theorder in which the modules are loaded in /etc/modules (the first one gets eth0 in that case, etc.) but that does not seem to work anymorewith newer kernels (ca. 2.6.14 and up, I think).Hi Florian,Thanks for the tip. I installed ifrename, and after a bit of try-and-error I figured out that I could specify the device name from the driver loaded (I couldn't specify the MAC address of a firewire device, so I was stumbling for a minute). This appears to be my best solution so far. My only concern is this warning that I get: Warning: Interface name is `eth0' at line 1, can't be mapped reliably.So it's still not perfect. I'm actually amazed there's no reliable standard solution to this problem, this can be critical on boxes with multiple interfaces (picture a firewall letting everything in and nothing out, or a file server sharing files to the entire internet instead of the local network). Cheers,Ketil
Re: Determine order of network interfaces
On 2/23/06, Magnus Therning <[EMAIL PROTECTED]> wrote: My temporary solution was to change the pattern for one of the devices.My problematic device is a rt2500-based PCI wireless card. I put thefollowing in a file in /etc/modprobe.d/ : options rt2500 ifname=wlan%d That renamed my wireless to wlan0, instead of jumping between eth0 andeth1.Maybe something similar is possible with the drivers you're using?I tried this:options eth1394 ifname=ieee%d but it didn't work. I actually think the ifname= parameter is specific to the rt2500 driver, and I couldn't see that the eth1394 driver had something similar. But it was worth a shot anyway, I'm getting frustrated...:p Cheers,Ketil
Re: Determine order of network interfaces
On 2/23/06, Alex Nordstrom <[EMAIL PROTECTED]> wrote: Information on which version of Debian and which kernel you use would beuseful in answering this question. Whether or not you use udev wouldalso be relevant.Thanks for the quick reply! I use debian 3.1 with the 2.6.8-2 kernel. Thursday, 23 February 2006 19:01, Ketil Froyn wrote: > Specifically, my problem is that the firewire driver suddenly started> using eth1 instead of eth2 yesterday. It hadn't done this before, and> I had to change my interfaces file as a result. The issue is that I > want a normal interface to be on eth1, and I want to be certain that> this never changes again. I have tried to edit /etc/modutils/aliases> and added (near the top)>> alias eth2 eth1394 If you use a 2.6 kernel, the modutils package which providedthe /etc/modutils/alias file for 2.4 kernels has, as the package'sdescription states, been superseded by the module-init-tools package.It appears the corresponding file therein is /etc/modprobe.d/aliases. I've tried changing the interface there too, with no luck... If, however, you use udev, or if using udev is an option for you, itwould probably be easiest and most elegant to write a rule for thenaming of the different components based on their characteristics (suchas MAC address). Here is a good introduction to this approach: http://dev-loki.blogspot.com/2005/12/forcing-network-interface-names.htmlI'm not aware that I use udev (dmesg | grep -i udev gives nothing). But I'm concerned that if I start using udev other stuff might change/break, so I'm not too keen actually. If this is the only option I have I'll give it a go, but surely there's another way? My requirement is pretty basic... Cheers,Ketil
Determine order of network interfaces
Hi,I have a computer with some network interfaces, and I am unable to determine which physical interface gets assigned to which ethX during boot.Specifically, my problem is that the firewire driver suddenly started using eth1 instead of eth2 yesterday. It hadn't done this before, and I had to change my interfaces file as a result. The issue is that I want a normal interface to be on eth1, and I want to be certain that this never changes again. I have tried to edit /etc/modutils/aliases and added (near the top) alias eth2 eth1394but this didn't help at all (I remembered to run update-modules, and saw the added lines in /etc/modules.conf). Then I tried creating /etc/modprobe.conf, with the linealias eth2 eth1394 and as a result I could remove all the network modules and successfully run "modprobe eth2", which had never worked before.However, the order was still wrong after rebooting, then eth2 was my normal ethernet interface, while eth1 had been "stolen" by the firewire driver again. So now I wonder how I can force my debian system to let me specify which interfaces get which names during boot.Cheers,Ketil
Re: 3c575_cb missing from debian install
On Fri, 2004-10-08, Andrea Vettorello wrote: > Mmh, don't know if can help, anyway try to look here > (http://linuxmafia.com/faq/Debian/installers.html) if you find a > Debian netinstaller that "seems" good enough for your laptop. Thanks for the tip, but I had a look, and I couldn't find any floppy images for debian testing that have more drivers... I have since found out that the 3c575_cb allegedly was merged into the 3c59x driver, but that doesn't work either, and I've tried it on 2 different 3c575 pcmcia cards now. *phew* Ketil -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
3c575_cb missing from debian install
Hi, I'm trying to install debian testing on my laptop, but I need the 3c575_cb for my PCMCIA network adapter, and for some reason it is not included on the boot, root, net-drivers or even cd-drivers images. I don't know why, the net-drivers has over 200kb free space. It's difficult to install without it, my Thinkpad 240 doesn't have a CD player, and I don't have a USB device, so network is the only practical way to install debian. Anyway, has anyone managed this? How? Or can anyone suggest how I can get that driver onto my system? I even tried downloading the pcmcia-modules package on my desktop system, and manually copying the module to my laptop, but it didn't work. Thanks. Ketil -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Problems with lstat() and large files
Hehe, talk about your well composed mail. That's what I get for sending it off in a rush... Anyway, I've tested with both glibc 2.2.5 (from woody) and glibc 2.3.2 (from sarge) on the same system, with both kernel 2.2.20 (from woody) and kernel 2.4.26 (vanilla kernel, self-compiled), and I figure debian glibc is broken. What happens is that with kernel 2.4.26, when glibc receives a successful lstat64(), glibc decides that something has gone wrong, whereas with 2.2.20, lstat64() fails, and lstat() appears to return the correct values to glibc. Here is a table of what I've tested: GLIBC 2.3.2 (from sarge): strace kernel 2.2.20: lstat64("largefile", 0xbc34) = -1 ENOSYS (Function not implemented) lstat("largefile", {st_mode=S_IFREG|0644, st_size=2644920320, ...}) = 0 ltrace kernel 2.2.20: __lxstat(3, 0xbebe, 0xbd20, 0x4014a550, 0xbd78) = 0 strace kernel 2.4.26: lstat64("largefile", {st_mode=S_IFREG|0644, st_size=2644920320, ...}) = 0 ltrace kernel 2.4.26: __lxstat(3, 0xbebe, 0xbd20, 0x4014a550, 0xbd78) = -1 GLIBC 2.2.5 (from woody): strace kernel 2.2.20: lstat64("largefile", 0xbc34) = -1 ENOSYS (Function not implemented) lstat("largefile", {st_mode=S_IFREG|0644, st_size=2644920320, ...}) = 0 ltrace kernel 2.2.20: __lxstat(3, 0xbebe, 0xbd14, 0x4012de48, 0xbd6c) = 0 strace kernel 2.4.26: lstat64("largefile", {st_mode=S_IFREG|0644, st_size=2644920320, ...}) = 0 ltrace kernel 2.4.26: __lxstat(3, 0xbebe, 0xbd14, 0x4012de48, 0xbd6c) = -1 Anyone heard of this before? Ketil On Tue, 27 Jul 2004 19:08:26 +0100, Ketil Froyn <[EMAIL PROTECTED]> wrote: > Hi, > > I'm having some problems with lstat() on debian stable and debian > testing with 2.4 kernels. A program that tries to do lstat on a file > bigger than 2gb gets an error. Has anyone else experienced this? I'm > running third party software that requires a functional ltrace(), and > doesn't handle this well. > > The lstat() appears to work with on debian stable with a 2.2 kernel, > but I need 2.4 for hardware support etc. > > I've tested on debian sarge as well, and I get the same results there. > Is this a known problem? Is there a known fix? The same program worked > fine on redhat 8. > Do I have to go back? > > Here is a sample program. Compile like this: > $ gcc -o testlstat testlstat.c > and run like this > $ ./testlstat /path/to/largefile > and watch it fail with ltrace: > $ ltrace ./testlstat /path/to/largefile > > -- testltstat.c -- > #include > #include > #include > #include > #include > #include > #include > > int main (int argc, char * argv[]) { > struct stat buf; > if (lstat(argv[1], &buf) == -1) { > printf("error: %s\n",strerror(errno)); > return 1; > } > printf("Success\n"); > return 0; > } > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Problems with lstat() and large files
Hi, I'm having some problems with lstat() on debian stable and debian testing with 2.4 kernels. A program that tries to do lstat on a file bigger than 2gb gets an error. Has anyone else experienced this? I'm running third party software that requires a functional ltrace(), and doesn't handle this well. The lstat() appears to work with on debian stable with a 2.2 kernel, but I need 2.4 for hardware support etc. I've tested on debian sarge as well, and I get the same results there. Is this a known problem? Is there a known fix? The same program worked fine on redhat 8. Do I have to go back? Here is a sample program. Compile like this: $ gcc -o testlstat testlstat.c and run like this $ ./testlstat /path/to/largefile and watch it fail with ltrace: $ ltrace ./testlstat /path/to/largefile -- testltstat.c -- #include #include #include #include #include #include #include int main (int argc, char * argv[]) { struct stat buf; if (lstat(argv[1], &buf) == -1) { printf("error: %s\n",strerror(errno)); return 1; } printf("Success\n"); return 0; } -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: apache2 sends reply for / and redirect for ~ (setup Q)
On Mon, 5 Jul 2004 01:53:02 +0300, Micha Feigin <[EMAIL PROTECTED]> wrote: > GET /~ HTTP/1.0 Try GET /~/ HTTP/1.0 > I get a Moved Permanently reply That is standard when you give a directory without the trailing slash. > with the server's external > host-name. How do I set it up so that it works also for connections > from the localhost? Try setting up localhost as a ServerAlias, and do Get /~/ HTTP/1.0 Host: localhost Disclaimer: I didn't test this. Cheers, Ketil Froyn [EMAIL PROTECTED] http://ketil.froyn.name/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: syslog
On Wed, 30 Jun 2004, Matthew Joyce said: > I have been asked by a 3rd party (who managed some > comms equipment for us) if we have a syslog server. > > If so, they say they can direct some logs to it. > > We have some Debian boxes, don't they all have syslog > running ? > > How does it work ? man syslogd, look for "-r". If set, syslogd receives messages on the network from whoever sends them, and puts them in the local log. Makes it easier to log from devices without a disk, and makes it harder for a hacker to hide what he has done (he'll need to break in to the syslog server too). Ketil Froyn [EMAIL PROTECTED] http://ketil.froyn.name/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: [d-i] Hard drive not detected (4GB) on IBM Thinkpad 560X
On Mon, 28 Jun 2004 17:28:27 +0200, Sylvain Vedrenne <[EMAIL PROTECTED]> wrote: > > Hello, > > I'm stuck trying to network-install Debian Woody (in order to install > 'Unstable') on an IBM Thinkpad 560X. > > I'm using the 'boot', 'root' and 'net-drivers' floppies found here on saturday > 2004/6/26: > http://ftp.debian.org/debian/dists/testing/main/installer-i386/current/images/floppy/ > > partman says something like "here are the partitions on your hard drive" but > there is nothing! but the menu asking me what I want to do next... > > It seems the hard drive (4GB, C,H,S: 993,128,63) is not detected. I also tried > "linux hd=993,128,63" when booting with the Sarge installation > disks, but the result is the same. Try loading the drivers on the disk cd-drivers.img, I seem to recall that it has a lot of IDE drivers that aren't included on the other disks. I had the same problem recently, and that solved the problem for me. Cheers, Ketil Froyn [EMAIL PROTECTED] http://ketil.froyn.name/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Install CD does not boot on LITE-ON model LTN-403
On Wed, 23 Jun 2004 19:06:11 +0100, Rich <[EMAIL PROTECTED]> wrote: > > > I have been trying to help a colleague install debian, but we ran into > > some serious trouble booting the CD. None of the debian install CDs > > worked (tried netinst beta 4, test candidate 1, and woody cd 1), > > whereas redhat and all other CDs he has tried worked fine. The CD-ROM > > drive in question was a LITE-ON model LTN-403. We changed it with > > another CD-ROM drive, and then everything went smooth. > > > > I just wanted to share this experience for the archive, and perhaps > > someone will find out why this happened and fix the debian bootable > > CDs. :) > > Interesting- I've been trying to install Sarge in the same manner and can't > get past the first CD. > I'm using a Liteon model LTN-526D. > > I'd love the chance to get to grips with Debian, but so far I've just found > the experience rather frustrating. > > Knoppix, on the other hand detects everything except my sound card. A search on google turned up a couple of other people having problems booting stuff (not necessarily debian) on a LITE-ON drive, but Knoppix and Red Hat still seem to boot without trouble. Perhaps Debian can steal some of the working stuff from another open source project. Ketil Froyn [EMAIL PROTECTED] http://ketil.froyn.name/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Install CD does not boot on LITE-ON model LTN-403
Hi, I have been trying to help a colleague install debian, but we ran into some serious trouble booting the CD. None of the debian install CDs worked (tried netinst beta 4, test candidate 1, and woody cd 1), whereas redhat and all other CDs he has tried worked fine. The CD-ROM drive in question was a LITE-ON model LTN-403. We changed it with another CD-ROM drive, and then everything went smooth. I just wanted to share this experience for the archive, and perhaps someone will find out why this happened and fix the debian bootable CDs. :) Cheers, Ketil Froyn [EMAIL PROTECTED] http://ketil.froyn.name/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
apache-1.3.26-0woody3 secure?
It seems debian is using apache 1.3.26 still, and it doesn't appear to have been updated for quite some time. Is it secure/stable? What about these fairly recent problems with apache since 1.3.29, that are listed in the changelog for apache 1.3.x? CAN-2003-0987 CAN-2004-0174 CAN-2003-0993 CAN-2003-0020 Ketil Froyn [EMAIL PROTECTED] http://ketil.froyn.name/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: dpkg: Killed
On Wed, 2004-03-10 at 13:46, Colin Watson wrote: > On Wed, Mar 10, 2004 at 01:01:30PM +0000, Ketil Froyn wrote: > > > > o "apt-get upgrade" worked fine > > o "apt-get install gaim" downloaded packages, but then complained that > >a subprocess was killed unexpectedly > > o "dpkg -i gaim-.deb" says: Killed > > That suggests to me that you're out of memory. Check your system logs > for out-of-memory killer messages. Thanks for the help, Colin! This wasn't directly it, but you were really close... I (or the snapshot netinst CD) must have screwed something up during installation, because my swap partition was not in use, and I had no swap. I turned it on, and now everything works like a charm! Damn, that wasn't the the most obvious place for me to look...:) Too bad a clearer error message wasn't given. I've also looked in my logs, but didn't find anything. I'd think this should show any out of memory stuff: debian:/var/log# grep -i out * | grep -i mem debian:/var/log# grep -i oom * debian:/var/log# Anyway, it's sorted now, so thanks again! Ketil -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
dpkg: Killed
Hi, I'm a beginner with debian, but not with linux. I installed debian from sarge-i386-netinst.iso dated 3 March. The install was fine (including installation of X and some other stuff), but now it seems I am unable to install software. I tried a couple of packages, and got the same problems. Here are the symptoms I see: o "apt-get upgrade" worked fine o "apt-get install gaim" downloaded packages, but then complained that a subprocess was killed unexpectedly o "dpkg -i gaim-.deb" says: Killed I tried strace-ing the dpkg -i process, but I wasn't able to make out anything obvious. I've tried searching in the archives and on google, but didn't find anything that seemed related. Any pointers would be great. Thanks. Ketil -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]