LVM + multipath
Hi folks, I have a curiosity question about using LVM w/ multipath devices. I've got some SAN disk on which I made a volume group, and a couple of logical volumes. Each volume has an ext3 filesystem. I've added the filesystems to my fstab using their UUID's. UUID=f9b573fd-9457-4150-9007-fedbe708f866 /dbfs/db1ext3 nodev,grpid0 2 UUID=8909ad2b-9990-4f8d-a544-93f4781929df /dbfs/db2ext3 nodev,grpid0 2 What's curious is that while they mount just fine, the 'df' command lists them inconsistently: /dev/mapper/dbfs-db1 385G 195M 365G 1% /dbfs/db1 /dev/dm-3 385G 195M 365G 1% /dbfs/db2 Now, the /dev/dm-3 device is the same device (major/minor) as the corresponding /dev/mapper/dbfs-db2 device, so I suppose it really doesn't matter what df thinks they should be called, but it's aethetically unpleasing at best. Does anyone have any idea what's happening here? Debian Lenny amd64. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso - Mount Holyoke College officials will *NEVER* ask for your password. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Re: etch/lenny ext3 compatibility
2009-03-19_14:22:45-0400 Sven Joachim : > > The inode size of existing filesystems is always left alone. According > to mke2fs(8) it is not even possible to change it after the filesystem > has been created. > > For the record, the etch kernel should not have any problems with 256 > byte inodes, but etch's grub version cannot deal with them, see > http://bugs.debian.org/463236. I ran into a similar grub issue already, trying to use knoppix on a lenny install. Made me crazy for a few hours until I finally found a reference to this. That's how I found out that the default inode size had changed. Interesting to note that mke2fs on RHEL doesn't include the -I option at all. It creates 128 byte inodes, and that's that. There shouldn't be any trouble with the kernel I imagine. Are there any userspace applications that could be affected if, on etch, I try to use a 256 byte inode ext3 filesystem created on lenny? I'm thinking ACL's. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso facebook: http://tinyurl.com/d63r5c - I wish my computer would do what I want it to do - not what I tell it to do. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Re: etch/lenny ext3 compatibility
2009-03-19_11:59:59-0400 rpeterso: > As I understand things, lenny's ext3 implementation uses 64 bit inodes, > to be forward compatible with ext4, etc. Sorry, I meant 256 byte inodes. Same issue, despite my brain fizzle. > I have FC disks that I'd like to use on either of two servers I have. > One is in production, running etch. The other is new, and I'd like to > use lenny. Will lenny read/write ext3 filesystems created on etch in > such a way that I can still use the filesystem on etch? I imagine going > the other way would be impossible. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso facebook: http://tinyurl.com/d63r5c - I wish my computer would do what I want it to do - not what I tell it to do. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
etch/lenny ext3 compatibility
As I understand things, lenny's ext3 implementation uses 64 bit inodes, to be forward compatible with ext4, etc. I have FC disks that I'd like to use on either of two servers I have. One is in production, running etch. The other is new, and I'd like to use lenny. Will lenny read/write ext3 filesystems created on etch in such a way that I can still use the filesystem on etch? I imagine going the other way would be impossible. ?? -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso facebook: http://tinyurl.com/d63r5c - I wish my computer would do what I want it to do - not what I tell it to do. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Re: stability of multipath
2008-01-07_01:21:47-0500 David <[EMAIL PROTECTED]>: > Paul Johnson wrote: > >On Jan 6, 2008 9:55 PM, Alex Samad <[EMAIL PROTECTED]> wrote: > >>I believe HP is becoming more debian friendly > > > >I believe this is one of their ongoing goals, at one point they were a > >major donor to Software in the Public Interest, if not Debian > >directly. > > > Debian directly. > Bdale Garbee still handles all their Linux consultation work. > > http://www.gag.com/~bdale/nc4000/ Thanks for the feedback. I'm afraid this is getting a little OT though. While I am quite interested in the level of official support commercial vendors provide debian, today I'm more interested in community support... ;) I already have my hardware (Dell 2950 + Qlogic 2432 + Promise VTrak), I'm just wondering about using multipath. Anyone using multipath in production on etch? Happy with it? I'm at least going to be doing some testing with it. I'm just fishing for a statement like: "Yes, I've used multipath on at least twenty heavily used servers for quite some time and it's like a rock." Or maybe "Don't even think about it unless you're insane, because I lost my job, my family, and even my dog doing that." That kind of thing. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: stability of multipath
2008-01-06_21:12:01-0500 Alex Samad <[EMAIL PROTECTED]>: > On Sun, Jan 06, 2008 at 08:00:00PM -0500, Ron Peterson wrote: > > Is anyone on this list using multipath and/or multipath-tools in > > production at an enterprise level? I'm considering using multipath > > for a new fileserver and a new mailserver, running etch, using > > qlogic 2432 hbas. This would the first time I did this in > > production, however, so it would be nice to have some reassurance > > that I'm not sticking my neck out too far. > who provide you storage, probably best to talk to them, most vendor > have matrix's out what combo of storage/hw/firmware/sw they have > certified, I believe HP have looked at multipath, not sure about the > others In general, I find most such matrixes exclude Debian, preferring to focus on corporate distro's like Suse, Redhat, etc. That hasn't scared me away from using Debian where I know the pieces well enough. But multipath is new to me, so I'm a little out of my comfort zone. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
stability of multipath
Hi, Is anyone on this list using multipath and/or multipath-tools in production at an enterprise level? I'm considering using multipath for a new fileserver and a new mailserver, running etch, using qlogic 2432 hbas. This would the first time I did this in production, however, so it would be nice to have some reassurance that I'm not sticking my neck out too far. TIA -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: software suspend and system cloning
2007-08-29_20:50:49-0400 Ron Peterson : > I was cloning a software raid system to a hardware raid system (Dell > 2650 to 2850). I took out the mdadm startup scripts, but apparently > software suspend was somehow trying to restore some kind of state > information about the system that was no longer valid. BTW, this is the where the boot process stops: Begin: running /scripts/local-premount kinit: name_to_dev_t(/dev/md8) = md8(9,8) kinit: trying to resume from /dev/md8 Attempting manual resume I/O error reading swsusp image Of course there is no swsusp image on my swap partition, because this is a newly cloned system. The new machine also has more memory, so I booted, with 'noresume', and installed a bigmem kernel, but that really crapped out. I went home after that, so I don't have much more detail, but from what I'm reading, it sounds possible that my filesystems were corrupted by what I was doing. This is starting to remind me of how tightly coupled Windows is to the hardware it's installed on. Yuck. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: PAM + LDAP and SSH
2007-07-31_12:51:35-0400 Allan Senna Porto <[EMAIL PROTECTED]>: > Hi, > Anyone know about problems with /etc/security/access.conf and > SSH in Etch? > I'm trying to block ldap users to access my servers, but > everything that I made don't block user to loggin. > If I use this to /etc/pam.d/login and /etc/pam.d/gdm, the > user can't loggin. It would help others to help you to see what you actually did. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
ssh x11 forwarding on etch
Debian Etch I just discovered that I had to add the following to my /etc/ssh/ssh_config in order for X11 tunnelling to work when I logged into remote systems. Host * XAuthLocation /usr/bin/xauth According to the installed man page for ssh_config: XAuthLocation Specifies the full pathname of the xauth(1) program. The default is /usr/bin/X11/xauth If the xbase-clients package is going to put xauth in a non-default location, shouldn't the ssh package be updated to reflect that? Anyway, maybe posting this will save someone else some aggravation. I'm wondering why this doesn't appear to be affecting more people. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: an ssh auth. question++
On Mon, May 15, 2006 at 02:33:31PM +0200, Krassen Deltchev wrote: > what is better to be used:: > a root authentification to a remote server over ssh or just a user > authentifaction to this very server+ #su root and then authentificate > the root access? I'd say 'which is better' depends on the context. Whether you are using ssh or not, if you can avoid being root, that's better. Sometimes of course, you must be root. If the set of tasks you are remotely performing via ssh is limited, you might consider using the 'command' option (see the AUTHORIZED_KEYS section of 'man sshd' for details) in your remote host's authorized_keys file to set up a restricted set of operations. For example, on the local machine, you might like to create a special key dedicated to a particular purpose, e.g. ssh-keygen -t rsa -b 2048 -f ~/.ssh/id_rsa_transfer If you wanted to transfer a file from the remote machine that was only readable by root, you might set up the remote machine's authorized_keys file to include something like: command="cat /path/to/afile" Now, from the local machine, you can do something like: ssh -i ~/.ssh/id_rsa_tranfer.pub [EMAIL PROTECTED] > /a/local/file Depending on your application, you may want/need to include other options like 'no_pty' in your authorized_keys file (see man page for options). -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Locating Debian kernel .config
On Mon, Dec 19, 2005 at 10:20:08AM +, Jon Dowland wrote: > Redhat puts the configs used in the srpm under ./configs ; perhaps > putting the debian config file in a seperate package (kernel-config-*) > would be a good idea. It seems like the easiest solution would be to simply ensure that the distributed kernels are compiled to put config.gz in /proc. That would certainly remove any ambiguity about which config file to look at. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://pks.mtholyoke.edu:11371/pks/lookup?search=0xB6D365A1&op=vindex -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: quickly partition disk, copy an image, install grub
On Tue, Sep 20, 2005 at 09:20:33AM -0400, Matt Price wrote: > Hi, > > I have to install ubuntu or something similar on about 20 aging > workstations without cd drives. THese are donated boxes with small hard > drives (as small as 2.1 gig, but not all identical) all wiped clean. Have you looked at System Installation Suite? http://wiki.sisuite.org/ apt-cache search systemimager Basic idea: * Make a golden master * Make an install CD that gets clones minimally booted and networked * Image the machine from the master Takes care of hostname tweaks etc. The instructions provide enought detail to make it look complicated, but when you get down to it, it's really easy. dd is also an option, be careful about bad sectors etc. Also presumes identical disks. Otherwise, do something like: * create master. make note of it's ip/network hostname * boot clone w/ knoppix cd * get shell, and 'sudo -s' to be root * ssh [EMAIL PROTECTED] 'sfdisk -d /dev/hda' | sfdisk /dev/hda * maybe do 'fdisk /dev/hda' and a write (type 'w') to make sure you call ioctl so kernel re-reads partition table * mke2fs as req'd * mkswap as req'd * mount /hda/ * mount /hda/ so you end up with something like /mnt/newroot /mnt/newroot/home /mnt/newroot/var etc. Then rsync data over: * rsync -avz --hard-links [EMAIL PROTECTED]:/ /mnt/newroot Then edit /mnt/newroot/etc/ files that make specific mention of hostname or ip address as req'd. This is from memory, so I'm sure I skipped or mistyped some crucial step that will drive you mad. Oh, like the mbr... Maybe something like this to copy the boot record: * ssh [EMAIL PROTECTED] 'dd if=/dev/hda bs=512k count=1' | dd of=/dev/hda -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Experiences with LSI / Adaptec cards
On Tue, Sep 20, 2005 at 04:34:03PM +1000, David Harrison wrote: > I was wondering if anyone had any experience with either of the > following cards (do they work, do they suck etc) : > > MegaRAID SCSI 320-1LP > > MegaRAID SCSI 320-2E > > I'm interested to hear if anyone has any good experiences with Adaptec > cards or models to recommend, I haven't seen a single list posting that > seems to recommend them. I've lately been using the MegaRAID cards DELL ships in it's 2800's and 2850's. We're using the PCI-E riser, which equates to the 320-2e. You can purchase a version of this adapter built into the riser, and you can purchase the adapter as an expansion card. We're a little concerned about the expansion card because under load, as it continually throws battery fast recharging and battery temperature out-of-range errors. The adapter built into the riser doesn't do this. I think this may be related to the case design, which doesn't seem to provide very good air circulation to the expansion cards. For comparison, we also purchased an Adaptec 39320A. The LSI cards are more configurable. In our early testing, the Linux drivers for the Adaptec weren't up to snuff. I think the drivers have been improved since, but we were past our decision point by then. If anyone has a MegaRAID 320-2e in production which you're monitoring w/ megamon, I'd love to know if you're getting any error reports. Those two cards seem to be it for the internal Ultra320 RAID market. I wouldn't say either of them 'suck', as you say, but I wouldn't give either one an A+ either. I'd like to spend more time w/ the Adaptec now that the drivers seem to have been fixed, but haven't found the time... -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: runlevel policy
On Mon, Sep 19, 2005 at 01:45:19PM -0400, Joe Smith wrote: > > "Paul E Condon" <[EMAIL PROTECTED]> wrote in message > news:[EMAIL PROTECTED] > >On Mon, Sep 19, 2005 at 11:11:55AM -0400, Ron Peterson wrote: > >>On Mon, Sep 19, 2005 at 10:40:46AM -0400, Ron Peterson wrote: > >> > >>> Can anyone explain why Debian's runlevel policy seems to have strayed > >>> so far from traditional System V? Why is xdm/gdm/kdm etc. in runlevel > >>> three, for example? > > > >Debian, as a distribution, really doesn't use runlevels 3-5, but it does > >set them up in skeletal fashion as a convenience to a local sysadmin who > >wants to use them for local purposes. Mostly what is there is just what > >Debian puts in runlevel 2. It saves a bit if copying by the sysadmin. I > >don't think it is governed by policy. The policy is that 3-5 are for > >local use. > Basically the debian point of view is that runlevels are the wrong solution > to the problem. Perhaps the BSD's are better in this regard. However, to date, Debian remains a System V derivative. I would rather see Debian chuck the runlevel concept altogether than capriciously cripple it (out of spite, perhaps.. :) > If you need to be able to interactive decide what services and subsystems > start up automatically, then you should use a truely interactive booting > system. At the moment, I'm not really interested in an interactive boot. But I do think it's a bit crazy that booting single mode attempts to start network services and nfs. That can be a crippling position to be in if you have nfs problems... > Part of the reason for the trditional system is that packages are qquite > hard to remove in RedHat-style systems. If a debian sysadmin does not what > a display manager (Funny name, as they are more like a login-manager) they > can just uninstall it. > > The vast majority of the time a 'traditional' system is booted into either > runlevel 5 or runlevel 3. So often in fact that making all debian runlevels > equivlent to that makes good sense. Basically it makes life easier on the > end user. In the quite rare case that some of the services are unwanted the > services can be stopped by hand or 'init=/bin/bash' can be used. I usually have a banana for breakfast. So the only breakfast food I should have in the house is a banana. If I don't want a banana, I should make a special trip to the store, or scrounge around for a couple of raisins. ?? The (hopefully) quite rare case you mention can be the case when you have to troubleshoot an important production server. In such a case, doing an init=/bin/bash doesn't give you a lot to work with. I'd rather have init=init, unless I'm truly desperate. Sure, once a machine is in production, you generally leave it alone. But during development/testing/troubleshooting, it's convenient to be able to easily forklift a number of services in or out of use. Personally, I like something like: 1 - no networking 2 - networking, local filesystems only 3 - networking + nfs 4 - 5 - networking + nfs + display manager There may be other non-runlevel ways to do the same thing. Point is, Debian currently provides neither; it has to be done by hand. Not terribly difficult work, but of course that's all relative to how much experience you have. I don't agree that it makes sense to make all the runlevels the same. Whether runlevels are the best way to determine boot options or not, they are currently the de-facto way it's done, and making them different is useful. I am certainly open to other better ways of accomplishing the same thing. Best. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: runlevel policy
On Mon, Sep 19, 2005 at 10:40:46AM -0400, Ron Peterson wrote: > Can anyone explain why Debian's runlevel policy seems to have strayed > so far from traditional System V? Why is xdm/gdm/kdm etc. in runlevel > three, for example? I realize the concept of 'traditional' System V runlevel policy is dubious, at best... -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
runlevel policy
To my mind, there is too much stuff in /etc/rcS.d/. In particular, I don't understand what network applications are doing there. If I boot to single user mode, I don't want networking, I certainly don't want NFS, etc. So what's all that cruft doing in there? Can anyone explain why Debian's runlevel policy seems to have strayed so far from traditional System V? Why is xdm/gdm/kdm etc. in runlevel three, for example? -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: scp: file download history
On Mon, Sep 19, 2005 at 02:11:31AM -0700, Jared Hall wrote: > Debian Users- > > I am trying to find a file which was uploaded to my server a couple of > days ago over ssh using the scp command syntax. I cannot remember the > name of the file, or where it was placed. > > Is there any place I can look to see what files were uploaded to my > computer via scp? > Is there a command or log file which will show me all files which have > been uploaded? > > I have another question too, about a certain missing php.ini file, but > I'll put that into a different email. You might check to see if the scp command is in your command history. In bash, something like: history | grep scp Best. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: my first raid disaster on reboot :o( update
On Thu, Sep 08, 2005 at 02:11:26PM +0100, Ken Walker wrote: > I'm getting confused again. Happens to me all the time... > I installed Debian 3.1 onto two SCSI drives set up as raid1. > > I also set-up the four ide drives, during installation and set them as > > /dev/md7 using /dev/hda,/dev/hdc > /dev/md8 using /dev/hab,/dev/hdd > > both ext3 > > they started, and sync'd > > on reboot, md7 and md8 didn't auto start. If you do 'cfdisk /dev/hda' (and b,c,d...), does it show the partition type set to 'Linux raid autodetect'? -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: sendmail trouble
On Wed, Sep 07, 2005 at 11:11:20PM -0600, Bob Proulx wrote: > While the m4 macros cover most common things they don't cover every > possible case. Here is an example. Years ago in sendmail I needed to > deliver mail in a particular way. I needed machines in the same > domain to be delivered directly with smtp but mail outside the domain > to be delivered to a smart host. I won't get into the DNS issues with > why using MX records was insufficient in my case. Here is the best > configuration I found at the time for sendmail. There was no m4 macro > support for this. That was a while ago and who knows but there might > be now. Unless I misunderstand what you were trying to do, these days you'd use sendmail's mailertable feature to do that. Your mc file would contain FEATURE(`mailertable')dnl and then you could construct a mailertable per the simple syntax described here: http://www.sendmail.org/m4/mailertables.html I don't know Postfix or Exim well enough yet to have an opinion about their relative merits. Getting to know them better is on my infinitely long to-do list though... -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: utility for temperature
On Wed, Sep 07, 2005 at 10:54:46AM +0200, roberto wrote: > actually i take control of proc temperature via the following command > > :~$ cat /proc/acpi/thermal_zone/THM/temperature > > but i'd like to ask if there exists some X utility to do this in KDE, > by some other program which uses that command or something similar... Not answering your question, but on a related note, the hddtemp package is nice for monitoring hard drive temps. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: eth1 (3com905B) question
On Wed, Sep 07, 2005 at 03:26:43PM -0700, Andrew Sackville-West wrote: > Arsalan Lodhi wrote: > >i guess not -- htis is what i've > > > >auto lo > >iface lo inet loopback > > > ># The primary network interface > >auto eth0 > >iface eth0 inet dhcp > > > >what should i add in ? > You must have an entry for eth1 in /etc/network/interfaces for it to be > activated. > > man interfaces > > is a good starting point. The entry will depend on what you are doing > with that interface. That is, is it a fixed ip or using dhcp? do you > want it to start automatically? or manually? etc etc etc. You can use ifconfig to do this stuff manually also, but that can be a pain. Your eth0 setup shows what to put in the 'interfaces' file if you want to use dhcp. To assign a static address, you'd do something like: auto eth1 iface eth1 inet static address 172.18.1.35 netmask 255.255.0.0 network 172.18.0.0 broadcast 172.18.255.255 gateway 172.18.100.1 pre-up /usr/local/sbin/up-firewall.sh The 'auto' line says to bring the interface up automatically at boot. Comment that out if you want to bring the interface up/down manually, e.g. 'ifup eth1'. There's more here than you need; you'll want to chop the pre-up bit unless you're familiar with how to set up an iptables script, for example. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Hardware RAID advices needed
On Wed, Sep 07, 2005 at 11:39:57PM +0200, Kjetil Kjernsmo wrote: > I'm not doing it for performance, just for reliability. Good idea. Make sure you get something with monitoring tools that will alert you if you lose a disk. RAID doesn't help much when you're running with a blown disk and don't know it. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: offline ?? Re: sendmail trouble
On Wed, Sep 07, 2005 at 04:40:11PM -0700, Alvin Oga wrote: > > FEATURE(`greet_pause', `1')dnl > > i don't use greet_pause .. i have no time to sit and test it out .. :-) It's one of the best anti-spam features in sendmail, and pretty fool proof. The main thing you need to remember is to properly modify your access file to set pause to zero for localhost and probably your local subnet. If you do happen to find it's blocking a legitimate sender, it's easy to whitelist them in the access db. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
what startup scripts attempt to update sendmail config?
Debian Sarge.. On boot, apparently right after the networking init script, and quite some time before the actual sendmail init script, something appears to be attempting to update various sendmail db's in /etc/mail. What's doing this? -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apache Sigterm every sunday at 6:25
On Mon, Aug 22, 2005 at 05:52:43PM +0200, Thomas Finkl wrote: > Every Sunday at 6:25 > > [Sun Aug 21 06:25:10 2005] [notice] caught SIGTERM, shutting down > -- > [Sun Aug 21 06:25:10 2005] [error] Init: Unable to read pass phrase > [Hint: key introduced or changed before restart?] > [Sun Aug 21 06:25:10 2005] [error] SSL Library Error: 218710120 > error:0D094068:asn1 encoding routines:d2i_ASN1_SET:bad tag > [Sun Aug 21 06:25:10 2005] [error] SSL Library Error: 218529960 > error:0D0680A8:asn1 encoding routines:ASN1_CHECK_TLEN:wrong tag > [Sun Aug 21 06:25:10 2005] [error] SSL Library Error: 218595386 > error:0D07803A:asn1 encoding routines:ASN1_ITEM_EX_D2I:nested asn1 error > [Sun Aug 21 06:25:10 2005] [error] SSL Library Error: 218734605 > error:0D09A00D:asn1 encoding routines:d2i_PrivateKey:ASN1 lib > > Where does the sigterm come from? cron. Look at what /etc/crontab runs at that time. > Where will I have to write the "pass phrease" Sounds like your SSL setup uses a cert with a passphrase. Probably the easiest way to avoid manual intervention would be to make a passphraseless cert. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: control-b in konsole
On Fri, Aug 12, 2005 at 11:28:51AM +0200, Maurits van Rees wrote: > On Tue, Aug 09, 2005 at 09:38:13AM -0400, Ron Peterson wrote: > > There is a ctrl-b function in konsole that brings up a little dialog > > "Add Bookmark - Konsole". Where it comes from, I don't know. How to > > shut it off, I don't know either. For an emacs bigot such as myself, > > it's extremely annoying. > > > > It's not always turned on. There's something I'm doing, or a > > configuration setting that gets activated somehow that turns it on. I > > can't find it to save my life. I'm dying. Please help. > > I don't have this problem and can happily use CTRL-B in Konsole to > move backwards. But you could look in Configuration -> Speedkeys (I'm > guessing the names as I have a Dutch language konsole) to see if Add > Bookmark is there somewhere, linked with CTRL-B. I don't see it > listed on my system though, not even with an empty entry for the > speedkey. I think the equivalant on my system is Settings/Configure Shortcuts... I've searched that dialog until my eyes hurt. Nothing there. :( > I do see the following in ~/.kde/share/config/kdeglobals: > AddBookmark=Ctrl+B > > I'm not sure where you would normally change this, but you could > probably just edit this with your favorite editor and either delete > this line or change it to use a different key combo. You may need to > restart KDE or at least fire up Konsole again. Thanks. I don't have that line in my kdeglobals. I recursively grep'd for addbookmark (-i of course) in /etc/kde3 also, w/ no luck. I added an AddBookmark entry in my kdeglobals, logged out and back into kde again. My newly defined shortcut doesn't seem to work. I've been trying Kubuntu on my laptop. This little gotcha is driving me so nuts, by itself it may drive me to revert to good ol' vanilla debian, which hasn't given me this problem on any system I'm running it on. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: samba open share
On Wed, Aug 10, 2005 at 12:42:53PM -0500, Rodney Richison wrote: > I'm having trouble getting samba to quit asking for a name and password > on an open share. > > How may I do this without affecting the other shares that do have name > and password verification? > The below does not work. > [public] >writable = yes >path = /backup/files >force group = rodney >guest ok = yes >force user = rodney >hosts deny = 0.0.0.0/0 >hosts allow = 127.0.0.1 192.168.1.0/24 >public = yes >browsable = yes See the 'map to guest' option in the smb.conf man page. You probably want to use the 'Bad User' option. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Serial comm program
On Wed, Aug 10, 2005 at 11:50:55AM -0400, rpeterso wrote: > Are you familiar with 'screen'? BTW, you can do serial stuff w/ screen directly, if you like. E.g.: 1017# cat /local/etc/screenrc.serial # This assumes that serialuser has proper # permissions to access the serial ports and to # write to the log files specified in the screenrc. # turn logging on for all windows deflog on # tell screen to log to /var/log/serial.$WINDOW logfile /var/log/serial.%n # open windows on the serial ports screen -t ttyS0 /dev/ttyS0 38400 screen -t ttyS1 /dev/ttyS1 19200 screen -t USB0 /dev/ttyUSB0 38400 screen -t USB1 /dev/ttyUSB1 38400 screen -t USB2 /dev/ttyUSB1 38400 screen -t USB3 /dev/ttyUSB1 38400 # if you like, launch at system startup rc script like: # su serialuser -c 'screen -dmS serial -c /local/etc/screenrc.serial' escape ^Tt The 'su ..' startup line above would launch screen using the specified config file, and put it in detached mode. Then you could log in and attach at any time by running 'screen -r serial'. Otherwise just 'screen -c ' I haven't done much serial stuff w/ screen myself, not sure how you set parity bits etc., but it's an interesting approach in any case.. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Serial comm program
On Wed, Aug 10, 2005 at 07:32:56AM -0400, Daniel D Jones wrote: > On Wednesday 10 August 2005 01:13 am, Karl O. Pinc wrote: > > On 08/10/2005 12:27:50 AM, Chris Palmer wrote: > ... I SSH into that server, then telnet into the device. When > I'm working on a device, being able to scroll back to stuff that's scrolled > off the screen is vital. We're required to log everything we do to the > devices, so logging at all times is necessary. There are a number of actions > that are highly repetitive. Scripting them makes it much easier and reduces > the chances of error. I also need to paste in long configs. If the info is > pasted in too fast, the router or switch will drop characters, so I need > something with a configurable delay after pasting characters or lines from > the clipboard. Most of the time, I upload new IOS files via TFTP. If a > router crashes or the flash memory croaks or for whatever reason it will only > come up to the rommon> prompt, it may require transfer via xmodem. etc, etc. Are you familiar with 'screen'? It's a bitch to google for help on the damn program, on account of the pedestrian name, but including the string 'screenrc' (the config file name) often helps. One hint: if you're use Ctrl-A a lot, e.g. in emacs or readline etc., bind screen's escape key to something else, like ctrl-T, by appending the following the /etc/screenrc (or ~/.screenrc, as you see fit): escape ^Tt I've been known to use netcat to upload text to cisco devices. Something like: netcat hostname 23 < script.txt where script.txt looks like: password term len 0 sho mac-address-table exit I know I'm not answering your original question; just thought I'd add further confusion by introducing even more options... ;) -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
control-b in konsole
There is a ctrl-b function in konsole that brings up a little dialog "Add Bookmark - Konsole". Where it comes from, I don't know. How to shut it off, I don't know either. For an emacs bigot such as myself, it's extremely annoying. It's not always turned on. There's something I'm doing, or a configuration setting that gets activated somehow that turns it on. I can't find it to save my life. I'm dying. Please help. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: use knoppix to rerun lilo
On Tue, Jul 26, 2005 at 09:28:19PM +0200, Felix Natter wrote: > Anyway, usually I can easily fix this by booting Knoppix (3.1 in my case), > mounting the root filesystem /dev/hda2 (it's all on one partition), > and running /sbin/lilo from a chroot-shell (chroot /mnt/hda2 /bin/bash). > But when I do this, I get the error: > Fatal: open /dev/hda: Permission denied Try the '-o dev' option to mount. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: pam userdb auth issue (pam_userdb can't open database) Sarge
On Sat, Jul 23, 2005 at 10:12:21AM -0300, Tom wrote: > > Im with the same problem (but in woody). I saw no one answered you > here. Did you solve the problme by some another way? The shadow password file is readable by the 'shadow' group. Make sure that any users (e.g. the user your program runs as) are in the shadow group. Make sure you don't inadvertently expose the shadow password file. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: first nfs mount in fstab fails
On Mon, Jul 11, 2005 at 08:46:09PM -0400, Ron Peterson wrote: > On Fri, Jul 08, 2005 at 11:16:19PM +0100, Wackojacko wrote: > > >I have an fstab, which after the usual local disk entries has nfs > > >entries that look like: > > > > > >aservername:/s /s nfs rw,bg,intr,proto=tcp 0 0 > > >aservername:/alum /alum nfs rw,bg,intr,proto=tcp 0 0 > > >aservername:/users /users nfs rw,bg,intr,proto=tcp 0 0 > > >aservername:/u2/u2nfs rw,bg,intr,proto=tcp 0 0 > > >etc. > > > > > >On boot, the first mount consistently fails to mount. It manually > > >mounts fine. If I switch entries around, the same thing happens - the > > >first entry fails to mount. > > > > > >Any ideas? > > > > > Is the NFS support compiled into the kernel or loaded as a module. If its > > a module then its possible it isn't loaded in time for the first line in > > fstab, but is loaded by the time the rest are called, or you mount > > manually. I had a similar problem with a Nvidia driver and X which was > > solved by putting the module in /etc/modules so it loads on boot. > > > > Could also try compiling support in the kernel if this doesn't work. > > Thanks. > > I'm using nfs-kernel-server, and I have the relevent nfs kernel stuff > statically linked... :( > > I'm mounting exports from a Tru64 alpha to Debian sarge on a Dell 2800, > BTW. I forgot to try making the first nfs mount a different OS at work > today, and I don't want to reboot from home (Murphy's law...). I'm > going to try that tomorrow and see if that works better. I have the same problem if the first mount is exported from another Debian box. Whichever mount I put first is not mounted at boot, but will mount fine manually afterwards. ?? -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: PostgreSQL 8.0 bug or my own stupidity?
On Mon, Jul 11, 2005 at 08:33:41PM -0400, Ron Peterson wrote: > mymap yourusername postgres > mymap postgres postgres > mymap root postgres (I would normally only do this for root and postgres, btw. I don't typically give my username special priviledges...) -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: first nfs mount in fstab fails
On Fri, Jul 08, 2005 at 11:16:19PM +0100, Wackojacko wrote: > >I have an fstab, which after the usual local disk entries has nfs > >entries that look like: > > > >aservername:/s /s nfs rw,bg,intr,proto=tcp 0 0 > >aservername:/alum /alum nfs rw,bg,intr,proto=tcp 0 0 > >aservername:/users /users nfs rw,bg,intr,proto=tcp 0 0 > >aservername:/u2/u2nfs rw,bg,intr,proto=tcp 0 0 > >etc. > > > >On boot, the first mount consistently fails to mount. It manually > >mounts fine. If I switch entries around, the same thing happens - the > >first entry fails to mount. > > > >Any ideas? > > > Is the NFS support compiled into the kernel or loaded as a module. If its > a module then its possible it isn't loaded in time for the first line in > fstab, but is loaded by the time the rest are called, or you mount > manually. I had a similar problem with a Nvidia driver and X which was > solved by putting the module in /etc/modules so it loads on boot. > > Could also try compiling support in the kernel if this doesn't work. Thanks. I'm using nfs-kernel-server, and I have the relevent nfs kernel stuff statically linked... :( I'm mounting exports from a Tru64 alpha to Debian sarge on a Dell 2800, BTW. I forgot to try making the first nfs mount a different OS at work today, and I don't want to reboot from home (Murphy's law...). I'm going to try that tomorrow and see if that works better. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: PostgreSQL 8.0 bug or my own stupidity?
On Mon, Jul 11, 2005 at 05:56:43PM +0200, Tom wrote: > Quite some lines like > > FATAL: user "tom" does not exist > FATAL: database "cds" does not exist > FATAL: user "root" does not exist > FATAL: user "root" does not exist > FATAL: user "tom" does not exist > FATAL: Ident authentication failed for user "postgres" etc. You can save yourself some trouble screwing around w/ the 'postgres' user account by creating an ident map which relates the shell account you normally use with the postgres user. Edit pg_ident.conf to look something like: mymap yourusername postgres mymap postgres postgres mymap root postgres Then edit pg_hba.conf to look something like: local all all ident mymap This says that when connecting to any local database as any user, do ident authentication, and refer to the 'mymap' entries in pg_ident.conf for doing the lookups. So now, logged in as yourusername or root, for example, you can do: psql -U postgres -d template1 -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Ethernet Speed and Duplex
On Mon, Jul 11, 2005 at 06:12:08PM -0400, Joe Grace wrote: > Ok, I am going crazy because I have not been able to figure this out. > I have googled this like crazy. I have a dsl modem that is connected > directly to an ethernet card. I can not figure out how to manually > set the duplex and speed. In the past I have used mii-tool, but when > I try to use it, it will say "no mii transceivers found". In my > google searches I found that there were some people who were using > ethtool. When I try setting the duplex and speed with ethtool it does > not display a word when I press return and then duplex and speed stay > the same, even after bringing the interface down and up. For some > reason with this card auto-negotation is very slow. Can someone point > me in the right direction? Can I set the duplex and speed in /etc/ > network/interfaces somewhere on the "auto eth0" line? What kind of card is it? What does 'ethtool eth0' (assuming your interface is eth0) tell you? What duplex/speed does your DSL modem run at? Are you using a crossover cable? -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
first nfs mount in fstab fails
I have an fstab, which after the usual local disk entries has nfs entries that look like: aservername:/s /s nfs rw,bg,intr,proto=tcp 0 0 aservername:/alum /alum nfs rw,bg,intr,proto=tcp 0 0 aservername:/users /users nfs rw,bg,intr,proto=tcp 0 0 aservername:/u2/u2nfs rw,bg,intr,proto=tcp 0 0 etc. On boot, the first mount consistently fails to mount. It manually mounts fine. If I switch entries around, the same thing happens - the first entry fails to mount. Any ideas? -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
thread indicator symbols in mutt
My mutt installation isn't displaying the character symbols used in thread mode to show relationships between messages; I'm seeing other special characters instead. What do I do to fix this? TIA. -- Ron Peterson Network & Systems Manager Mount Holyoke College http://www.mtholyoke.edu/~rpeterso -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: debian: size of stable, testing and unstable files
Ilya Martynov wrote: > > RP> If you'd like more precise info about my rsync inclusions and > RP> exclusions, let me know. > > Can you post your rsync inclusions and exclusions? I've recently tried > to setup local mirror and I'd like to compare mine rsync setup with > yours. You can view my mirror scripts at http://mill.mtholyoke.edu/mirror/. I'm not going to claim these are the best script in the world, but they seem to work for me. The include/exclude files are there also, they're just not visible in the index. Just explicitly enter their name in the URL to see them. -- Ron Peterson Network & Systems Manager Mount Holyoke College GPG and other info at http://www.mtholyoke.edu/~rpeterso
Re: debian: size of stable, testing and unstable files
Brendan J Simon wrote: > > I'm considering mirroring Debian for our company. I'd like to know how > much diskspace I would require. Can someone tell me the size of the > following Debian Distros. Is there a webpage with this information ?? > Is it automagically updated ??? I mirror debian, debian-non-US, and Ivan's KDE work for i386 and alpha. Here's today's stats: 113$ du -h --max-depth 2 . 5.3G./debian/pool 7.0G./debian/dists 3.8M./debian/doc 3.1M./debian/indices 668k./debian/tools 12k ./debian/project 12G ./debian 84M ./debian-non-US/dists 192k./debian-non-US/indices-non-US 158M./debian-non-US/pool 12k ./debian-non-US/project 242M./debian-non-US 12k ./debian-kde/project 296M./debian-kde/debian 296M./debian-kde 13G . If you'd like more precise info about my rsync inclusions and exclusions, let me know. -Ron- GPG and other info at: http://www.yellowbank.com/
xfree86 and woody
What's the story with xfree86 and woody? I don't see any debs for xfree86 in that tree. There are various .deb's in pool, but the most recent ones don't seem to be associated with any dist, including unstable. I've tried searching the list archives on this, but the search page keeps stalling on me. I'm trying to convert to Debian from Red Hat. I'm contemplating installing Debian on a pool of workstations in a computer lab at the university where I work. The issue that's concerning me is that I need to get an up-to-date desktop running. I'd prefer to use Debian because I like it's package management system better. But if I have to kludge a desktop, that kind of obviates the advantage. The machines will all be new, so I expect xfree86-4.x will support the hardware better. Ivan's excellent KDE work appears to be geared to run on 3.x, the stable potato version of xfree86. I haven't checked out the Gnome situation yet. So let's see, what's my question? ;) Any advice? -Ron- GPG and other info at: http://www.yellowbank.com/
duplicate packages found via sources.list
What happens if multiple URI's in sources.list can be used to find a particular package? Does the first one win? What if you find the same package twice, but different versions? Is there a way to say 'apt-get install most-recent'? In my case, I want to install the most recent kdebase. I have URI's to both the debian and kde packages in my sources.list. Ivan's kde packages are more recent, and are the ones I want to use. Do I need to comment out everything but the KDE URI in sources.list? What's the 'proper' way to handle this situation? -- Ron Peterson Network & Systems Manager Mount Holyoke College GPG and other info at http://www.mtholyoke.edu/~rpeterso
what is 'pool'?
What is 'pool'? The Debian FAQ refers to doc/, project/, tools, etc. but makes no mention of pool/. I've searched the debian-user archive for 'pool', and only found incidental reference to it. -- Ron Peterson Network & Systems Manager Mount Holyoke College GPG and other info at http://www.mtholyoke.edu/~rpeterso
debian kernel modifications
Debian FAQ item 7.2 says that "The kernel (filesystem) in Debian GNU/Linux systems supports replacing files even while they're being used." How is this accomplished? If I roll my own kernel, do I need to patch it first to support this functionality? Are there any other Debian-specific kernel patches to be aware of? -- Ron Peterson Network & Systems Manager Mount Holyoke College GPG and other info at http://www.mtholyoke.edu/~rpeterso
Re: Booting past 1024?
Peter Welte wrote: > > Hey... > im installing Debian on a system that has a 30 GB > hard drive, and the first 10 GB is used by NTFS (which > i cant delete). If i install Debian past 8 GB, and > lilo on the mbr, is it true that lilo wont be able to > start linux? > if so... i can still make a /boot on the first part > of the hard drive, right? (by specifying [Beginning] > durring the debian install)? If your BIOS supports large hard drives, then you may use lilo's 'linear' or 'lba32' option to enable use of the drive. These options are mutually exclusive. I'm running potato, and 'man lilo.conf' doesn't provide information about the lba32 option. You can find out more by reading the lilo manual, though. This is probably in /usr/share/doc/lilo. This is a good document to read anyway if you'd like to know how disk partitioning and addressing works. You shouldn't need to make /boot in it's own partition. If you specify 'beginning', as you say, when creating the a /boot partition, it will be created at the beginning of the as-of-yet unpartitioned space. Which is what you'd want. Otherwise you'd be overwriting your existing NTFS partition. This might have its advantages, but I'm guessing that's not what you want to do. -- Ron Peterson Network & Systems Manager Mount Holyoke College GPG and other info at http://www.mtholyoke.edu/~rpeterso
Re: upgrading libc6 from 2.1 to 2.2 - trouble brewing?
Colin Cashman wrote: > > Tonight I tried to install the latest version of ssh onto my system, but I > wasn't paying close enough attention to the dependencies. Ssh 2.5.1p1 > requires libc-2.2.1-2 or higher, and potato uses libc6-2.1.3-15. > > I have a vague recollection of problems running apps using libc-2.1 with > libc-2.2, so I thought I'd ask the list since I'm sure at least one person > has tried it. :) > > Should I expect any trouble if I simply upgrade to libc-2.2? If so, what is > the best strategy to resolve those problems? I believe you can compile ssh 2.5.1p1 on a system using libc-2.1. If you don't mind compiling. -Ron- GPG and other info at: http://www.yellowbank.com/
Re: debian on del poweredge 4200
andy wrote: > > i took the resultant megaraid.o file, stuck it on a floppy in /boot, > then tried to load it at the appropriate point in the install. load failed. I believe the module needs to exist on the floppy in the same folder as in the final install. E.G. - /lib/modules/2.2.18pre21/scsi/ for scsi drivers .../net for network drivers, etc. -Ron- GPG and other info at: http://www.yellowbank.com/
Re: shell script experts pls help a lost soul
john smith wrote: > > hi, > > I am having some problems creating some sample scripts...hoping that > somebody can lend a helping hand. > > #!/bin/sh > if test -x $1 > then whoami; > echo $1 > fi > > Yes I know my script sucks but I'm just learning..anyway, my questions are: > > 1. my script doesn't seem to display the script name together with > whoami..actually, what the script was supposed to do is when it is invoked, > it will display the user name together with the scriptname should the > scriptname gets changed somehow..it displays the username but not the > scriptname itself. The name of the script is $0. $1 is the first argument. > 2. how do I add to this script the count of the number of lines in a file? > i.e. let's say cat's man page. man wc. > 3.lastly,let's say I want to print the first line of each 3-char.h file in > /usr/include/lib..how can I accomplish this? I'm not sure what you're asking exactly. Perhaps something like for i in *.h; do this; do that; do the other thing; done See 'man bash'. for name [ in word ] ; do list ; done See also 'man head'. > and Yes, this is part of my assignment I hope I still left something to the imagination... -Ron- GPG and other info at: http://www.yellowbank.com/
Re: [OT] postgresql triggers
will trillich wrote: > when i expand this and try to add a SELECT on other tables, or an > UPDATE to another table or two, postgresql gripes about > > NEW used in non-RULE query... Could you provide an example of what *doesn't* work? -Ron- GPG and other info at: http://www.yellowbank.com/
Re: unresolved symbols in 3dfx.o
Johnny Blade wrote: > > I've been attempting to prepare my machine to play > some quake, but I can't get the 3dfx.o module to work. > I downloaded the device3dfx source using apt-get, ran > the buildpkg script, and then installed the .deb > created. As the deb is trying to update-modules, I > see a message saying "Unresolved symbols in > lib/.../3dfx.o". When I depmod -e this (I've been > hitting the man pages...), the two symbols it mentions > are "register_chrdev_Rbbdcb79f" and > "printk_R1b7d4074". > > In the the archives someone mentioned running nm > 3dfx.o | grep printk, and then grep printk /proc/ksyms > to see if the kernel headers that were used match. > Mine don't. I'm not sure how to fix that, because my > kernel source is in /usr/src/linux (actually > kernel-source-2.2.18pre21, but it's symlinked) where > the programs should be able to find the right files. > I just recently recompiled my kernel because the > makefile wouldn't run without the kernel having been > configured and installed, so I dont' think restoring > my old kernel would help. You might try restoring your original kernel, and just installing the kernel headers, to allow you to compile. You'll probably need to -I/usr/src/linux/include. If you're compiling using make, try 'export CFLAGS=-I/usr/src/linux/include' before running make. When you build your own kernel, there's an option for whether you want the kernel to set version information on module symbols. The way that was set when you rolled your kernel could be what's giving you problems. -Ron- GPG and other info at: http://www.yellowbank.com/
Re: compiling a program
Philipp Bliedung wrote: > > Hi, > > I have simple question concerning compiling a programm > When I compile a program from the sources and it beakes for whatever > reason, how can I get rid of all the files that were created so far or > how can I remove this 'broken' program ? Is there anything like > 'apt-get -f install' which is used for .deb files ? If the maintainer put a make directive (e.g. 'make uninstall') to do so in their makefile perhaps. Otherwise not usually. You can review what happened during a 'make install' (or any other make command, for that matter), by using make's -n option. I.E. - 'make -n install'. This will just echo what would happen, without really doing it. Problems like these are why package management systems exist. > And how can I find out if a new version of a package, for example libc6, > depends on any other package? For example I tried to compile a > movie-player but it needed a newer version of libc6, which I don't have, > so how can I find out if I need any other package for libc6 to work? > (I'm still using the 2.2.17 kernel...) Build it and read the error messages. They are not always easy to decipher, especially if you don't know the language. Again a good reason to use package management systems. Of course, that's not always an option... -Ron- GPG and other info at: http://www.yellowbank.com/
Re: kill: cannot kill some processes
William T Wilson wrote: > > On Thu, 22 Feb 2001, brian moore wrote: > > > > does the process list "Z" under STAT ? if it is the process has gone > > > zombied and i don't think there is much you can do. sometimes zombie'd > > > processes die on their own eventually many times they will not die until > > > you reboot .. > > > > Not quite true... zombies don't ever die: they're already dead. > > While the description of zombie processes is accurate, I think another > likely situation is that the process is in "uninterruptible sleep," i.e. > the 'D' state. This happens when a process is blocked in a system call - > it will be 'D' until the kernel function returns. Kernel bugs, hardware > problems, and dead NFS mounts can cause these kernel functions to take > a long time or forever. > > In such a case, you really are stuck; unless the resource the process is > waiting for frees up, it's going to hang around until a reboot. > > One thing about zombie process: Don't worry about trying to "make" them go > away. They don't consume any CPU time, or any other resources other than > the slot in the process table and the less than 1K of memory required to > hold their state information. They are not worth worrying about. Not entirely true. Init can inherit enough zombie processes that it hits its process limit (1024, if I remember correctly). Can you 'shutdown'? Nope. Not unless you can free up a slot. And if something's going haywire and spawning zombies quickly, this can be a problem. Not a common occurance, though... -Ron- GPG and other info at: http://www.yellowbank.com/
Re: after 'su -', 'Can't open display'
> When I start X Windows, using the KDE window manager, I change to root > (with su - ) > for administrative tasks. However I seem unable to run X window > applications. > > Whatever X application I try to run, I inevitably get a message similar > to "... unable to open display". Try this. While logged in as yourself, type 'xauth nexport - $DISPLAY > myxauth' (sans quotes, of course). Then become root, and type 'xauth nmerge myxauth'. Then delete myxauth. -Ron- GPG and other info at: http://www.yellowbank.com/
Re: 3c905/network problems
Pete Meyer wrote: > > Hi all...I'm having difficulty getting my network card (3c905c) work under > debian. I've tried using the 3c59x driver (I was using the wrong one > before). It installs ok, but then DHCP won't configure the network. I know > that there's a server present, because that's how windows is configured. > I've also tried the 3c90x driver, but can't get it to compile. I was recently trying to get a 3c905cx-txm working. I compiled the module as gcc -c 3c90x.c -O2 -Wall -Wstrict-prototypes -fomit-frame-pointer \ -fno-strength-reduce -pipe -m486 -malign-loops=2 \ -malign-jumps=2 -malign-functions=2 -DCPU=486 \ -DMODULE -D__KERNEL__ \ -DMODVERSIONS \ -I/usr/src/kernel-headers-2.2.18pre21/include That compiled with no errors. I installed the module, updated /etc/modutils/aliases, did update-mdules and depmod -a. No errors. Still didn't work. You have a different card, though, so maybe you'll have better luck. (You might not need the -DMODVERSIONS, I can't remember.) You should have a line like the following in /etc/network/interfaces. You probably know that, but thought I'd mention it. iface eth0 inet dhcp Also make sure your BIOS setting for Plug-and-Pray OS is turned off. -- Ron Peterson Network & Systems Manager Mount Holyoke College GPG and other info at http://www.mtholyoke.edu/~rpeterso
Re: newbie postgresql question
Charles Lewis wrote: > psql: FATAL 1: SetUserId: user 'chas' is not in 'pg_shadow' > > I imagine that I have to add myself to some group or something, but I have > absolutely no idea where to start looking. > > I've read some postgres documentation, but debian apparently does things a > little differently. Any suggestions? This is a PostgreSQL issue, not a debian issue. You need to add 'chas' to the PostgreSQL user database. Do 'createuser chas'. See 'man createuser'. -- Ron Peterson Network & Systems Manager Mount Holyoke College GPG and other info at http://www.mtholyoke.edu/~rpeterso
Re: importing gnuplot eps into microsoft word
Mark Mackenzie wrote: > > Hi, > > I plot a graph with gnuplot3.7, and have: > set terminal postscript eps > set output "prop.eps" > giving me an eps that I can view with gv. > > However, importing this into ms word gives: > Title: > prop.eps > Creator: > gnuplot... > Preview: EPS not saved with preview... > Comment: > This EPS picture will print to a Postscript printer > ... > in an outlined box. > > This text is also printed on my PS printer (hjlj2100). It _will_ print to a postscript printer, but what goes where the ...'s are? I think it says something like _will not_ print to a non-postscript printer. The HP Laserjet 2100 does not support postscript. The 2100M and 2100TN do. A postscript printer should not require an EPS preview be saved in order to print. Just the opposite: a non-postscript printer would require the preview to print. The preview is just a raster image. A non-postscript printer, by definition, does not know how to rasterize postscript, but it might be able to print the (already rasterized) preview. > I never have this problem with lyx, but unfortunately the graph is for > someone else. The only thing that works is to print to a .png, but this > makes a pigs breakfast of the beautiful postscript.' Speaking of pig's breakfast (nice phrase :-), an EPS preview can be enormous. Especially if it's high enough resolution that you get a nice print from it. I've seen people assemble less than 1MB worth of EPS files in Pagemaker, and end up with Pagemaker files that were many hundreds of megabytes, all because they had no idea what was going on with their EPS previews. -Ron- GPG and other info at: http://www.yellowbank.com/
Re: 3com 3c90x install
Nathan E Norman wrote: > > On Wed, Feb 21, 2001 at 10:48:13AM -0500, Ron Peterson wrote: > > I lied. Well, I didn't look closely enough. It's a 3c905cx-txm. > > Anyway, I tried the 3c59x module first, actually, but it didn't work. > > Now I'm not even sure the 3com provided driver will work. I might just > > put in a different card... > > Go get the diagnostic disks from 3com and test the card. You might > find that it's bad; a friend reported a DOA 3c905c the other day. It's a multi-boot machine and it works under the other (unmentionable) OS's. -Ron-
Re: 3com 3c90x install
I tried a different network card - 3c905-tx - with the 3c50x driver module. This doesn't work either. It does, however, spit out some kind of error message. Trouble is, I can't read it, because it appears on the blue screen behind the install dialog. Is there any way to read this? I can see | ot alloc| eth0: Ca| | This a new computer using a new ASUS motherboard (Asus CUSL2-C). Am I having chipset or bios configuration problems? -Ron- Ron Peterson wrote: > > I'm trying to do a network install of debian on a computer which has a > 3com 3c905c-tx network card. 3com kindly provides GNU licenced source > code for this card on their website. I succesfully compiled the module > on another debian system, and put it on a floppy, in directory > /lib/modules/2.2.18pre21/. I inserted this floppy when appropriate > during the 'Configure Device Driver Modules' portion of the > installation. > > However, no network interfaces are found. I imagine I need to pass some > kernel or insmod parameters for the specified module. Where/How do I do > this? > > -- > Ron Peterson > Network & Systems Manager > Mount Holyoke College > GPG and other info at http://www.mtholyoke.edu/~rpeterso > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] -- Ron Peterson Network & Systems Manager Mount Holyoke College GPG and other info at http://www.mtholyoke.edu/~rpeterso
Re: 3com 3c90x install
The horses are already on the track... The error I got was that neither DHCP or manual IP configuration resulted in a system that was able to do a network install. DHCP simply times out and fails. Manual network configuration is simple enough, but doesn't get me anywhere. Link light's on, but nobody's home. Joris Lambrecht wrote: > > hold yr horses, i would be really sursprised if this card would not work > can't you refer to any error's you get ? > > -Original Message- > From: Ron Peterson [mailto:[EMAIL PROTECTED] > Sent: Wednesday, February 21, 2001 4:48 PM > To: John Kuhn > Cc: debian-user@lists.debian.org > Subject: Re: 3com 3c90x install > > I lied. Well, I didn't look closely enough. It's a 3c905cx-txm. > Anyway, I tried the 3c59x module first, actually, but it didn't work. > Now I'm not even sure the 3com provided driver will work. I might just > put in a different card... > > Still wouldn't mind knowing how the third-party module install thing > works, just to satify my curiousity, though. > > -Ron- > > John Kuhn wrote: > > > > You don't need to compile this driver from source. During the 'Configure > > Device Driver Modules' load the '3c59x' module. Yes, the current 3c59x > > module does work with the 3c905c-tx. I used the 3com source driver > > with older versions of the 2.2.x kernels, but it is not needed for > > 2.2.18pre21. > > > > John > > > > On Wed, Feb 21, 2001 at 09:59:40AM -0500, Ron Peterson wrote: > > > > > I'm trying to do a network install of debian on a computer which has a > > > 3com 3c905c-tx network card. 3com kindly provides GNU licenced source > > > code for this card on their website. I succesfully compiled the module > > > on another debian system, and put it on a floppy, in directory > > > /lib/modules/2.2.18pre21/. I inserted this floppy when appropriate > > > during the 'Configure Device Driver Modules' portion of the > > > installation. > > > > > > However, no network interfaces are found. I imagine I need to pass some > > > kernel or insmod parameters for the specified module. Where/How do I do > > > this?
Re: 3com 3c90x install
I lied. Well, I didn't look closely enough. It's a 3c905cx-txm. Anyway, I tried the 3c59x module first, actually, but it didn't work. Now I'm not even sure the 3com provided driver will work. I might just put in a different card... Still wouldn't mind knowing how the third-party module install thing works, just to satify my curiousity, though. -Ron- John Kuhn wrote: > > You don't need to compile this driver from source. During the 'Configure > Device Driver Modules' load the '3c59x' module. Yes, the current 3c59x > module does work with the 3c905c-tx. I used the 3com source driver > with older versions of the 2.2.x kernels, but it is not needed for > 2.2.18pre21. > > John > > On Wed, Feb 21, 2001 at 09:59:40AM -0500, Ron Peterson wrote: > > > I'm trying to do a network install of debian on a computer which has a > > 3com 3c905c-tx network card. 3com kindly provides GNU licenced source > > code for this card on their website. I succesfully compiled the module > > on another debian system, and put it on a floppy, in directory > > /lib/modules/2.2.18pre21/. I inserted this floppy when appropriate > > during the 'Configure Device Driver Modules' portion of the > > installation. > > > > However, no network interfaces are found. I imagine I need to pass some > > kernel or insmod parameters for the specified module. Where/How do I do > > this?
3com 3c90x install
I'm trying to do a network install of debian on a computer which has a 3com 3c905c-tx network card. 3com kindly provides GNU licenced source code for this card on their website. I succesfully compiled the module on another debian system, and put it on a floppy, in directory /lib/modules/2.2.18pre21/. I inserted this floppy when appropriate during the 'Configure Device Driver Modules' portion of the installation. However, no network interfaces are found. I imagine I need to pass some kernel or insmod parameters for the specified module. Where/How do I do this? -- Ron Peterson Network & Systems Manager Mount Holyoke College GPG and other info at http://www.mtholyoke.edu/~rpeterso
