serial console problem
I have a collocated server which had serial console access. The server centre re-located the server and left off the serial console. The server worked fine for a few hours, but then access got slower and slower and eventually died. Could it be that the lack of a console is causing that? I'm think the console may want to flush its buffer or is waiting on a question etc etc. Whilst I was able to communicate with it I could see that an fsck was happening so perhaps it's asking permission to clean up or something. Of course I have asked the centre staff to re-connect the consol so I can see what's going on, but that is happening as fast as hell freezes. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
geometry does not match label
I'm just setting up a software raid mirror using geom; everything seems to be working fine and the mirror is slowly synchronizing. Because of cabling constraints this is a mirror rather than a duplex (I haven't used mirror before) so the gm0 components are ad0/ad1 rather than ad0/ad2. In dmesg I see .. GEOM_MIRROR: Device mirror/gm0 launched (1/2). GEOM_MIRROR: Device gm0: rebuilding provider ad1. GEOM: mirror/gm0s1: geometry does not match label (16h,63s != 255h,63s). .. Trying to mount root from ufs:/dev/mirror/gm0s1a .. anyone know if I should worry about the "geometry does not match label" message? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Undefined symbol libintl_bindtextdomain
I am trying to build (FreeBSD 8.0) a bleeding edge python (latest unladen swallow) with gcc44 and am getting mostly good results. However, one of the standard modules is failing to load at runtime because of an undefined symbol libintl_bindtextdomain I looked in /user/local/lib/libintl.so and see that the symbol is there. In the module of interest we don't seem to refer to libintl $ LD_LIBRARY_PATH=`pwd` ldd _locale_failed.so _locale_failed.so: libutil.so.8 => /lib/libutil.so.8 (0x281ad000) libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x281bc000) libc.so.7 => /lib/libc.so.7 (0x28091000) however, there are references in the text of _locale_failed.so. Is there an implication that a library reference is missing during the build of the module? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: binary package dependencies
2009/12/20 Lowell Gilbert : .. >> Somehow I had naively assumed that apache20 and apache22 were >> incompatible and could not simultaneously be installed. Did the binary >> package load ignore all conflicts? What's the proper way to approach >> these issues. Looking in the apache20 Makefile I see it conflicts with >> earlier apache, but how can it conflict with a later one? > > I think that it should. As I read it, apache22 registers a conflict > with apache20, but the reverse is not true. If you had installed them > in the other order, it would've refused to install. apache20 is the > default, so the official package was built depending on that. > > I think this should be entered as a bug, but I'm not quite positive... > clearly A conflict B is supposed to be a symmetric relation, but I guess in this case when I install apache22 it's the entire ports system that needs to record the conflicts. I don't think it's reasonable for an individual port to know that a future conflict may arise. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
binary package dependencies
I would like to know how to use self compiled ports made with make install together with the packages which can be downloaded with pkg_add -r I'm in the process of upgrading an old freebsd 6.0 server to 8.0 and have decided to try and use apache22. For various reasons I built the apache22 server using ports (mainly to force usage of a particular BSDB). Then added subversion also using ports. After setting up the new apache to act as an svn source and getting that working I decided to add viewvc. Rather stupidly I used pkg_add -r viewvc which seemed to work. However, my apache setup stopped working. After much faffing about I learned that the pkg_add -r viewvc had also installed another version of apache (a 2.0 version). All my apachectl commands were directed at the 2.0 version and my edits to the httpd.conf were bing entirely ignored. Somehow I had naively assumed that apache20 and apache22 were incompatible and could not simultaneously be installed. Did the binary package load ignore all conflicts? What's the proper way to approach these issues. Looking in the apache20 Makefile I see it conflicts with earlier apache, but how can it conflict with a later one? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: atom based servers
David Rawling wrote: Hi Brian Indeed, I have a FreeBSD 8.0RC1 system running as my primary time server for the home network. Since it's an Atom 330, it fully supports 64-bit mode (an opportunity I have grasped with both hands). I have one of the Acer ION gadgets running at home and that also uses the Atom 330. I cannot find any nice way to reduce the power consumption though as the 330 doesn't seem to support speedstep and my cpu is always running at 68C. Does your board provide any power control opportunity? The board I happen to be using is an Intel DG945GCLF2 - a clone board with just 1 DIMM slot and two SATA ports. Everything I need to have supported Just Worked out of the box. . I can provide the output of most any other commands if you'd like to see anything specific. I rather suspect that the Supermicro and other server-class Atoms will still be using the Intel 945 or similar chipsets. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
binary upgrade 6.1 - 7.2/8.0
Is it feasible to upgrade a system from 6.1 to 7.2 or 8.0-RC1 and if yes what sequence of upgrades should I actually carry out ie is it feasible to do 6.1-6.2 and then 6.2 - 7.2 or should it be done in small steps? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: fix remote degraded gmirror
mojo fms wrote: ... > > > What about shrinking the old mirror drive a few megs so its smaller than > the new one? > The original problem has gone away for the moment as my hoter found an AAJS drive with the same number of sectors and the mirror synchronized fine. I looked around for ways to shrink the slice, but didn't discover anything very authoritative or easy. Is there a slice reduction beast? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: fix remote degraded gmirror
Adam Vande More wrote: . >> > > This has been answered already but you should be aware that when using > gmirror, your back disk should either be the same model or bigger as hard > drive capacities differ between manufacturers. As in Seagate 250GB ~= WD > 250GB. So this is not an uncommon problem. > > yes this is a fairly old WD original was Device Model: WDC WD1600JS-22MHB0 Firmware Version: 02.01C03 User Capacity: 160,041,885,696 bytes they replaced with Device Model: WDC WD1600JS-75NCB2 Firmware Version: 10.02E03 User Capacity: 160,000,000,000 bytes which looks quite strange as it looks like something got lopped off somehow. I didn't think bad blocking/sectoring was done any more. Anyhow they've removed this disk and are considering options (at least I hope they are). -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: fix remote degraded gmirror
. a simple dd if=/dev/zero of=/dev/adX bs=1M count=1 would fix that if nothing else if there's was no existing label, there is nothing to do except gmirror man instructions. Gmirror is block level mirroring which why no other steps are needed. well now I really need advice, the data centre guy successfully inserted another disk in the right place and it appears to be the same model. However, smartctl -a /dev/ad4 seems to show a number of errors and many of the indicators are prefail or old_age. In addition the disk doesn't have exactly the same length so when I tried to insert gmirror complains that it is too small. What are my options? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: fix remote degraded gmirror
Thanks for the information re reboot; I think I knew it was a sata already. My one remaining question is what preparation does the hd need prior to gmirror insert. I see various people recommending clearing out various chunks of the disk (to make the disk unambiguously not in sync?), but that doesn't seem to be required. I guess the whole idea is to be able to just connect a new disk and start the system up and then insert back into the mirror. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: fix remote degraded gmirror
krad wrote: ... probably not but possibly. What is the server you have is it a custom build one? Does it have a drive cage? Can you make the hd light blink? I've never seen the machinery. Is the remote hands going to go inside the case for you, as I think he might to have to trace the cables to work out what drive is the correct one? I thought the serial number would help with that, Personally I would recommend that you or one of your trusted engineers goes down an replaces the drive yourself. Remote hands engineers in my experiance are a bit crude, and at the end of the day dont care about your box as much as you would. but you're probably right about the remote hands being uncaring :( While you or your guy is down there he could do a proper job and label up all the drives so in future you can get remote hands to do it all for you. it's still under discussion anyway whether we just abandon the machine; the current colo company took over our freebsd machines from another and they don't provide freebsd support anyway. The alternative ugh is to switch to ubuntu. I suspect if we do go ahead it will come down to one of our guys going down there to do the hands on. I'm still looking for good examples of exactly how to do the actual recovery. I assume I can just do gmirror remove gm0 /dev/ad4 to eliminate the bad disk. However, will the system still reboot from the disk on /dev/ad6 or should the good disk be swapped into the /dev/ad4 position. I don't know enough about sata booting, but I suppose it can be configured through our serial console. If the system does boot then the other disk can be inserted with gmirror insert gm0 /dev/ad6, but I am not sure whether any partitioning/formatting etc etc is required, Dru Lavigne was my guide when we installed this stuff, but this is the first time it's failed and she doesn't mention the repair process. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
fix remote degraded gmirror
I have a freebsd 6.1 system with a gmirror raid 1 which is degraded. Originally I had /dev/ad4 & /dev/ad6, but I'm getting a degraded status and log messages reporting dergraded status and uncorrectable block for ad4. Obviously I need to remove the device and get it replaced and then insert back into the mirror. However, this is a remote machine and I need to get the right disk fixed. Is there a way I can get the colocation engineer to replace the right disk? I guess I need to do 1) figure out what kind of disk this is and 2) figure out some information to allow them to identify. I looked at the smartctl output and see this /dev/ad4 Device Model: WDC WD1600JS-22MHB0 Serial Number:WD-WCANM4438410 /dev/ad6 Device Model: WDC WD1600JS-22MHB0 Serial Number:WD-WCANM4434657 is this sufficient to enable a data centre engineer to fix my problem? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Problem building htop
I'm trying to build htop-0.8.1 from ports on a FreeBSD 6.1 system. It seems htop requires ncurses and I see this error In file included from RichString.h:14, from Object.h:12, from Process.h:14, from ProcessList.h:17, from Settings.h:13, from AvailableMetersPanel.h:6, from AvailableMetersPanel.c:2: /usr/include/curses.h:289: error: conflicting types for 'wchar_t' /usr/include/stdlib.h:58: error: previous declaration of 'wchar_t' was here *** Error code 1 do I need a more modern system or should I update the ports tree and try again? I guess that stdlib comes with the OS and maybe ncurses is modernized. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
apache halting during log rotation
I have the following in my newsyslog.conf /var/log/httpd-*.log644 9 8@T01BG /var/run/httpd.pid 30 and am seeing crashes with apache-2.0.63_2 which I think are associated. The relevant error log entries appear to be these lines [Thu Apr 02 01:00:00 2009] [notice] Graceful restart requested, doing restart [Thu Apr 02 01:00:00 2009] [emerg] (2)No such file or directory: Couldn't initialize cross-process lock in child [Thu Apr 02 01:00:00 2009] [emerg] (2)No such file or directory: Couldn't initialize cross-process lock in child [Thu Apr 02 01:00:00 2009] [notice] FastCGI: process manager initialized (pid 93707) [Thu Apr 02 01:00:01 2009] [warn] RSA server certificate CommonName (CN) `*.xxx.org' does NOT match server name!? [Thu Apr 02 01:00:01 2009] [warn] RSA server certificate CommonName (CN) `*.xxx.org' does NOT match server name!? ... [Thu Apr 02 01:00:01 2009] [warn] RSA server certificate wildcard CommonName (CN) `*.xxx.org' does NOT match server name!? [Thu Apr 02 01:00:01 2009] [notice] Apache/2.0.63 (FreeBSD) mod_fastcgi/2.4.6 mod_ssl/2.0.63 OpenSSL/0.9.8e configured -- resuming normal operations [Thu Apr 02 01:00:01 2009] [alert] Child 93701 returned a Fatal error... Apache is exiting! [Thu Apr 02 01:00:01 2009] [alert] FastCGI: read() from pipe failed (0) [Thu Apr 02 01:00:01 2009] [alert] FastCGI: the PM is shutting down, Apache seems to have disappeared - bye I have used this rotation scheme with 6.0/6.1/6.2 for some years and it has never failed till the last few days when I'm now running 7.0-RELEASE. Can anyone adivise what could be done to ameliorate this. I see mention of rotatelogs, is that a better solution to this problem? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
smartd errors again
I'm seeing these sorts of messages in syslog output Mar 12 15:06:10 app3 smartd[519]: Device: /dev/ad4, 6 Currently unreadable (pending) sectors Mar 12 15:06:10 app3 smartd[519]: Device: /dev/ad4, 6 Offline uncorrectable sectors Mar 12 15:29:39 app3 kernel: ad4: FAILURE - READ_DMA status=51 error=40 LBA=151607951 Mar 12 15:29:39 app3 kernel: g_vfs_done():ad4s1f[READ(offset=72458502144, length=16384)]error = 5 Mar 12 15:30:25 app3 kernel: ad4: FAILURE - READ_DMA status=51 error=40 LBA=151607951 Mar 12 15:30:25 app3 kernel: g_vfs_done():ad4s1f[READ(offset=72458502144, length=16384)]error = 5 Mar 12 15:35:43 app3 smartd[519]: Device: /dev/ad4, 6 Currently unreadable (pending) sectors Mar 12 15:35:43 app3 smartd[519]: Device: /dev/ad4, 6 Offline uncorrectable sectors Mar 12 15:55:03 app3 kernel: ad4: FAILURE - READ_DMA status=51 error=40 LBA=151607951 Mar 12 15:55:03 app3 kernel: g_vfs_done():ad4s1f[READ(offset=72458502144, length=114688)]error = 5 Mar 12 16:05:28 app3 smartd[519]: Device: /dev/ad4, 6 Currently unreadable (pending) sectors Mar 12 16:05:28 app3 smartd[519]: Device: /dev/ad4, 6 Offline uncorrectable sectors by searching through the filesystem I located a file which appears to contain the error. However, I am unable to write to this file (the recommended cure for bad blocks) I used # dd if=/dev/zero bs=12544487 count=1 of=wow-22048.dmp.gz conv=notrunc dd: wow-22048.dmp.gz: Input/output error 1+0 records in 0+0 records out 0 bytes transferred in 1.894545 secs (0 bytes/sec) so am I forced to abandon this hard disk or is their some easy way to actually force a write that will make the disk bad block the wrong sectors. I looked in vain for my old bad blocking tools and indeed it seems lost+found is no longer here. This disk is supposedly only 433 hours old acording to the smartctl output (assuming I do need to regard the Hitachi numbers as being minutes). Any help would be appreciated. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
problem rotating apache logs
I'm trying to get newsyslog to rotate my apache2 logs (this in 7.0 & 7.1 release). I added this to the end of newsyslog.conf /var/log/httpd-*.log644 9 8@T01BG /var/run/httpd.pid 30 this is what is currently working for me in 6.0 & 6.1. Signal 30 is supposed to be USR1 which is intended to cause a graceful restart. However, I find that although I do get a rotated log file the new logs always seem to be empty. The logs seem to start growing only after I restart the apache server. What am I doing wrong? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: install freebsd from inside another operating system
After some messing around with various efforts on the depenguinator I came up with this scheme for getting a freebsd boot using an external freebsd machine ie not using the target to create the boot. ##freebsd machine #get the freebsd bootonly location wget [use your local location]/7.1-RC1-i386-bootonly.iso #determine its size (in my case 39M) ls -l 7.1-RC1-i386-bootonly.iso #make an image file slightly larger dd if=/dev/zero of=fbsd.img bs=1k count=4 #create /dev/md0 connected to the image file mdconfig -a -t vnode -f fbsd.img -u 0 #lable it and make a file system thereon bsdlabel -w -B md0 auto newfs -m 0 md0a #make a mount point and then mount the new mem fs on it mkdir img mount /dev/md0a img #and check its usage etc etc df img #now make a mount point for the iso and connect it to md1 mkdir iso mdconfig -a -t vnode -f 7.1-RC1-i386-bootonly.iso -u 1 #and mount as cd mount_cd9660 /dev/md1 iso df iso #copy the iso to the image cp -pr iso/* img/ #clean up umount img iso mdconfig -d -u 0 mdconfig -d -u 1 #copy the image to the target scp fbsd.img m...@target:/tmp/ ### target machine ubuntu 8.10 in my case Now after logging in to the target and sudoing su mv /tmp/fbsd.img /boot cp /usr/lib/syslinux/memdisk /boot then edited /boot/grub/menu.lst to contain a new entry title FreeBSD root (hd0,0) kernel memdisk initrd /boot/fbsd.img On reboot I selected the FreeBSD entry and was pleasantly surprised to see the freebsd boot process happening just fine. I suppose with this setup (provided all stays connected) one can do a full install from a network whose parameters may be set up during the install. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: install freebsd from inside another operating system
My experiments with the depenguinator seem to show it has a hidden dependency on the partition used for the constructed disk image. On my ubuntu 8.10 the original install constructed /dev/sda1 /dev/sda2 extended /dev/sda5 swap Not sure why the extended, but it wasn't really needed so I deleted sda5 & sda2 and then created a new primary partition /dev/sda2 for the swap. When using /dev/sda5 grub gave an error 12: invalid device, now with exactly the same build process /dev/sda2 gives a boot that works. However, the result crashes when the next stage kicks in. I don't know where the dependency is, perhaps it must use a primary partition or perhaps there's some way to specify the boot device that's not being used. The bootcode is fixed and not a function of the boot target (which is suspicious). -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: install freebsd from inside another operating system
Jan Henrik Sylvester wrote: Robin wrote: > I'm just wondering if it is possible to install FreeBSD from inside a > bootstrap operating system such as a knoppix or ubuntu preinstall? http://www.daemonology.net/blog/2008-01-29-depenguinator-2.0.html Cheers, Jan Henrik well I tried pretty hard with the depenguinator, but it doesn't work for me. I have an 8.10 ubuntu system with swap on sda5. I had to fix a buglet related to ARG_MAX being missing while compiling glob.c, but otherwise the build seemed to work. Have done all the suggested fixes ie making the partition bootable, changing the type to freebsd, but at boot time I get "12 : Invalid device requested.". I suspect there's some residual problem related to not having exactly the same setup as the author's. One thing I noticed is that the bootsector is not built with knowledge of the target partition, but I don't know exactly how I could fix that, but presumably it's supposed to boot as freebsd. I see it sets partition to 4, 5 sectors size. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: install freebsd from inside another operating system
Jan Henrik Sylvester wrote: Robin wrote: > I'm just wondering if it is possible to install FreeBSD from inside a > bootstrap operating system such as a knoppix or ubuntu preinstall? http://www.daemonology.net/blog/2008-01-29-depenguinator-2.0.html ... I've just finished building an image using it. Bit of a struggle as I had to patch one of the files and my lack of understanding of the chattr command in ubuntu caused me no end of problems. Basically when you run the build again after the first failure you need to clean up properly using chattr -R -i before you can even remove one of the folders. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: install freebsd from inside another operating system
Wojciech Puchar wrote: Reason I'm asking is that my boss is asking about getting another commodity server using a provider that doesn't offer freebsd. We do get KVM over IP however, and I assume that really clever people might know how to handle this sort of thing if they have console access. stupid/strange solution: run qemu under linux giving FReeBSD iso as cdrom, and linux swap partition as disk (disable swap on linux temporarily), install it WITHOUT making slices and labels (manually) - minimum install, make sure whole "disk" (actually linux swap partition) is bootable, then exit qemu, run linux fdisk to set swap partition as bootable instead of linux root, reboot. you will get FreeBSD working, repartition everything else, optionally copy FreeBSD to partition that has to be /, reboot, ready That almost sounds doable :) not sure my ubuntu/qemu skills are really there yet. Can you run qemu in a tty/console? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
install freebsd from inside another operating system
I'm just wondering if it is possible to install FreeBSD from inside a bootstrap operating system such as a knoppix or ubuntu preinstall? Reason I'm asking is that my boss is asking about getting another commodity server using a provider that doesn't offer freebsd. We do get KVM over IP however, and I assume that really clever people might know how to handle this sort of thing if they have console access. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: gcc-4.3.3 problem
Jeremy Chadwick wrote: . >>> >>> Is there some way to make these libraries available without this? >> I think this is just a question of adding /usr/local/lib/gcc-4.3.3 to the ldconfig hints. At least when I ran >> >> ldconfig -R /usr/local/lib/gcc-4.3.3 >> >> things improved. > > The ports framework has support to do this automatically, so the > question is why it didn't happen. Could you file a PR on the matter? > yes. I looked in the makefile and it certainly has USE_LDCONFIG= ${TARGLIB} so I suppose the ldconfig stuff is supposed to get done. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: gcc-4.3.3 problem
Robin Becker wrote: I'm trying to run some benchmarks for a modified version of Python 2.5.2, unfortunately the modification requires a newer version of GCC so I have installed gcc version 4.3.3 from ports. When I try to run my make command I'm seeing this error *** WARNING: renaming "_codecs_hk" since importing it failed: Shared object "libgcc_s.so.1" I see that the gcc 4.3.3 libs were installed in /usr/local/lib/gcc-4.3.3 and I can get things to work by setting up the appropriate -L option and LD_LIBRARY_PATH. Is there some way to make these libraries available without this? I think this is just a question of adding /usr/local/lib/gcc-4.3.3 to the ldconfig hints. At least when I ran ldconfig -R /usr/local/lib/gcc-4.3.3 things improved. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
gcc-4.3.3 problem
I'm trying to run some benchmarks for a modified version of Python 2.5.2, unfortunately the modification requires a newer version of GCC so I have installed gcc version 4.3.3 from ports. When I try to run my make command I'm seeing this error *** WARNING: renaming "_codecs_hk" since importing it failed: Shared object "libgcc_s.so.1" I see that the gcc 4.3.3 libs were installed in /usr/local/lib/gcc-4.3.3 and I can get things to work by setting up the appropriate -L option and LD_LIBRARY_PATH. Is there some way to make these libraries available without this? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
gcc 3.4.4 -fno-gcse
I'm trying to do some benchmarks for a new/improved version of CPython and would like to know if gcc 3.4.4 as distributed with FreeBSD 6.1 handles the -fno-gcse option reasonably. I looked in the man page, but don't see that option explicitly so perhaps the main thrust of the optimisation approach is going wrong. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: no reverse dns
Chuck Swiger wrote: On Oct 29, 2008, at 11:10 AM, Robin Becker wrote: We have just moved offices and our freebsd machine has started complaining in the following terms Oct 29 17:14:39 int kernel: arplookup ww.xx.yy.zz failed: host is not on local network We have an external router connected as a dhcp server at 192.168.0.2 which apparently has external address ww.xx.yy.zz. I am using a fixed ip address ie 192.168.0.6 I have this in my rc.conf defaultrouter="192.168.0.2" hostname="int.myoffice.com" ifconfig_em0="inet 192.168.0.6 netmask 255.255.255.0" and have dns mapping int.myoffice.com --> ww.xx.yy.zz, If you tell the machine that it is int.myoffice.com and you set up DNS which claims that hostname has an external IP, it will be sad because it doesn't know how to reach that IP. You can use DNS split-horizon / views to return the internal IP when the machine asks, or simply keep your external and internal names separate. Ie, set up DNS like: int.myoffice.com A 192.168.0.6 ext.myoffice.com A ww.xx.yy.zz Regards, On the machine I have set the local names to point to 192.168.0.6 in the hosts file. I have not set up any dns except externally. I suppose that packets are arriving and being routed via NAT into the internal server which claim to be addressed to the router's external address. Can I add some simple route that fixes this? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
no reverse dns
We have just moved offices and our freebsd machine has started complaining in the following terms Oct 29 17:14:39 int kernel: arplookup ww.xx.yy.zz failed: host is not on local network We have an external router connected as a dhcp server at 192.168.0.2 which apparently has external address ww.xx.yy.zz. I am using a fixed ip address ie 192.168.0.6 I have this in my rc.conf defaultrouter="192.168.0.2" hostname="int.myoffice.com" ifconfig_em0="inet 192.168.0.6 netmask 255.255.255.0" and have dns mapping int.myoffice.com --> ww.xx.yy.zz, but our ISP will not make the reverse mapping. I assume that we're trying to reverse lookup something and the lack of reverse dns is causing this issue. What can I add to my rc.conf to stop this arplookup problem? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
change of behaviour after reboot
I had to perform a stupidity related reboot after allowing my /var fs to get to 109%. All seems well except for the behaviour of some of my rather ancient cgi scripts which are for serving up moinmoin wikis. Prior to the reboot these all seemed to work fine they hd the #! line #!/usr/bin/env python after the reboot these scripts refused to work. The log messages say env: python: No such file or directory Premature end of script headers: zzz-wiki.cgi I tried to check that env was there and that it gave me a reasonable python (at least at the terminal). However, something had obviously changed as the scripts now only work if I use #!/usr/local/bin/python I'm worried now in case my reboot or the over usage has caused some hidden damage. Anybody have any idea what could cause this misbehaviour? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: network problem
OutBackDingo wrote: whats redirecting the requests to the apache server? is it on piublic or private ip block ? whats it set to listen on IP wise ? what about a changed netmask ? The server is listening to a local ip address which hasn't changed eg 192.168.0.x The external IP address was changed as was the external name server and the netmask. Our local net is provided by an ethernet router that has been configured with the new external IP address in exactly the same way as the old one. I've just reviewed the settings again and both https(443) and http(80) appear to be NAT'd to 192.168.0.x which is both the http & https server. Https works & http doesn't. The external IP is xxx.yyy.zzz.240 with netmask xxx.yyy.zzz.224 and gateway xxx.yyy.zzz.225 and I think that works. The dns server ip is setup in the gateway and it acts as both a dhcp server and static router. When I connect from the lan side to http://xxx.yyy.zzz.240 it seems to work fine. From outside it fails. On Saturday 16 August 2008 15:55:11 Robin Becker wrote: After a recent change of ISP and hence our external IP address I find that our FreeBSD 6.1 apache http server appears to be invisible from outside our site. Port 80 requests just seem to hang whether done via a dns lookup or using the ip address. Checking the logs seems to indicate that no requests now arrive, but that we have seen some external requests since the ISP changed. The https/sshd/mysql etc servers are still visible externally and the http server is visible on its local address and if you visit the external address from inside the network. I suspect that either I've still got some misconfiguration on the server, gateway or or dns(although this seems less likely). Alternatively the ISP might have some other block in place. I've looked for the old IP address in /etc and found no usages the gateway appears to be correctly set up. What tests can I do internally/externally to see what happens to my port 80 packets? ... -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
network problem
After a recent change of ISP and hence our external IP address I find that our FreeBSD 6.1 apache http server appears to be invisible from outside our site. Port 80 requests just seem to hang whether done via a dns lookup or using the ip address. Checking the logs seems to indicate that no requests now arrive, but that we have seen some external requests since the ISP changed. The https/sshd/mysql etc servers are still visible externally and the http server is visible on its local address and if you visit the external address from inside the network. I suspect that either I've still got some misconfiguration on the server, gateway or or dns(although this seems less likely). Alternatively the ISP might have some other block in place. I've looked for the old IP address in /etc and found no usages the gateway appears to be correctly set up. What tests can I do internally/externally to see what happens to my port 80 packets? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: shutdown forensics
Robert Huff wrote: Robin Becker writes: I'm trying to figure out what caused an unplanned shutdown of a freebsd 6.0 server; I see this in the last output reboot ~ Thu Apr 24 06:15 shutdown ~ Thu Apr 24 06:10 but cannot see anybody having been logged in at the time. Is there any way I can trace the origin of the shutdown command? I see a time like that, my gut suspicion for the shutdown is "something drive y a cro job". Robert Huff ... I've looked at the various cron jobs and cannot see anything that looks suspicious. Root only has an ntp job which runs at 01:17. The one other user with a crontab has jobs running at *:19 and 03:13. Looked in /var/crash and see minfree from 2005. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
shutdown forensics
I'm trying to figure out what caused an unplanned shutdown of a freebsd 6.0 server; I see this in the last output reboot ~ Thu Apr 24 06:15 shutdown ~ Thu Apr 24 06:10 but cannot see anybody having been logged in at the time. Is there any way I can trace the origin of the shutdown command? I looked in the root shell history various other logs, but couldn't see anything obvious. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
ldconfig vs /usr/local/lib/mysql
One of our FreeBSD 6 servers got rebooted and when it came back we found that /usr/local/lib/mysql was not amongst the directories on the standard ldconfig paths. Can someone gently explain how this is supposed to work. On our other servers I see the both /usr/local/lib/apache2 and /usr/local/lib/mysql are on the standard paths, but I don't know where those folders get set up as there are no ld.so.conf (or ld-elf.so.conf) files in /etc. It's easy to get those folders into the /var/run/ld.so.hints etc files, but how does one make them persist? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
how to respond to possible attacks
Sorry if this is too off topic, but I would like to find out what to do when you suspect a possible dos attack on your system. I know there are many experienced sysadmins here. Although my system (freebsd 6.0/apache 2.0.x) did in fact hold up, what steps should I be taking? The originating ip doesn't seem to be reverse mappable. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
apache/fastcgi/django kern.maxfiles
I'm having a problem related to a freebsd 6.x apache 2.0.x+mod_fastcgi --> django+flup forked server. Basically under heavy load the system seems to reach some kind of limit on the number of open files ie I see messages related to user 80 & (the django server uid) having reached kern.maxfiles. Has anyone got any experience with this kind of setup and how to decide what the fault really is. In particular since the web can always out request any particular limit is there some way to make apache start refusing gracefully before things go critical. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: 7.0 kde 3 build patch error
Matthias Apitz wrote: El día Wednesday, March 05, 2008 a las 04:02:11PM +, Robin Becker escribió: I'm trying to build kde3 from ports with 7.0 release and get this patch error whilst trying to build cups ... After doing portsnap update I get further with building kde3, but am now getting stuck with some error related to docbook .. Making all in doc rm -rf html mkdir html jade -t sgml -d ./website.dsl\#html ./manual.sgml jade:/usr/local/share/sgml/docbook/4.2/dbcentx.mod:308:0:E: cannot open "/usr/local/share/sgml/docbook/4.2/iso-amsa.gml" (No such file or directory) jade:/usr/local/share/sgml/docbook/4.2/dbcentx.mod:312:0:E: cannot open "/usr/local/share/sgml/docbook/4.2/iso-amsb.gml" (No such file or directory) jade:/usr/local/share/sgml/docbook/4.2/dbcentx.mod:316:0:E: cannot open "/usr/local/share/sgml/docbook/4.2/iso-amsc.gml" (No such file or directory) ... I guess this is some bug building docbook, but is there a better way to get kde up and running? Or should I try again with make config-recursive and remove the docbook related stuff. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: 7.0 kde 3 build patch error
Matthias Apitz wrote: El día Wednesday, March 05, 2008 a las 04:02:11PM +, Robin Becker escribió: I'm trying to build kde3 from ports with 7.0 release and get this patch error whilst trying to build cups ===> Running ldconfig *** Error code 1 Stop in /usr/ports/print/cups-base. *** Error code 1 ... I run into the same problem and fetched the tar file of the cups-base-1.3.6 port from FreeBSD which compiles fine; .. thanks updating the ports tree seems to have fixed that. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
7.0 kde 3 build patch error
I'm trying to build kde3 from ports with 7.0 release and get this patch error whilst trying to build cups ===> Running ldconfig /sbin/ldconfig -m /usr/local/lib ===> Registering installation for libmng-1.0.9 ===> Returning to build of qt-3.3.8_6 ===> qt-3.3.8_6 depends on shared library: png - found ===> qt-3.3.8_6 depends on shared library: jpeg - found ===> qt-3.3.8_6 depends on shared library: Xft.2 - found ===> qt-3.3.8_6 depends on shared library: cups.2 - not found ===>Verifying install for cups.2 in /usr/ports/print/cups-base ===> Vulnerability check disabled, database not found ===> Found saved configuration for cups-base-1.3.5_2 => cups-1.3.5-source.tar.bz2 doesn't seem to exist in /usr/ports/distfiles/. => Attempting to fetch from http://ftp.easysw.com/pub/cups/1.3.5/. cups-1.3.5-source.tar.bz2 100% of 3986 kB 190 kBps 00m00s ===> Extracting for cups-base-1.3.5_2 => MD5 Checksum OK for cups-1.3.5-source.tar.bz2. => SHA256 Checksum OK for cups-1.3.5-source.tar.bz2. ===> Patching for cups-base-1.3.5_2 ===> Applying FreeBSD patches for cups-base-1.3.5_2 Ignoring previously applied (or reversed) patch. 10 out of 10 hunks ignored--saving rejects to cups/ipp.c.rej => Patch patch-CVE-2007-4351 failed to apply cleanly. *** Error code 1 Stop in /usr/ports/print/cups-base. *** Error code 1 Stop in /usr/ports/x11-toolkits/qt33. *** Error code 1 Stop in /usr/ports/x11-toolkits/qt33. *** Error code 1 Stop in /usr/ports/x11/kde3. /usr/ports/x11/kde3: -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
duplicate message removal
We have a bunch of FreeBSD 6.x servers which we administer remotely. As part of that we get the normal root job mails emailed to a mailing list which the admins(mostly me) can inspect at leisure and also use for historical purposes. Trouble is many of the emails get huge because of repeated messages typically stuff like xxx.yyy.com login failures: Feb 22 20:07:54 app3 sshd[56886]: reverse mapping checking getaddrinfo for 216-194-26-66.ny.ny.metconnect.net failed - POSSIBLE BREAKIN ATTEMPT! etc etc All these servers are running denyhosts, but we still see lots of these messages. I was wondering if there's any simple compression script which notices the repetitions (apart from timestamp) and can remove the many duplicates etc etc. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: FBSD or PCBSD?
Zbigniew Szalbot wrote: Which is a better solution for a home user? Thank you very much for your practical recommendations! ... I have had great success using PCBSD on various acers, I eventually got fed up with it being not quite identical to FreeBSD and I now have FreeBSD 6.2 on my current acer laptop (a 1680). -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
software raid 1 and recovery
I'm using software raid 1 on a 6.1 freebsd. This is a so called cold swap system, but I wonder how much it actually improves reliability. First off what should I be doing to detect error conditions and secondly what happens if the machine refuses to boot. I set this system up using Dru Lavigne's recipe, but I don't really understand what happens if one of the drives starts to fail. I think there was some discussion about HD monitoring recently, but I can't seem to locate it. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: genuine bulk email
Jeffrey Goldberg wrote: On Oct 12, 2007, at 9:40 AM, Robin Becker wrote: these all sound very reasonable. However, we use the same IP for several virtual hosts ie we have more than one domain name so the reverse DNS is not clear to me. Is the from address inspected for comparison with the RDNS ie if I claim to be sending from xxx.com should my RDNS point back to xxx.com? Presumably I can have only one IP-->domain ptr. I suspect it will be easier to set up the front end machine as that is supposed to be for the same client. The checking will work fine with virtual domains. What matters is that DNS(rDNS(IP)) = IP = DNS(vhost) I think I've got that right. (It's a bit more complicated to state when MXes and multiple A records for the same name are considered, but this is the general idea.) Also it's not so much the header FROM or the envelope FROM, but the HELO string that is checked here. For SPF and DomainKeys, it is the envelope FROM that is checked against the IP. Presumably your mailer gives a constant HELO irrespective of the vhost that is in the envelope FROM. OK this makes better sense. So long as my internal machine name matches the equation above then I should be OK no matter what. I'm fairly sure I've got stuff set up properly. I do need to set up the appropriate rDNS and check the HELO on the test machine though. The HELO looks right, but my rDNS is shot and I am currently in the CBL for some reason. And how do the sales people acquire the data? I'm sure that it's okay, but you may want to have a small description of the process on your web page that you could point postmasters to if a question arises. .. these are telephone sales people so far as I know. When they're booking the room(s) for the client they ask if they want to receive an emailed document describing the hotel etc etc so it's not bulk in the sense of database --> email and certainly we're not recording any of the details at all. I suppose we could be used by our client as a spurious mail sender, but it would not be terribly fast and it's unlikely as they likely could do it much easier themselves. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: genuine bulk email
Jeffrey Goldberg wrote: On Oct 12, 2007, at 7:05 AM, Robin Becker wrote: At present I have reduced the email to a textual format with an embedded textual link. So the email looks like .. When you personalize that give the date and IP address of the request. Something like ... the brochures you requested at TIME from IP. this sounds reasonable but is there anything at a system level that can be done to make emails less likely to be classified as spam? The most crucial thing is the status of IP of the host sending the mail o Does it have a proper DNS PTR (reverse DNS) record? o Are you using SPF or DomainKeys to show that that IP address is authorized to send mail in the sending domain's name? o Do you have working postmaster and abuse addresses for the domain you are sending from? o Do you have a static IP address? o Are you clear of any major blacklists? o Can you demonstrate that every recipient really did request the mail? these all sound very reasonable. However, we use the same IP for several virtual hosts ie we have more than one domain name so the reverse DNS is not clear to me. Is the from address inspected for comparison with the RDNS ie if I claim to be sending from xxx.com should my RDNS point back to xxx.com? Presumably I can have only one IP-->domain ptr. I suspect it will be easier to set up the front end machine as that is supposed to be for the same client. Each of those are far more important than whether you attach a PDF. (By the way, say it's PDF or even Adobe's PDF, but not "Adobe Acrobat format".) OK that's good. I assume that spammers try very hard and fail, so is this kind of email application effectively dead in the water before it starts? Automatic mailing is fine. What is important is how the email addresses were acquired. .. this isn't automatic, the sales people manually enter all the details. The attached document is what the application generates and then the combined email. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
genuine bulk email
We have a project to make a graphical tool which allows salespersons to send an email with an attached PDF document. We are in testing and although the tool seems to work we have the problem of not looking like spam/phish etc etc. I have tried numerous tricks to avoid being classified as spam, but so far nothing seems to work. At present I have reduced the email to a textual format with an embedded textual link. So the email looks like Your Document, Thank you for your inquiry. below is a link to the brochures as requested, in Adobe Acrobat format. It includes the YYY Airport Hotel and other information which may be useful. We thank you for your query. Your Document may be found here http://host/path/aa-hrcc-20071012113659-20zi0rfoknv6gdi1w4bls0psd0.pdf Sales Team It could be personalised a bit more, but is there anything at a system level that can be done to make emails less likely to be classified as spam? I assume that spammers try very hard and fail, so is this kind of email application effectively dead in the water before it starts? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: temporary su login
Eric Crist wrote: I'm sure nobody will mention this, so I will. On most systems with support ACPI, your colo provider can simply press the power button on the front of your server. FreeBSD's kernel will pick up the signal and shut down cleanly. Once you're moved, they can press the same button to power the system on. There is *NO* need to give them login access to the box. Also, they could simply call you to have you shut it down. .. many good ideas; thanks. I guess since they ask for an ip based mechanism that I'll create a special user in the operator group and do the chmod trick on shutdown. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
temporary su login
My collocation supplier is about to move our FreeBSD box and wants some way to shut it down cleanly. Is there a simple way to allow a non-root user to have shutdown rights without just giving them the world. At present I don't even allow login via ssh on that box ie it's purely key based. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
hda driver
I have an acer 1644WLMi laptop which has an HDA sound card. I tried using the binary modules from http://tomoyo.mybsd.org.my/HDA/kmod/i386/ If I add snd_hda_load="yes" sound_load="yes" to loader.conf then I have a crash when starting kdm I am running 6.2Release with Xorg 7.2 and latest port kde 3.5. I have seen that others have this HDA sound working; so what's the proper way to get it working? My reading of the related threads in freebsd-multimedia left me a bit confused. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Yes i need to Ask A question
Christian Walther wrote: ... On a side note: Having a FreeBSD based Laptop with you in a store where people know only about Windows is an experience one should have. They need their time to get used to the idea that there are other Operating Systems as Windows. And that these don't need graphics to work. And that there are people out there that can work with a computer without graphics... ;-D ... here in the UK you'd be lucky to be allowed to try even one card in your own machine. Only reason I bought an Acer is that someone else had already managed to get most of the kinks ironed out (and had written about it). Unfortunately improvements in FreeBSD have taken me backwards in terms of understanding over the last year. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
iwi won't switch on at boot
I have an acer laptop with an intel 2200 wifi; when I had PCBSD 1.3 installed I could boot from cold and have the wireless start up properly (it's using wpa with a hidden ssid ap). After I installed freebsd 6.2 I cannot get the wireless to come up on first boot. I seem to have to switch the wireless on and then reboot. I think under PCBSD is used to have to do something like sysctl dev.iwi.0.radio=1 in one of the startup scripts to make this work. Now however when I execute that I see a message saying that oid is read only. Is there some way to get this wifi starting properly again? -- Robin Becker various bits of information $ uname -a FreeBSD xxx.net 6.2-RELEASE FreeBSD 6.2-RELEASE #0: Fri Jan 12 11:05:30 UTC 2007 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/SMP i386 $ kldstat Id Refs AddressSize Name 1 15 0xc040 70794c kernel 21 0xc0b08000 1adb8linux.ko 31 0xc0b23000 e1a8 if_iwi.ko 42 0xc0b32000 2d60 firmware.ko 51 0xc0b35000 2cf4 wlan_acl.ko 61 0xc0b38000 59f20acpi.ko 71 0xc4f27000 4000 logo_saver.ko $ pkg_info | grep iwi iwi-firmware-2.4_7 Intel PRO/Wireless 2200 Driver Firmware iwi-firmware-kmod-3.0_2 Intel PRO/Wireless 2200 Firmware Kernel Module $ cat /boot/loader.conf if_iwi_load="YES" firmware_load="YES" wlan_load="YES" wlan_acl_load="YES" wlan_wep_load="YES" wlan_ccmp_load="YES" wlan_tkip_load="YES" messages == initial boot Jun 10 08:53:01 xxx kernel: iwi0: mem 0xb010 1000-0xb0101fff irq 17 at device 4.0 on pci6 Jun 10 08:53:01 xxx kernel: iwi0: Ethernet address: 00:13:ce:a1:2e:8e Jun 10 08:53:01 xxx kernel: iwi0: radio turned off == after switch pressed Jun 10 08:55:17 xxx kernel: iwi0: radio turned on == after reboot Jun 10 09:14:44 xxx kernel: iwi0: mem 0xb010 1000-0xb0101fff irq 17 at device 4.0 on pci6 Jun 10 09:14:44 xxx kernel: iwi0: Ethernet address: 00:13:ce:a1:2e:8e ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: failure in xorg 7.2 compile of vblank.c
On 09/06/07, Robin Becker <[EMAIL PROTECTED]> wrote: I am trying the build from scratch route of xorg 7.2. > cc -c -I. -I../../../../../src/mesa/drivers/dri/common -Iserver -I../../../../../include -I../../../../../include/GL/internal -I../../../../../src/mesa -I../../../../../src/mesa/main -I../../../../../src/mesa/glapi -I../../../../../src/mesa/math -I../../../../../src/mesa/transform -I../../../../../src/mesa/shader -I../../../../../src/mesa/swrast -I../../../../../src/mesa/swrast_setup -I../../../../../src/egl/main -I../../../../../src/egl/drivers/dri `pkg-config --cflags libdrm` -I/usr/local/include -Wall -O -Wmissing-prototypes -std=c99 -Wundef -ffast-math -DUSE_X86_ASM -DUSE_MMX_ASM -DUSE_3DNOW_ASM -DUSE_SSE_ASM -I/usr/local/include -DPTHREADS -DUSE_EXTERNAL_DXTN_LIB=1 -DIN_DRI_DRIVER -DGLX_DIRECT_RENDERING -DGLX_INDIRECT_RENDERING -DHAVE_ALIAS ../common/vblank.c -o ../common/vblank.o > ../common/vblank.c: In function `driGetCurrentVBlank': > ../common/vblank.c:306: error: `DRM_VBLANK_SECONDARY' undeclared (first use in this function) > ../common/vblank.c:306: error: (Each undeclared identifier is reported only once > ../common/vblank.c:306: error: for each function it appears in.) > ../common/vblank.c: In function `driWaitForVBlank': > ../common/vblank.c:363: error: `DRM_VBLANK_SECONDARY' undeclared (first use in this function) ignore this I figured out it was caused by an unistalled libdrm package. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
detect stderr writes when combined
This is possibly a stupid question, but I would like to have a particular sh script stdout and stderr combined, but at least detect when stderr has been used. In particular for my cron scripts it seems that error messages get wrapped up and emailed, but they are then out of context with the normal informative output. If I could have the combined output go to my logs in the normal way and detect that some error output had occurred I could then email the whole cron output. I'm fairly sure this is doable with some C, but is there an easier way? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Synaptics trackpad problem
Paul Fraser wrote: On 6/3/07, Robin Becker <[EMAIL PROTECTED]> wrote: Any help appreciated Hi Robin, I know this isn't going to be much of a comfort - but I've had great difficulty trying to get Xorg to work nicely with - surprise - the Synaptics touchpad on my Acer notebook as well. I posted to the list almost a week ago reporting that while Xorg worked brilliantly one day - suddenly it stopped working. I did blow the installation away and started again (unrelated issue) the other day and the touchpad is working much as it was before. I don't have my specific Xorg config in front of me as the missus has stolen the laptop today unfortunately, however I've flagged this thread so I'll attach it and send it over to you tonight. Thanks I appreciate it. With my machine I can get the 1280x800 mode using 915resolution hack and I do have a mouse if I drop the touchpad entirely (from loader.conf and xorg.conf) and use moused_enable="YES" in rc.conf. For whatever reason the synaptics stuff is a bit flaky on my machine as I have had a couple of forced reboots with hw.psm.synaptics_support=1. If all else fails I will try and do the new 7.2 Xorg build from scratch. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Is it better to just rebuild Xorg 7.2
Norberto Meijome wrote: On Mon, 04 Jun 2007 19:45:48 + Robin Becker <[EMAIL PROTECTED]> wrote: I've just installed 6.2R and have just about got Xorg going. Is it better for me to just remove all my X related packages and try and build from the xorg-7.2 meta package or to go through the pain of trying to update? I think it'll be the same. Upgrade was the only option for me as I had a fully functionally X setup in my laptop - it went pretty smoothly (for the amount of changes done to the system). Make sure you follow the instructions in /usr/ports/UPDATING to the letter. In either case I need to sync the ports tree etc. well, if u dont upgrade your tree, you may be able to install xorg 6.9 from packages (MUCH faster) ; get it working; then sync the ports after this and upgrade. There is nothing particularly wrong with 6.9, specially if you have a clean machine and you plan to do the whole upgrade path. just dont go overkill with setting up KDE / Gnome and other things that depend on X. twm would do for now ;) ... I have Xorg from 6.2R working now, but I have already cvsup'd the ports so I guess I'll have to do the upgrade from the 6.2R version. I haven't yet got KDE etc going yet anyway. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
complete static link
I'm trying to get a particular utility compiled statically so I can embed it in a distribution package. It's just a part of lib tiff. I tried various things with ./configure --disable-shared --enable-static, but couldn't eliminate the dynamic dependencies. In the end I tried this gcc -O2 -Wall -W -o tiffcp tiffcp.o ../libtiff/.libs/libtiff.a ../port/.libs/libport.a /usr/lib/libz.a /usr/lib/libm.a /usr/lib/libc.a and although the libz/libm dependencies are gone I still depend on libc.so.6 (or 4). Is there some esy way to do this? Also I get warnings about f_prealloc anyone know what that means? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Is it better to just rebuild Xorg 7.2
I've just installed 6.2R and have just about got Xorg going. Is it better for me to just remove all my X related packages and try and build from the xorg-7.2 meta package or to go through the pain of trying to update? In either case I need to sync the ports tree etc. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Synaptics trackpad problem
Paul Fraser wrote: On 6/3/07, Robin Becker <[EMAIL PROTECTED]> wrote: Any help appreciated Hi Robin, I know this isn't going to be much of a comfort - but I've had great difficulty trying to get Xorg to work nicely with - surprise - the Synaptics touchpad on my Acer notebook as well. I posted to the list almost a week ago reporting that while Xorg worked brilliantly one day - suddenly it stopped working. I did blow the installation away and started again (unrelated issue) the other day and the touchpad is working much as it was before. I don't have my specific Xorg config in front of me as the missus has stolen the laptop today unfortunately, however I've flagged this thread so I'll attach it and send it over to you tonight. Thanks I appreciate it. With my machine I can get the 1280x800 mode using 915resolution hack and I do have a mouse if I drop the touchpad entirely (from loader.conf and xorg.conf) and use moused_enable="YES" in rc.conf. For whatever reason the synaptics stuff is a bit flaky on my machine as I have had a couple of forced reboots with hw.psm.synaptics_support=1. If all else fails I will try and do the new 7.2 Xorg build from scratch. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Synaptics trackpad problem
I'm trying to get XORG setup on my acer laptop with 6.2 release. Things are fine except I cannot get the trackpad to work; x starts, but the mouse is frozen. I have had pcbsd on the same machine and the mouse worked there, but not as a trackpad. Any help appreciated Xorg.0.log (**) Option "CorePointer" (**) Synaptics_Touchpad: Core Pointer (II) XINPUT: Adding extended input device "Synaptics_Touchpad" (type: MOUSE) (II) XINPUT: Adding extended input device "Keyboard0" (type: KEYBOARD) Synaptics DeviceInit called SynapticsCtrl called. Synaptics DeviceOn called (--) Synaptics_Touchpad synaptics touchpad found (--) Synaptics Touchpad, model: 0 (--) Firmware: 6.2 (--) Sensor: 18 (--) new absolute packet format (--) Touchpad has extended capability bits (--) -> multifinger detection (--) -> palm detection (WW) fcntl(8, O_ASYNC): Inappropriate ioctl for device Synaptics DeviceOff called the corrsponding sections in the xorg.conf are Section "ServerLayout" .. InputDevice"Synaptics_Touchpad""CorePointer" EndSection Section "Module" Load "synaptics" . EndSection Section "InputDevice" Identifier"Synaptics_Touchpad" Driver"synaptics" Option"Device""/dev/psm0" Option"Protocol""psm" Option"LeftEdge""1700" Option"RightEdge""5300" Option"TopEdge""1700" Option"BottomEdge""4200" Option"FingerLow""25" Option"FingerHigh""30" Option"MaxTapTime""180" Option"MaxTapMove""220" Option"VertScrollDelta""100" Option"HorizScrollDelta""100" Option"MinSpeed""0.06" Option"MaxSpeed""0.06" Option"AccelFactor""0.0010" Option"ScrollButtonRepeat""100" Option"UpDownScrolling""on" Option"UpDownRepeat""on" Option"LeftRightScrolling""on" Option"LeftRightRepeat""on" # "SHMConfig on" seems good works with synclient(1). But this # options is insecure. I recommended "off" as default. Option"SHMConfig""off" # If you use circular touchpad, uncomment them. #Option"CircularPad""on" #Option"CircularScrolling""on" #Option"CircScrollDelta""0.5" EndSection and certainly dmesg | grep Synaptics psm0: model Synaptics Touchpad, device ID 0 as /boot/loader.conf has hw.psm.synaptics_support="1" -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: ezjail ip conflicts
Joe Holden wrote: >> >> how do I fix this or perhaps I don't need to? > syslogd_flags="-ss" in rc.conf > sshd is configured in /etc/ssh/sshd_config. . I looked in vain in /etc/rc.d/syslogd for references to syslogd_ and didn't find any, but now I see \$rc_flags which I guess must be what is used. Thanks Joe and Karol. I now get a message saying Warning: IP 209.67.217.27 not configured on a local interface. but I think that just means I don't have an alias set up yet. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
ezjail ip conflicts
I'm getting these ip conflicts whilst trying to create a jail ezjail-admin create xxx.xxx.xxx.27 Warning: IP xxx.xxx.xxx.27 not configured on a local interface. Warning: Some services already seem to be listening on all IP, (including xxx.xxx.xxx.27) This may cause some confusion, here they are: mysqlmysqld 505 10 tcp4 *:3306*:* root syslogd291 6 udp4 *:514 *:* my rc.conf has ifconfig_fxp0="inet xxx.xxx.xxx.26 netmask 255.255.255.248" defaultrouter="xxx.xxx.xxx.25" inetd_flags="-wW -a xxx.xxx.xxx.26" so I believe the xxx.xxx.xxx.27 address is OK, but I guess I need to make mysqld and syslogd listen only on xxx.xxx.xxx.26. I don't actually understand what's preventing sshd from listening on all the addresses in range unless it's the inetd flags, but I thought sshd is started by init nowadays. Anyhow I think I can fix the mysqld problem by having mysql_args="--bind-address=xxx.xxx.xxx.26" in the rc.conf, but I don't see any easy way to configure syslogd to start with a -b xxx.xxx.xxx.26 how do I fix this or perhaps I don't need to? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: cron mystery
Chuck Swiger wrote: Robin Becker wrote: [ ... ] before ## SHELL=/bin/sh MAILTO=user 13 3 * * * $HOME/bin/daily 19 * * * * $HOME/bin/hourly after ## SHELL=/bin/sh MAILTO=user 13 3 * * * /home/user/bin/daily 41 * * * * /home/user/bin/hourly and at 41 past the hour the hourly job came back. Is it the HOME variable or the act of rewriting? User did have home defined in /etc/passwd. I suspect that $HOME isn't being defined as one might expect-- cron provides a very minimal shell environment for scripts it runs. except that I have exactly the same script running on another box with the same freeBSD version and that runs things fine. Looking in man 5 crontab seems to suggest that SHELL=/bin/sh & HOME, LOGNAME are set from the user passwd entry. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
cron mystery
Can anyone think of something that can stop cron working for a particular user? I just noticed on one of our 6.1 machines the crontab for a particular user wasn't run properly since dec 21. There were hourly and daily jobs, but neither seemed to be running. Looked in var/cron and see no deny or allow files. The user x had an proper crontab. In the end I modified the users crontab and rewrote it before ## SHELL=/bin/sh MAILTO=user 13 3 * * * $HOME/bin/daily 19 * * * * $HOME/bin/hourly after ## SHELL=/bin/sh MAILTO=user 13 3 * * * /home/user/bin/daily 41 * * * * /home/user/bin/hourly and at 41 past the hour the hourly job came back. Is it the HOME variable or the act of rewriting? User did have home defined in /etc/passwd. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: binary updating 6.0-6.2
Jerry McAllister wrote: Hello Robin, I see from the latest announcement that freebsd-update is now in the base system for 6.2. Following the links I see there's a procedure for 6.0-61. and another for 6.1-6.2 binary updating. By the way, I see that your post says binary update, but I really don't have confidence in that and prefer updating the source and doing the build as I described in my response. jerry no problem, I just replied to that and I almost guessed you might have missed the binaryness. My experience of the binary update process is fairly positive, but there have been glitches eg my lsof didn't get updated along with a kernel and I now see stuff like lsof: WARNING: compiled for FreeBSD release 6.0-RELEASE; this is 6.0-SECURITY. when using it. I suppose things like lsof needing to be recompiled are an implied dependency and may be hard to find. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: binary updating 6.0-6.2
Jerry McAllister wrote: On Thu, Feb 22, 2007 at 03:38:18PM +, Robin Becker wrote: . What you want to do is a cvsup (Mathew Seaman recently pointed out csup that is supposed to be part of the system, but I don't seem to have it on the machine I am presently on running 6.1) so you would have to install cvsup from ports (/usr/ports/net/cvsup or /usr/ports/net/cvsup-without-gui) if you need it. You seem to be suggesting just what I wish to avoid ie all the trouble of a complete system compile and install with a full ports rebuild. Rather than the src build I would prefer a complete dump and iso reinstall. I looked further into the binary update script and it seems pretty comprehensive. I still expect to go through my ports rebuild, but perhaps I wouldn't need to copy my files back. Even so I see comments in the 6.0-6.1 upgrade.sh like "# XXX For some reason, /usr/lib/liblwres.so had its version bumped. # XXX Leave behind the old version just in case." so we end up with some minor cruft left lying about. If the ports need to be rebuilt anyhow, I think I prefer to start from scratch. I just posted a basic rundown on doing this (but with slightly different numbers) just a couple days ago. You might want to check the questions archive.Briefly, you install cvsup (if not already installed); run the cvsup (or csup if you have it) using a supfile set for *default tag=RELENG_6_2 (I will append my working supfile at the end. It is a simple one) Then build and install everything. Do this stuff in a script(1) session to preserve messages and warnings. cvsup -g -L 2 /etc/supfile62 ...good stuff elided -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
binary updating 6.0-6.2
I see from the latest announcement that freebsd-update is now in the base system for 6.2. Following the links I see there's a procedure for 6.0-61. and another for 6.1-6.2 binary updating. In the past I have always dumped all my files and /etc /usr/local/etc done a completely fresh install, reinstalled files and manually hacked any of the modified system files. Then follows a long process of updating the ports tree and rebuilding the ports I need. Clearly I will still dump all my files in case anything horrific goes wrong. But can I update 6.0-6.1 6.1-6.2 with a reasonable hope of success. The very first step seems the hardest since I have been using freebsd-update already in my 6.0 system and the kernel is no longer a 6.0-RELEASE, but 6.0-SECURITY FreeBSD and a warning in the notes on the 6.1-6.2 update process says I need a GENERIC or SMP kernel. Does the updater know about updates? Does the updater squirrel away copies of the updated files so a last ditch attempt at status quo restoration can be made? Also do my existing ports continue to work with the new kernel? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: BBC debate "Battle of the operating systems"
Bill Moran wrote: In response to Robin Becker <[EMAIL PROTECTED]>: The BBC is to host a debate on multiple OSes. http://news.bbc.co.uk/2/hi/technology/6288119.stm They want one individual to represent each OS. Apparently they only want Vista, OS X and Linux, but I don't see why we can't press for FreeBSD. I sent them a comment about how FreeBSD should be included. I think that if enough of us send in comments, they'll re-open the submission. Where did you mail~ usually there's a link like "tell us your opinion" etc etc, but I could not see one on that page. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: BBC debate "Battle of the operating systems"
Robin Becker wrote: The BBC is to host a debate on multiple OSes. http://news.bbc.co.uk/2/hi/technology/6288119.stm They want one individual to represent each OS. Apparently they only want Vista, OS X and Linux, but I don't see why we can't press for FreeBSD. unfortunately they already closed the applications so I guess FreeBSD will have to remain partially represented by some McFanBoy. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
BBC debate "Battle of the operating systems"
The BBC is to host a debate on multiple OSes. http://news.bbc.co.uk/2/hi/technology/6288119.stm They want one individual to represent each OS. Apparently they only want Vista, OS X and Linux, but I don't see why we can't press for FreeBSD. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: make world for a jail
.. I fixed up my /usr/src tree to be latest RELENG_6_0 in line with my machine and ran cd /usr/src make world DESTDIR=/usr/jails/mailserver however, this seems to produce a lot of stuff in /usr/obj (the normal make world destination). Eventually the make fails with this message ## make: don't know how to make /usr/jails/mailserver/usr/lib/libc.a. Stop *** Error code 2 Stop in /usr/src. *** Error code 1 Stop in /usr/src. ## In line with my earlier observation that the sub makes were being called like this make -f Makefile.inc1 DESTDIR=/usr/obj/usr/src/tmp par-all is there some magic that's required to make world into a specified DESTDIR? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
make world for a jail
I'm trying to make world for a jail build following the recipe in http://www.onlamp.com/pub/a/bsd/2003/09/04/jails.html First off I started by updating the src tree. I copied /usr/share/examples/cvsup/stable-supfile edited the host name then I ran SERVER=`fastest_cvsup -q -c ca,us` cvsup -L2 -h $SERVER /root/bin/stable-supfile that seemd fine and stuff appeared in /usr/src as expected. mkdir /usr/jails mkdir /usr/jails/mailserver cd /usr/src make world DESTDIR=/usr/jails/mailserver Now I get an error building sendmail chmod 444 freebsd.cf rm -f freebsd.submit.cf m4 -D_CF_DIR_=/usr/src/etc/sendmail/../../contrib/sendmail/cf/ /usr/src/etc/sendmail/../../contrib/sendmail/cf/m4/cf.m4 /usr/src/etc/sendmail/freebsd.submit.mc > freebsd.submit.cf chmod 444 freebsd.submit.cf ERROR: Required audit group is missing, see /usr/src/UPDATING. *** Error code 1 I'm not sure what this means. I looked in UPDATING and see references to FreeBSD 6.1-RELEASE; so am I trying to build a 6.1 kernel with my 6.0 system? I'm guessing that *default release=cvs tag=RELENG_6 should have been *default release=cvs tag=RELENG_6_0 One other thing is that during the build I saw sub makes going on that had a different DESTDIR. Eg make -f Makefile.inc1 DESTDIR=/usr/obj/usr/src/tmp par-all how do I get things built into my desired location or is the make world eventually going to put them in the right place? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: find port install options
Warren Block wrote: On Wed, 20 Dec 2006, Warren Block wrote: I suppose its safest to assume that I need to reinstall perl, but how easy is it to do that? I think it was installed by default. Easy, as long as you've updated ports with cvsup or portsnap. As root: [corrected by adding make deinstall] cd /usr/ports/lang/perl5.8 make deinstall make -DWITH_THREADS install make clean .. yes I understood that part :) but I have a bunch of other things dependent on perl so I guess I would need to rebuild all of those as well. And I still need to find out what options were used in building perl. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: find port install options
N.J. Mann wrote: On Wednesday, 20 December, 2006 at 12:42:09 +, Robin Becker wrote: Robin Becker wrote: [...] thread sample script. But I guess I still need to know if any other options were used during the install. /var/db/ports e.g. > cat /var/db/ports/portupgrade/options # This file is auto-generated by 'make config'. # No user-servicable parts inside! # Options for portupgrade-2.1.3.2_1,2 _OPTIONS_READ=portupgrade-2.1.3.2_1,2 WITH_BDB4=true WITHOUT_BDB1=true Cheers, Nick. unfortunately I think my perl came as part of the base install. When I look in /var/db/ports I see only gettext python python23 rsync I'm sure I have built and installed other ports, but it seems ports doesn't store the options for them automatically. In fact Dru Lavigne recommends that if using portupgrade you modify the pkgtools.conf to include the special options. So I have lots of stuff for apache/subversion etc in the make options section, but nothing about perl and as a python person I don't think I intentionally installed perl anywhere. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: small mail server
Jeff Palmer wrote: At 06:23 AM 12/20/2006, you wrote: Robin, I've had much success with the following guide. http://www.thekeyboardcowboys.org/help/fbsd_postfix/FreeBSD_Postfix.html It mentions everything you have above, with SMTP AUTH (you don't have to check mail before sending) Hope it helps, ... I've heard lots about Postfix, but have never used it. Since the box is currently only using sendmail for outgoing stuff I guess it shouln't be too much of a pain to try it out. I've actually been considering using a jail to do this so it should be fairly easy to do. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: small mail server
Dave wrote: Hi, You want to set up a separate mail server from your isp? If that's the case I can do this if you want. I've got this working on a FreeBSD 6.1 box and i quite like it. This would actually give me the excuse i mean motivation to get webmail working on my own box as well. I would base this server on a postfix solution, and for the amound of users your not likely to need a database, that's overkill. HTH Dave. .. Thanks for the offer Dave, I don't think my boss would allow non-employees to access our servers. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: find port install options
Robin Becker wrote: Is it possible to determine what options were used during port installation. One of the recipes I'm trying to follow calls for perl to be installed with USE_THREADS=yes. well I found out that the installed perl doesn't have threads by running a thread sample script. But I guess I still need to know if any other options were used during the install. I suppose its safest to assume that I need to reinstall perl, but how easy is it to do that? I think it was installed by default. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
find port install options
Is it possible to determine what options were used during port installation. One of the recipes I'm trying to follow calls for perl to be installed with USE_THREADS=yes. pkg_info | grep perl produces perl-5.8.8 Practical Extraction and Report Language but that doesn't tell me how it was installed. Trawling through the masses of port reated docs doesn't give me an obvious answer either and make.conf contains only PERL_VER=5.8.8 PERL_VERSION=5.8.8 I suppose its safest to assume that I need to reinstall perl, but how easy is it to do that? I think it was installed by default. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
small mail server
I'm looking for some advice on using one of our existing freebsd 6.x servers as a mail server for a small number (<20) of users. Our existing provider gives us 1) pop3/IMAP for reading mail 2) SMTP for sending, but we need to read mail before using smtp; I guess this implies we don't need to authenticate directly. 3) web based interface for adding users and redirections etc etc 4) spam filtering (presumably based on their large user mail volume). 5) white/grey listing -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: python 2.3 binary package bug
Gerard Seibert wrote: On Wednesday December 06, 2006 at 11:34:40 (AM) Alex Zbyslaw wrote: Robin Becker wrote: Gerard Seibert wrote: .. them. If it is a bug what's the reporting mechanism? man send-pr what an awful interface that was ;( made me grumpy using it. Perhaps I've grown too used to these new fangled web applications :) http://www.freebsd.org/send-pr.html You are right Alex. I had forgotten about that never having used it personally. I have a question though. The PC based send-pr captures certain system attributes. Does the web based one function similarly? I don't believe it does. you're right; part of the cli version has bits already filled in with the release version and (I believe) output from uname -a. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: python 2.3 binary package bug
Gerard Seibert wrote: .. them. If it is a bug what's the reporting mechanism? man send-pr what an awful interface that was ;( made me grumpy using it. Perhaps I've grown too used to these new fangled web applications :) -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
python 2.3 binary package bug
I installed this binary package ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages-6-stable/All/python23-2.3.5_1.tbz on my freebsd 6.1 system and while testing had a problem related to /usr/local/lib/python2.3/lib-dynload/_socket.so (it didn't appear to be there). When I build from the source of the package it seems OK. Is this a bug? The binary install mentioned certain bits might be missing, but socket wasn't one of them. If it is a bug what's the reporting mechanism? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: shell test for stdout=stderr
Matthew Seaman wrote: Robin Becker wrote: Is there a way for a shell script to test if 2> is the same as 1>? I want to put messages in both when they are connected to different files, but would like to avoid duplicating the message when they are the same. You could try using fstat(1) to print out the open file descriptors from your process: fstat -p $$ and then compare the values in the DEV and INUM columns -- unfortunately fstat has no way to map back from those device and inode values to filenames. thanks for the fstat tip this code seems to do what I want fstat -p $$ | awk '{if($4=="1"||$4=="2"){F[$4]=$5+$6+$8}} END{ if(F["1"]==F["2"]) print 1 }' I suppose there might be a problem if F never gets created or one of 1/2 were closed prior to execution. I probably need a BEGIN{F[""]=""} or something to make it a bit more robust. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
shell test for stdout=stderr
Is there a way for a shell script to test if 2> is the same as 1>? I want to put messages in both when they are connected to different files, but would like to avoid duplicating the message when they are the same. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: sysv semaphores
Damian Wiest wrote: ... No, it's my fault; I checked things on the wrong system. OpenBSD uses seminfo, FreeBSD uses ipc. aa the joy of forking :) -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: sysv semaphores
Damian Wiest wrote: . You'll want to use either sysctl(8) to change the settings dynamically, or use /etc/sysctl.conf to modify the settings permanently. I'm not sure if there's a benefit to rolling a new kernel versus using sysctl.conf, or if things even work that way anymore. # sysctl -a | grep seminfo kern.seminfo.semmni=10# number of semaphore identifiers kern.seminfo.semmns=60# number of semaphores in system kern.seminfo.semmnu=30# number of undo structures in system kern.seminfo.semmsl=60# max number of semaphores per id kern.seminfo.semopm=100 # max number of operations per semop call kern.seminfo.semume=10# max number of undo entries per process kern.seminfo.semusz=100 # size in bytes of undo structure kern.seminfo.semvmx=32767 # semaphore maximum value kern.seminfo.semaem=16384 # adjust on exit max value Those comments are from /usr/include/sys/sem.h -Damian Sorry, I forgot to mention a few things. You should become familiar with ipcs(1) as it will allow you to query the current state of SysV IPC facilities. You'll probably find yourself manually deleting semaphores depending on how well that extension cleans up after itself during testing. IIRC, the kernel maintains some in-memory datastructures to keep track of semaphores. I believe increasing the maximum number of semaphores will take up a negligible amount of main memory. If you're interested, the "Design and Implementation of the 4.4 BSD Operating System" (ISBN # 0201549794) has great coverage of this stuff. thanks very much all very useful info. Someone else recommended looking at these options kern.ipc.semmap=180 kern.ipc.semmni=160 kern.ipc.semmns=210 kern.ipc.semmnu=180 kern.ipc.semmsl=210 kern.ipc.semopm=250 kern.ipc.semume=160 kern.ipc.semusz=92 kern.ipc.semvmx=32767 kern.ipc.semaem=16384 and on my 6.1 system I see these with sysctl -a | grep ipc, however, # sysctl -a | grep seminfo /usr/RL_HOME/users/robin: # I guess they've been renamed. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
sysv semaphores
I'm trying to test a python extension (POSH) that uses semaphores. When testing I get a run time error that indicates it requires too many semaphores. Is it possible to adjust the allowed number of semaphores without rebuilding the kernel? What are the costs of having semaphores ie are they memory/cpu intensive? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
kick off a post boot job
I have a number of servers which don't have console access, but I would like to have apache started automatically if the server is rebooted. However, it seems that if https is used then I need to type in a secret at boot time (on the console). Is there a way to start processes up automatically after the boot is finished? I could imagine asking another, trusted, server to supply the magic string using scp or some other secure transport and then using the decoded result to start up apache. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
IP address impersonation
We have a remotely hosted 6.0 server that has apparently been impersonated by a colocated server. The provider allows root access and we have set up our server from a base 6.0 installation. We were allocated an ip address and mostly we have had a good experience with this setup. However, twice in three weeks we have had difficulty in logging in and have had to crash boot the server. Analysis of the logs revealed that another machine on the hoster's network had assigned itself our ip address. Even when we provided the suspect mac address it seemed the hoster had trouble in finding out/appreciating what the problem was. I have little experience of this sort of thing, but can anyone else offer some advice on 1) is this a recognized form of attack? I can see that it could be used for password harvesting and traffic interception, but are there other implications. 2) Are there ways to mitigate this kind of problem? We have other hosted servers on machines with similar (root) access. They presumably could also be impersonated. We found this out by inspection of our own log files; could the provider be doing something more to prevent this? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: denyhosts problems
Daniel Bye wrote: On Thu, Sep 28, 2006 at 05:22:43PM +0100, Robin Becker wrote: I'm trying to get denyhosts-2.5 to work in 6.0 and have inserted a line in hosts.allow ALL: xxx.myoffice.com : allow sshd: /etc/hosts.deniedssh : deny ALL: ALL : allow but am finding that this causes my home ip to be denied even though I log in with a pre-shared key. sshd will still avail itself of libwrap's functionality /before/ the client even has a chance to offer its key. Anyone who manages to get a copy of your key will need also to satisfy your /etc/hosts.allow rules before they can use it. The /etc/hosts.deniedssh file is being created, but my home ip is not present (it would be hard as I have a dynamically allocated one anyhow). The hosts.deniedssh file contains entries like ... ALL : ALL : 61.219.xx.250 : deny : deny which, clearly, is nonsense! I am not writing this file, denyhosts is. Make sure that denyhosts.cfg has a blank value for BLOCK_SERVICE and that it points HOSTS_DENY to the right file. I guess that at least is correct, though. My BAD I have the value ALL for BLOCK_SERVICE, I suppose that's the problem. I read further and it seems I do indeed need to set an empty value. Thanks. DenyHosts will then correctly record only the IP address of blocked hosts, which will result in much saner rule expansions! I have the same setup in 6.1 and it seems to work. But I still see messages related to line 24 from that setup. Does denyhosts work properly? I suspect it is not quite the same - check the BLOCK_SERVICE setting on that machine. You're probably right. Check out the DenyHosts FAQ - it's very useful. http://denyhosts.sourceforge.net/faq.html And the FreeBSD hosts_options(5) man page as well, which, as I said earlier, contains the full story on setting up your /etc/hosts.allow. Thanks again. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: linux recovery tool
Ahmad Arafat Abdullah wrote: ... One of my hosters says they provide a ram based linux recovery tool, but I'm actually using freebsd on the hardware so am wondering if linux knows how to mount UFS partitions? -- Robin Becker Yes, it does .. thanks -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
denyhosts problems
I'm trying to get denyhosts-2.5 to work in 6.0 and have inserted a line in hosts.allow ALL: xxx.myoffice.com : allow sshd: /etc/hosts.deniedssh : deny ALL: ALL : allow but am finding that this causes my home ip to be denied even though I log in with a pre-shared key. The /etc/hosts.deniedssh file is being created, but my home ip is not present (it would be hard as I have a dynamically allocated one anyhow). The hosts.deniedssh file contains entries like ... ALL: 61.219.xx.250 : deny ALL: 209.8.xx.242 : deny ... I am getting an error in the auth log related to the denial that looks like this /etc/hosts.allow, line 24: can't verify hostname: getaddrinfo(xxx-yyy-.., AF_INET) where /etc/hosts.allow line24 corresponds to the sshd: /etc/hosts.deniedssh : deny line in hosts.allow. I have the same setup in 6.1 and it seems to work. But I still see messages related to line 24 from that setup. Does denyhosts work properly? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
linux recovery tool
One of my hosters says they provide a ram based linux recovery tool, but I'm actually using freebsd on the hardware so am wondering if linux knows how to mount UFS partitions? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: gmirror HD failure detection
John Nielsen wrote: .. Use both approaches since they tell you different things which just happen some of the time to coincide. If you happen to be one of the smart admins who actually reviews the output of the periodic scripts, then simply adding daily_status_gmirror_enable="YES" to /etc/periodic.conf will give you a daily health check. If you want more granularity than a single day, you could use the contents of the periodic script as a starting point for rolling your own. JN That's good as I actually get all of our servers to email to me. I have also done the smartd thing with an email output as well. Thanks to all for useful input. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: gmirror HD failure detection
Dave wrote: Hi, I've got smartd going on a gmirror system, however when smartd starts up it says it can't find the various drives. I've tried both the autodetection line as well as specifying the individual drives. If this does work i'd like to know about it as i believe i might have one failing drive, but am not sure which one. Thanks. Dave. well as root I can certainly run smartctl -a /dev/ad4 (or /dev/ad6) so I assume smartd could. I like the idea of using gmirror status -s , but I don't know what the results would be if one of the disks were going bad. Would it change from COMPLETE to DEGRADED suddenly? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
gmirror HD failure detection
After using Dru Lavigne's excellent article http://tinyurl.com/da66a about Raid-1 I have a full Raid-1 mirror on a new rack server. I'm wondering if anyone can tell me how best to monitor the hardware status to detect imminent failure of one of the disks? Do I use something like smartctl in a cron or what? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
serial console install
Hi, my helpful supplier says "If you connect to the serial console and then turn on the machine from the power cycler the machine should boot up to a syslinux boot prompt, which will let you fire up the FreeBSD installer." and when this syslinux thing boots up I see the possibility to boot into a freebsd installer. I assume that I will be able to see the installer in my serial console and will be able to do the normal install, but do I get the opportunity to do some editing of /boot/loader.conf & /etc/ttys to allow the HD boot to show up in my console? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
AMD 64 3000
Anybody have any strong opinions on this for a freebsd 6.1 web server? We are currently using a 2.40GHz celeron which is fairly slow. I'm hesitant to switch to 64 bits, are there any gotchas for freeBSD? -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: vmware on freebsd?
Norberto Meijome wrote: > On Tue, 15 Aug 2006 02:58:23 -0400 (EDT) > Peter <[EMAIL PROTECTED]> wrote: > >> Is it possible to install VMWare Server on FreeBSD 6.0? I'm looking >> for comments from people who may have done this. > > I may be completelly off the mark here...but I think the VMWare Server is more > like XEN rather than Qemu ( which is more VMWare workstation). > > AFAIK, VMWare does not support FreeBSD as a host (YES as a guest) in their > latest versions of the Workstation line. I havent heard of host as for the > Server line, but I could be wrong. > > Have you looked into Xen instead? > > Beto well I'm trying vmware3 at the moment using 6.1 as host; the port seems to build fine, and the wizard now runs to build a setup. I need to get a key in order to actually run a machine, but I think overall it will eventually run XP. Had to change my linux compatibility etc etc. The difficulty with Xen at present is that it won't run XP/Win2k etc etc which vmware will. Xen will eventually get there now that they've got agreements with M$. The real problem with these emulations is how far they are from the real hardware. USB is still "new" so I don't expect to get my usb dvbt working even if a virtual XP runs fine. I know the qemu can do some kind of device control under linux, but I suspect it's harder under freeBSD. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: vmware library problem
dick hoogendijk wrote: On Thu, 10 Aug 2006 07:59:53 + Robin Becker <[EMAIL PROTECTED]> wrote: I fixed this by adding LD_LIBRARY_PATH=/usr/X11R6/lib to the environment, but since I actually have a choice of /usr/compat/linux/usr/X11R6/lib /usr/X11R6/lib which should I actually use? For some reason even though I'm using a fresh portupgraded system the libraries the two locations are not the same. Sure they're not the same. They belong to a different OS! The first are linux libraries, the second thos of FreeBSD. hmmm, I thought they both come from Xorg, it's obvious that different systems might apply different patches, but in practice shouldn't there be just one copy of libX11.so.6 on any given system? Presumably my KDE is operating with the freeBSD patched one so why should linux compatible apps use a different version? If a different compiler were required or different calling conventions assumed then it would be reasonable, but I couldn't see any obvious differences in the compilations. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: vmware library problem
Robin Becker wrote: I have built and installed vmware3 from a recent(3/Aug) cvsup'd ports # vmware Setting TMPDIR=/var/tmp. vmware-mks: error while loading shared libraries: libX11.so.6: cannot open shared object file: No such file or directory vmware-ui: error while loading shared libraries: libX11.so.6: cannot open shared object file: No such file or directory find indicates that the libraries are indeed present. # find / -name libX11.so\* Anyone got any ideas what I'm doing wrong? What exactly should I be doing to get an XP machine image to run? I fixed this by adding LD_LIBRARY_PATH=/usr/X11R6/lib to the environment, but since I actually have a choice of /usr/compat/linux/usr/X11R6/lib /usr/X11R6/lib which should I actually use? For some reason even though I'm using a fresh portupgraded system the libraries the two locations are not the same. -- Robin Becker ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"