Re: Intel EMGD driver on FreeBSD?
On Tue, 27 Dec 2011, Mark Felder wrote: Are you sure that the problem isn't just that the current releases of FreeBSD don't have GEM/KMS support which the newer Intel drivers require? I'm running a test build on my Google CR48 and have OpenGL acceleration without any major hiccups except not being able to get back to a console after starting X. Test here: http://lists.freebsd.org/pipermail/freebsd-current/2011-November/029219.html I converted the image into full install on my hard drive. More details here: http://wiki.freebsd.org/Intel_GPU Thank you for the reply. I've been unable to lay hands on a USB drive big enough to hold that entire image, so I can't easily test with it. From what I'm finding on the web, your Google CR48 is probably an Intel GMA 3150, which looks like it's covered by the normal open source Intel driver even though it's not explicitly listed on the manpage. http://intellinuxgraphics.org/documentation.html I think my GMA 500 is a different situation. I'm pessimistic. ___ 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
Intel EMGD driver on FreeBSD?
I've recently acquired an Asus Eee PC 1201HAB netbook with GMA500 Poulsbo video. My research and experiments show that this hardware is not supported by the usual mainstream Intel driver. The only driver that I seem to have any luck with at all under FreeBSD is vesa, but even vesa has problems. The latest news I've read from the Linux world is that Intel's EMGD Embedded Media and Graphics Driver is the current best solution. If I'm reading the FAQ correctly, the bits I need the most are closed source. http://www.intel.com/p/en_US/embedded/hwsw/software/emgd#faqs Aside from being nailed down to a specific version of OpenGL, I had good experiences using the Nvidia binary blob driver under FreeBSD. I'd be willing to endure that same kind of situation again if that's the only way to get any kind of video acceleration out of this hardware. Is anyone working on making the EMGD work under FreeBSD? Would it make sense to try? ___ 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: kern.geom.journal.stats.low_mem refers to what?
I've recently set up gjournal on top of gmirror on FreeBSD 8.2. I understand that this setup has a lot of redundant writing. It is working, but I'm not sure I've set it up as efficiently as I should. During prolonged writes, such as copying large files to the file system across the network or producing very large logfiles, the low_mem and skipped_bytes statisics rise rapidly and the system becomes less responsive. top always reports free memory, so I don't think that's the issue. journal_full and wait_for_copy have never occurred. Here's a sample of what happens after a couple hours of intense writing... kern.geom.journal.stats.low_mem: 5379 kern.geom.journal.stats.journal_full: 0 kern.geom.journal.stats.wait_for_copy: 0 kern.geom.journal.stats.switches: 7543 kern.geom.journal.stats.combined_ios: 265318 kern.geom.journal.stats.skipped_bytes: 935712768 low_mem sounds like a bad thing. What could I do to remedy that? Did I make the journal too small? The stats say that journal_full has never happened, so maybe not? Is there a setting I should tweak? The Handbook says 1GB is good enough most of the time, but it also says that 3x the amount of physical memory is a good size as well. I compromised between the two and made an 8GB journal for this system that has ~4GB of memory. Replying to myself for posterity... I think I figured it out. I believe low_mem means its running out of gjournal cache, defined by kern.geom.journal.cache.limit, which is initially set by kern.geom.journal.cache.divisor, which is initially 2, which means half of the kernel memory, which is vm.kmem_size, which is 330MB on my i386 8.2 system by default. I don't know where this default comes from, but it seems common. This gives me about 167MB of cache for journalling. During big file transfers, I can watch the journal_data statistic in vmstat -m. It climbs close to the 167MB cache size. That's when I see kern.geom.journal.stats.low_mem tick over and the stutter happens. So I think I understand what I'm seeing now. I could try increasing vm.kmem_size and possibly the KMEM_SIZE option that's compiled into my kernel as described in the ZFS tuning documents, since it's a similar issue. This would allow the kernel to have more memory and thus allow gjournal to use more memory for caching. If I'm consistently filling the cache faster than I can empty it, making the cache bigger probably won't help. This journal is on top of a mirror, so I'm always expecting it to read faster than it writes... Copying large files locally might always max out the buffer? No harm in trying, I guess. Or I could just live with it. It's doing what I told it to do. ___ 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
kern.geom.journal.stats.low_mem refers to what?
I've recently set up gjournal on top of gmirror on FreeBSD 8.2. I understand that this setup has a lot of redundant writing. It is working, but I'm not sure I've set it up as efficiently as I should. During prolonged writes, such as copying large files to the file system across the network or producing very large logfiles, the low_mem and skipped_bytes statisics rise rapidly and the system becomes less responsive. top always reports free memory, so I don't think that's the issue. journal_full and wait_for_copy have never occurred. Here's a sample of what happens after a couple hours of intense writing... kern.geom.journal.stats.low_mem: 5379 kern.geom.journal.stats.journal_full: 0 kern.geom.journal.stats.wait_for_copy: 0 kern.geom.journal.stats.switches: 7543 kern.geom.journal.stats.combined_ios: 265318 kern.geom.journal.stats.skipped_bytes: 935712768 low_mem sounds like a bad thing. What could I do to remedy that? Did I make the journal too small? The stats say that journal_full has never happened, so maybe not? Is there a setting I should tweak? The Handbook says 1GB is good enough most of the time, but it also says that 3x the amount of physical memory is a good size as well. I compromised between the two and made an 8GB journal for this system that has ~4GB of memory. ___ 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: Shutting down X with control+alt+backspace
On Sun, 7 Jun 2009, Manolis Kiagias wrote: Manolis Kiagias wrote: Luke Dean wrote: This is an answer to a question I started to post, but then decided to research instead. I know many readers of this list use the feature I'm describing. When Xorg was upgraded to version 7.4, the historic ability to shut down X with Control+Alt+Backspace became a non-default option. The solution to re-enabling this behavior was to add Option DontZap off to the ServerLayout or ServerFlags section of xorg.conf as documented in a note in the Handbook http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/x-config.html A few days ago, x11/xkeyboard-config was upgraded to 1.6 and the solution in the Handbook is no longer sufficient. The new solution that gets Control+Alt+Backspace working for me again is to add Option XKbOptions terminate:ctrl_alt_bksp to the InputDevice section of xorg.conf. Thanks for mentioning this. I have not yet upgraded to the new version of xkeyboard-config, but will try this and update the Handbook accordingly. This gets even more complicated - the setting in xorg.conf will only be effective when AutoAddDevices is false (or AllowEmptyInput is false). On systems that totally rely on HAL for device detection, the setting has to be moved to an XML file like this: ?xml version=1.0 encoding=UTF-8? deviceinfo version=0.2 device match key=info.capabilities contains=input.keyboard merge key=input.x11_driver type=stringkbd/merge merge key=input.xkb.Model type=stringpc105/merge merge key=input.xkb.Layout type=stringus/merge merge key=input.xkb.Rules type=stringxorg/merge merge key=input.xkb.Options type=stringterminate:ctrl_alt_bksp/merge /match /device /deviceinfo which should be named i.e. keyboard.fdi and placed in /usr/local/etc/hal/fdi/policy In light of the above, I feel we probably need to add a section on Configuring Additional Options Using HAL to the Handbook. Oh, okay. I'm not running HAL and I do have AllowEmptyInput and AutoAddDevices off, so just modifying the xorg.conf was enough for me. So far I've avoided HAL because it seems complicated and scary, and I don't know what the benefits of using it would be aside from some automounting tricks that I can live without. It does seem to be the wave of the future however, so any documentation that might help ease me into that transition would be appreciated. ___ 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
Shutting down X with control+alt+backspace
This is an answer to a question I started to post, but then decided to research instead. I know many readers of this list use the feature I'm describing. When Xorg was upgraded to version 7.4, the historic ability to shut down X with Control+Alt+Backspace became a non-default option. The solution to re-enabling this behavior was to add Option DontZap off to the ServerLayout or ServerFlags section of xorg.conf as documented in a note in the Handbook http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/x-config.html A few days ago, x11/xkeyboard-config was upgraded to 1.6 and the solution in the Handbook is no longer sufficient. The new solution that gets Control+Alt+Backspace working for me again is to add Option XKbOptions terminate:ctrl_alt_bksp to the InputDevice section of xorg.conf. ___ 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
How to recover disk space after filesystem full
Yes, it sounds like a stupid question, but let me tell the story. The log for my dhcp server filled up /var last night, which meant that dhcpd was also unable to hand out new leases, which meant that I had effectively been DOSed. I'll have to look into changing my logging policies. So, to correct the problem, I log into the router, removed the big log and several other files in /var to free up some space, and assumed this would correct the problem. It did not. Several minutes after freeing up a lot of space on /var, I continued to get filesystem full messages and df continued to show the capacity at 100%. I checked df -i for the inodes too. That was fine. I ran a quick fsck to see if that might shock the system into seeing all the space that I'd freed up, but no good. I ended up rebooting the box. Was there any other possible solution I could've tried? Why wouldn't the free space immediately appear as free? ___ 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: How to recover disk space after filesystem full
On Fri, 22 May 2009, Steve Bertrand wrote: Luke Dean wrote: I ended up rebooting the box. Was there any other possible solution I could've tried? You have to restart the service that was holding the log file(s) open. The system does not release the space while an application is 'using' the file, even after it's been deleted. Oh yeah! I forgot. I've got it configured to use syslogd to handle the dhcp logging, so I probably just needed to restart that one. I probably could've left dhcpd running. This kind of emergency always seems to happen before I get a chance to make coffee. Thank you. ___ 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
ath_hal problem on slow hardware, can be tuned?
I'm following 7-STABLE with my laptop. As soon as ath_hal appeared, I started having trouble with my wireless connection dropping every few hours. +ath0: ath_chan_set: unable to reset channel 11 (2462 Mhz, flags 0x480 hal flags 0xc0), hal status 3 The ath manpage documents this error as: ath%d: ath_chan_set: unable to reset channel %u (%u Mhz) The Atheros Hardware Access Layer was unable to reset the hardware when switching channels during scanning. This should not happen. sys/contrib/dev/ath/ah.h documents the error 3 as: HAL_EIO = 3,/* Hardware didn't respond as expected */ I can run /etc/rc.d/netif start dhclient ath0 to reset the interface and reconnect to the wireless network after this happens. I'm thinking that ath_hal doesn't like my card as much as the old ath driver system did, or perhaps there's some tunable that I need to adjust. This is an old laptop. It's a Sony Vaio PCG-Z505S - a celeron 133MHz designed for MS Windows 98. In the best of circumstances, it can never establish a wireless connection before ntp comes up at boot time. wpa_supplicant always gives up the first time it attempts to connect because this hardware is so slow. I'm saying it's slow. I wonder if it's so slow that it's not responding as fast as ath_hal wants it to respond whenever it decides to rescan for channels. I know that such problems can sometimes be solved with sysctls or adjustment of constants. Would anyone have any suggestions? The wireless card is a DWL-G650 pccard, revision B2 pciconf -vlbc says this about it: a...@pci0:2:0:0:class=0x02 card=0x32021186 chip=0x0013168c rev=0x01 hdr=0x00 vendor = 'Atheros Communications Inc.' device = 'AR5212, AR5213 802.11a/b/g Wireless Adapter' class = network subclass = ethernet bar [10] = type Memory, range 32, base 0x8800, size 65536, enabled cap 01[44] = powerspec 2 supports D0 D3 current D0 ___ 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
Noisy GEOM_LABEL on boot of 7-STABLE
I just upgraded my old laptop from an older 7.1 release to the latest 7-STABLE for i386, and I started getting a lot of new GEOM_LABEL noise during the boot process. It says it's removing and adding labels every time it does the filesystem checks. What's that all about? Should I be concerned? This is just an old laptop with a simple out-of-the-box UFS filesystem with no frills or geom features that aren't default. I have never run tunefs or used glabel. kern.geom.label.debug=0 FreeBSD 7.2-PRERELEASE #1: Tue Apr 28 12:16:44 PDT 2009 ad0: 57231MB TOSHIBA MK6034GAX AC101A at ata0-master UDMA33 GEOM_LABEL: Label for provider ad0s1a is ufsid/47225356153f5b56. GEOM_LABEL: Label for provider ad0s1d is ufsid/472253592de7e9f5. GEOM_LABEL: Label for provider ad0s1e is ufsid/47225356139009e4. GEOM_LABEL: Label for provider ad0s1f is ufsid/47225356087c310e. Trying to mount root from ufs:/dev/ad0s1a . . snip . . Starting file system checks: GEOM_LABEL: Label ufsid/47225356153f5b56 removed. /dev/ad0s1a: FILE SYSTEM CLEAN; SKIPPING CHECKS /dev/ad0s1a: clean, 111842 free (1610 frags, 13779 blocks, 0.6% fragmentation) GEOM_LABEL: Label for provider ad0s1a is ufsid/47225356153f5b56. GEOM_LABEL: Label ufsid/47225356139009e4 removed. /dev/ad0s1e: FILE SYSTEM CLEAN; SKIPPING CHECKS /dev/ad0s1e: clean, 253810 free (34 frags, 31722 blocks, 0.0% fragmentation) GEOM_LABEL: Label for provider ad0s1e is ufsid/47225356139009e4. GEOM_LABEL: Label ufsid/47225356087c310e removed. /dev/ad0s1f: FILE SYSTEM CLEAN; SKIPPING CHECKS /dev/ad0s1f: clean, 23004564 free (148036 frags, 2857066 blocks, 0.6% fragmentation) GEOM_LABEL: Label for provider ad0s1f is ufsid/47225356087c310e. GEOM_LABEL: Label ufsid/472253592de7e9f5 removed. /dev/ad0s1d: FILE SYSTEM CLEAN; SKIPPING CHECKS /dev/ad0s1d: clean, 541653 free (589 frags, 67633 blocks, 0.1% fragmentation) GEOM_LABEL: Label for provider ad0s1d is ufsid/472253592de7e9f5. GEOM_LABEL: Label ufsid/47225356153f5b56 removed. Setting hostuuid: 9f2d90ee-5dba-11dd-b39c-080046044d85. Setting hostid: 0x51e2e234. Mounting local file systems: GEOM_LABEL: Label ufsid/47225356139009e4 removed. GEOM_LABEL: Label ufsid/47225356087c310e removed. GEOM_LABEL: Label ufsid/472253592de7e9f5 removed. ___ 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
boot0cfg: read /dev/ad1: Input/output error using nanobsd
I'm running FreeBSD 7.0 on a soekris 4801 with an image built by nanobsd. It's a small piece of headless hardware that boots from a compact flash drive - no moving parts. I wanted to update the operating system to 7.1 and install some more packages on it, so I built a new image with nanobsd and uploaded it to the second partition using nanobsd's updatep2 tool. The new image mounts fine, but the last line of updatep2, boot0cfg -s 2 -v ${NANO_DRIVE} fails with the message I've quoted in the subject line. The machine boots fine, but I can't provoke any kind of response from boot0cfg except for input/output errors. I'd like to make the machine start booting from the second slice. Any ideas? ___ 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: boot0cfg: read /dev/ad1: Input/output error using nanobsd
I'm running FreeBSD 7.0 on a soekris 4801 with an image built by nanobsd. It's a small piece of headless hardware that boots from a compact flash drive - no moving parts. I wanted to update the operating system to 7.1 and install some more packages on it, so I built a new image with nanobsd and uploaded it to the second partition using nanobsd's updatep2 tool. The new image mounts fine, but the last line of updatep2, boot0cfg -s 2 -v ${NANO_DRIVE} fails with the message I've quoted in the subject line. The machine boots fine, but I can't provoke any kind of response from boot0cfg except for input/output errors. I'd like to make the machine start booting from the second slice. Any ideas? More information... The number of heads that the diskinfo reports is different depending on whether the compact flash card is plugged into the soekris box or mounted in a USB card reader/writer. The usb reader/writer reports: da0 512 # sectorsize 2052513792 # mediasize in bytes (1.9G) 4008816 # mediasize in sectors 249 # Cylinders according to firmware. 255 # Heads according to firmware. 63 # Sectors according to firmware. The soekris box reports: ad1 512 # sectorsize 2052513792 # mediasize in bytes (1.9G) 4008816 # mediasize in sectors 3977# Cylinders according to firmware. 16 # Heads according to firmware. 63 # Sectors according to firmware. This is the same compact flash card. I gave up on boot0cfg and booting from the second partition, so I took the compact flash card out of the soekris box, connected it to my workstation with a USB card reader, and wrote a whole new image to it. Even after doing this, boot0cfg still won't work. I noticed some new messages on the console when I tried boot0cfg -v ad1: ata0: FAILURE - non aligned DMA transfer attempted ad1: setting up DMA failed boot0cfg: read /dev/ad1: Input/output error ad1 is attached to ata0, per dmesg: ad1: 1957MB LEXAR ATA FLASH CARD 20060911 at ata0-slave WDMA2 I think this means there is some kind of geometry problem here, but I don't know how to fix it. I'd like to find a solution to this, but it's not critical. ___ 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: X.org-update - screen garbled / settings for xorg.conf?
On Fri, 30 Jan 2009, Ewald Jenisch wrote: Hi, Today I upgraded my system (ports) so I got x.org 7.4 installed. Given the various entries in /usr/ports/UPDATING about the recent updates for xorg I wonder what the correct settings/combinations are for my installation: I've got hald running as per /etc/rc.conf with hald_enable=YES as well as moused (moused_enable=YES in rc.conf). My login is via console from where I manually start X11 (gnome) via startx. Here's what I've got in my xorg.conf: # Added after upgrade of X11 Section ServerFlags Option AllowEmptyInput false EndSection the following entries where there even before the upgrade: Section InputDevice Identifier Keyboard0 Driver kbd Option Corekeyboard Option XkbModel pc105 Option XkbRules xorg Option XkbLayout de Option XkbVariant nodeadkeys EndSection Section InputDevice Identifier Mouse0 Driver mouse Option Protocol auto Option Device /dev/sysmouse Option ZAxisMapping 4 5 6 7 EndSection Here are my questions: o) When I start X it basically works, but then I get out of gnome, ending up at the commandline and start X11 again (startx), the screen becomes garbled and everything freezes completely. Not only is any input via the console disabled but also no reaction on already established ssh-sessions - complete lockup. I'm using the radeonhd driver for my graphics card is a ATI Radeon HD 2400 XT (RV610 chipset). o) Are the above settings (i.e. AllowEmptyInput and the sections for keyboard/mouse) correct for my combination of hald and moused running? Thanks much in advance for your help, -ewald Me too. I started a thread a few days ago titled Restarting new Xorg freezes system. My symptoms are the same as yours. I'm not running hald at all. I'm not using full-blown gnome, just windowmaker. I'm using the radeon driver on an ATI HD 4350 I just tried switching to the vesa driver, and now I can stop and start X as much as I want with no problems whatsoever. I know that this was not a problem with the radeon driver prior to the Xorg update. ___ 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: Restarting new Xorg freezes system
On Fri, 30 Jan 2009, Alain G. Fabry wrote: On Thu, Jan 29, 2009 at 10:18:10PM -0800, Luke Dean wrote: On Wed, 28 Jan 2009, Glen Barber wrote: Luke Dean wrote: The operating system and ports were all cvsupped and built from source today. I'm running Xorg + windowmaker + the radeon driver on i386 7-STABLE My xorg.conf was built from scratch by Xorg -configure, plus I added Option AllowEmptyInput off to the ServerLayout section, per /usr/ports/UPDATING, since I'm not running hal and I can't get keyboard or mouse otherwise. Everything appears to be great. My applications launch much faster than they did before the upgrade. Shutting down and restarting X is the only problem I'm having. This was not a problem before I upgraded to the new Xorg. This is a desktop system. I launch X with startx from the console. I can shut down X either by exiting windowmaker or by killing the Xorg server with ctrl+alt+backspace. There appear to be some failure messages on the console when I shut down X, but I don't know what they mean or if they're important. I'm attaching a log below. I don't see any stuck processes after the shutdown, but perhaps I don't know what I'm looking for. If I later restart X with startx, some corrupted graphical junk appears on the screen and the system freezes solid. Keyboard and mouse are completely unresponsive. NumLock light won't change. I can't ssh into the system either. I can reproduce this at will, but my hard drives don't like it. Any ideas how to troubleshoot this? Do I need to switch to hal? Have you read /usr/ports/UPDATING? Well.. yeah... See my second paragraph above. Is there something in there that I'm not seeing that explains why restarting X would hang the whole system? ___ You by any chance don't have dual-headed X configured? I had a similar problem and after removing my second ServerLayout section + RgbPath + 'AllowEmptyInput' in xorg.conf, everything seems to work again (even my dual display ???) Section ServerLayout Identifier X.org Configured Screen 0 PANEL 0 0 InputDeviceMouse0 CorePointer InputDeviceKeyboard0 CoreKeyboard Option AllowEmptyInput off EndSection # Removed after upgrade xorg... #Section ServerLayout # Identifier X.org Configured # Screen 0 PANEL 0 0 # Screen 1 VGA_1 RightOf PANEL # InputDeviceMouse0 CorePointer # InputDeviceKeyboard0 CoreKeyboard #Option AllowEmptyInput off # Option Xinerama true #EndSection No, just one monitor and just one ServerLayout section. I just discovered that switching from radeon to vesa makes the problem go away for me. ___ 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: Restarting new Xorg freezes system
On Wed, 28 Jan 2009, Glen Barber wrote: Luke Dean wrote: The operating system and ports were all cvsupped and built from source today. I'm running Xorg + windowmaker + the radeon driver on i386 7-STABLE My xorg.conf was built from scratch by Xorg -configure, plus I added Option AllowEmptyInput off to the ServerLayout section, per /usr/ports/UPDATING, since I'm not running hal and I can't get keyboard or mouse otherwise. Everything appears to be great. My applications launch much faster than they did before the upgrade. Shutting down and restarting X is the only problem I'm having. This was not a problem before I upgraded to the new Xorg. This is a desktop system. I launch X with startx from the console. I can shut down X either by exiting windowmaker or by killing the Xorg server with ctrl+alt+backspace. There appear to be some failure messages on the console when I shut down X, but I don't know what they mean or if they're important. I'm attaching a log below. I don't see any stuck processes after the shutdown, but perhaps I don't know what I'm looking for. If I later restart X with startx, some corrupted graphical junk appears on the screen and the system freezes solid. Keyboard and mouse are completely unresponsive. NumLock light won't change. I can't ssh into the system either. I can reproduce this at will, but my hard drives don't like it. Any ideas how to troubleshoot this? Do I need to switch to hal? Have you read /usr/ports/UPDATING? Well.. yeah... See my second paragraph above. Is there something in there that I'm not seeing that explains why restarting X would hang the whole system? ___ 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
Restarting new Xorg freezes system
The operating system and ports were all cvsupped and built from source today. I'm running Xorg + windowmaker + the radeon driver on i386 7-STABLE My xorg.conf was built from scratch by Xorg -configure, plus I added Option AllowEmptyInput off to the ServerLayout section, per /usr/ports/UPDATING, since I'm not running hal and I can't get keyboard or mouse otherwise. Everything appears to be great. My applications launch much faster than they did before the upgrade. Shutting down and restarting X is the only problem I'm having. This was not a problem before I upgraded to the new Xorg. This is a desktop system. I launch X with startx from the console. I can shut down X either by exiting windowmaker or by killing the Xorg server with ctrl+alt+backspace. There appear to be some failure messages on the console when I shut down X, but I don't know what they mean or if they're important. I'm attaching a log below. I don't see any stuck processes after the shutdown, but perhaps I don't know what I'm looking for. If I later restart X with startx, some corrupted graphical junk appears on the screen and the system freezes solid. Keyboard and mouse are completely unresponsive. NumLock light won't change. I can't ssh into the system either. I can reproduce this at will, but my hard drives don't like it. Any ideas how to troubleshoot this? Do I need to switch to hal? This is the console log. The part after the wsoundserver warnings is the shutdown. Script started on Tue Jan 27 10:44:04 2009 [0:usr/home/lukas startx xauth: creating new authority file /home/lukas/.serverauth.1413 X.Org X Server 1.5.3 Release Date: 5 November 2008 X Protocol Version 11, Revision 0 Build Operating System: FreeBSD 7.1-STABLE i386 Current Operating System: FreeBSD sonata.lukas.is-a-geek.org 7.1-STABLE FreeBSD 7.1-STABLE #0: Tue Jan 27 08:36:33 PST 2009 lu...@sonata.lukas.is-a-geek.org:/usr/obj/usr/src/sys/CUSTOM i386 Build Date: 27 January 2009 08:59:48AM Before reporting problems, check http://wiki.x.org to make sure that you have the latest version. Markers: (--) probed, (**) from config file, (==) default setting, (++) from command line, (!!) notice, (II) informational, (WW) warning, (EE) error, (NI) not implemented, (??) unknown. (==) Log file: /var/log/Xorg.0.log, Time: Tue Jan 27 10:44:07 2009 (==) Using config file: /etc/X11/xorg.conf object id 0005 01 src object id 2115 21 record type 1 record type 4 object id 000f 01 src object id 2116 22 record type 4 record type 7 object id 0002 02 src object id 2121 33 src object id 2116 22 record type 1 record type 2 record type 4 finished output detect: 0 Dac detection success finished output detect: 1 finished all detect before xf86InitialConfiguration in RADEONProbeOutputModes Dac detection success after xf86InitialConfiguration failed to set mtrr: Invalid argument Output 68 disable success Blank CRTC 0 success Disable CRTC 0 success Disable CRTC memreq 0 success Blank CRTC 1 success Disable CRTC 1 success Disable CRTC memreq 1 success mc fb loc is 00ef00d0 (EE) RADEON(0): Acceleration initialization failed Output 68 disable success Blank CRTC 0 success Disable CRTC 0 success Disable CRTC memreq 0 success Blank CRTC 1 success Disable CRTC 1 success Disable CRTC memreq 1 success Output 68 disable success Blank CRTC 0 success Disable CRTC 0 success Disable CRTC memreq 0 success Mode 1280x1024 - 1688 1066 5 freq: 10800 best_freq: 10800 best_feedback_div: 48 best_ref_div: 2 best_post_div: 6 Set CRTC PLL success Set CRTC Timing success Not using RMX scaler 0 setup success Set CRTC 0 Source success Output DAC1 setup success Output 68 enable success Enable CRTC memreq 0 success Enable CRTC 0 success Unblank CRTC 0 success Blank CRTC 1 success Disable CRTC 1 success Disable CRTC memreq 1 success wsoundserver warning: could not locate specified file wsoundserver warning: could not locate specified file wsoundserver warning: could not locate specified file waiting for X server to shut down XIO: fatal IO error 35 (Resource temporarily unavailable) on X server :0.0 after 2149 requests (2149 known processed) with 0 events remaining. XIO: fatal IO error 35 (Resource temporarily unavailable) on X server :0.0 after 1376 requests (1376 known processed) with 0 events remaining. XIO: fatal IO error 35 (Resource temporarily unavailable) on X server :0.0 after 2344 requests (2342 known processed) with 0 events remaining. XIO: fatal IO error 35 (Resource temporarily unavailable) on X server :0.0 after 48875 requests (124 known processed) with 0 events remaining. XIO: fatal IO error 35 (Resource temporarily unavailable) on X server :0.0 after 18635 requests (18628 known processed) with 0 events remaining. Output 68 disable success wmCalendar: Fatal IO error 35 (Resource temporarily unavailable) on X server :0.0. Blank CRTC 0 success wmfire:
Re: Port forwarding behind two routers
On Wed, 19 Nov 2008, Jakub T wrote: 2008/11/15 Luke Dean [EMAIL PROTECTED] Port-forwarding through two NATs is something I've never had any success with. I have a few suggestions that have worked for me and my friends with this setup. A) Disable NAT on the ADSL router. I think the term is bridged mode. Turn it into a dumb box and shift all the NAT/firewall/routing responsibilities over to your wireless router. Depending on your ISP, the hardware, and the protocols involved, this may not be an option for you. B) Disable NAT on the wireless router. This allows it to be a simple switch and wireless access point. The price is that you're probably relying on the DHCP server in the wireless router for your wireless devices and you'll have to disable the DHCP when you disable NAT. This creates new problems to be solved. C) Plug the FreeBSD box into the ADSL router, skipping the wireless router. Your wireless devices will still be double-NATted, but if you're not running servers on them, you might be able to live with that. Luke, Thank you very much, your advices were very helpful and I now have a working port forwarding through two routers. Sorry for the delay in the answering, it took me some time to test various options... Actually your (A) advice is what did the job. I turned off DHCP server on ADSL router and enabled NAT - DMZ Host option on it (for which I realized that it was the closest to your description of bridged mode). Then I configured the wireless router to use static IP config instead of expecting DHCP server. The situation is now this: INTERNET | telephone/adsl-wire | | ADSL router wan : xx.xx.xx.xx FreeBSD box (wired) lan : 192.168.1.1 ip: 192.168.0.102 | laptopgateway: 192.168.0.1 | (wireless)| [internet plug]ip: 192.168.0.101 | Wireless router gateway: 192.168.0.1 | wan : 192.168.1.2:| lan : 192.168.0.1 . . . . . :| [ethernet plug] | | | +---+ DMZ host for ADSL router is 192.168.1.2 -- and it works! I have one question more (forgive my ignorance): now the wireless router is configured to use static IP config and I must provide one or more Static DNS servers to it. Is it ok to type just 192.168.1.1 as DNS (which works for now) or to copy DNS servers which are automatically provided to the ADSL router by the ISP? Your solution is a little different from what I was suggesting, but it might be a better solution in some ways. If 192.168.1.1 really works as a source of DNS, I would take that to mean that your ADSL router is passing your name requests along to the nameservers that the ISP provided it. That's good. If your ISP ever moves its nameservers, it will tell your ADSL box about it, and the changes should propogate. If you hardcoded your DNS addresses into your wireless router, you would have to change them by hand if a change was ever required. I believe your wireless router is now responsible for being the firewall for your network, so make sure you've set that up. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Port forwarding behind two routers
On Sat, 15 Nov 2008, Jakub T wrote: Good day people, I'm trying to get wireless Internet access for my laptop and to use this wireless router as a switch for my FreeBSD box at the same time. This wireless router has one Internet plug and for Ethernet plugs for wired boxes. Now I have this situation: INTERNET | telephone/adsl-wire | | ADSL router wan : xx.xx.xx.xx FreeBSD box (wired) lan : 192.168.1.1 ip: 192.168.0.102 | laptopgateway: 192.168.0.1 | (wireless)| [internet plug]ip: 192.168.0.101 | Wireless router gateway: 192.168.0.1 | lan : 192.168.0.1 . . . . . :| [ethernet plug] | | | +---+ The wireless router software configured the router like this: Destination LAN IP Subnet Mask GatewayInterface 0.0.0.0 0.0.0.0 192.168.1.1 WAN (Internet) 192.168.0.0 255.255.255.0 192.168.0.1 LAN Wireless 192.168.1.0 255.255.255.0 192.168.1.2 WAN (Internet) ... so it works as a switch for two boxes and as a router at the same time. The FreeBSD box is configured like this: ifconfig_XXX0=inet 192.168.0.102 netmask 255.255.255.0 defaultrouter=192.168.0.1 Now I have Internet connection on both computers. However, I can't get aMule and other apps that need port forwarding working on FreeBSD box. First, I tried to configure ADSL router (192.168.1.1) just to forward 4662 port to 192.168.0.102, doesn't work. Then, I tried this: 192.168.1.1 router: forward 4662 to 192.168.0.1 192.168.0.1 router: forward 4662 to 192.168.0.102 not working again. I have a feeling that I'm missing something very simple, but can't figure out what. (A note: before I acquired a wireless router, forwarding with one router was just working, with FreeBSD box configured as 192.168.1.101, so that side of things is ok. And, no, it's not possible to use just wireless router because I can't plug telephone wire in it.) Can anyone help me? Should I post more details? TIA, Jakub Port-forwarding through two NATs is something I've never had any success with. I have a few suggestions that have worked for me and my friends with this setup. A) Disable NAT on the ADSL router. I think the term is bridged mode. Turn it into a dumb box and shift all the NAT/firewall/routing responsibilities over to your wireless router. Depending on your ISP, the hardware, and the protocols involved, this may not be an option for you. B) Disable NAT on the wireless router. This allows it to be a simple switch and wireless access point. The price is that you're probably relying on the DHCP server in the wireless router for your wireless devices and you'll have to disable the DHCP when you disable NAT. This creates new problems to be solved. C) Plug the FreeBSD box into the ADSL router, skipping the wireless router. Your wireless devices will still be double-NATted, but if you're not running servers on them, you might be able to live with that. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: I've just found a new and interesting spam source - legitimate bounce messages
On Thu, 16 Oct 2008, Matthew Seaman wrote: Until the wonderful day that the entire internet abides by these rules[*], use of technologies like SPF and DKIM can discourage but not entirely prevent the spammers from joe-jobbing you. I just started getting these bouncebacks en masse this week. My mail provider publishes SPF records. If the names and numbers in the bouceback messages are to be believed, however, the spammers have defeated SPF by hijacking DNS. The poor recipients never see my SPF records because they're looking at the wrong IP address. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Nvidia Driver - OpenGL - Compiz
On Tue, 12 Aug 2008, Jeff Molofee wrote: Can anyone tell me why, how to fix, or even what is happening on my machine. I ran the 169.x driver for a long time with no issues at all. I decided to upgrade to 173, and noticed instantly that any time I run an OpenGL application X crashes instantly. The odd thing is that I am not seeing any errors on the screen (maybe going off too fast), and I do not see any errors in any of the log files. X shuts down instantly, my machine goes black, X starts back up, and I sit here staring at the log in screen. If I go back to the 169 driver, the issues does not occur. I'm not positive it's OpenGL applications, but it seems any game or GL application I try crashes it instantly. Not even a second to acknowledge the crash... screen goes black and that's it. I'm running FreeBSD 7.X (RELENG_7), nothing custom in the kernel, Intel 6600, GeForce 8800, compiz, gnome, and nothing really fancy. I've had other tell me to roll back to 169, but no one can tell me why this is happening, if it's a problem for everyone, if there is a fix, etc. Any bits of information would be greatly appreciated. There should be some useful information in /var/log/Xorg.0.log just after X crashes. If you restart X after the crash, I believe the old log would get rotated to /var/log/Xorg.0.log.old Since x11/nvidia-driver is a binary driver that was linked against a specific version of OpenGL, it ships that version of OpenGL as a replacement for whatever's on your system at the time. Maybe your OpenGL applications need to be rebuilt? I'm not even sure if that suggestion makes sense, but it's something to try. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: aterm whacked?
On Sat, 12 Jul 2008, Neil Short wrote: I just mem-wiped my laptop and am in the process of installing everything after. I have several issues (linux-firefox with linux-flash plugin crashes on sites with flash), not the least of which is aterm doesn't seem to behave very well. When I run aterm I get: $ aterm aterm has encountered the following problem interacting with X Windows : Request: 64,Error: 8(BadMatch (invalid parameter attributes)) in resource: 0x14D aterm has encountered the following problem interacting with X Windows : Request: 64,Error: 8(BadMatch (invalid parameter attributes)) in resource: 0x14D aterm has encountered the following problem interacting with X Windows : Request: 64,Error: 8(BadMatch (invalid parameter attributes)) in resource: 0x14D etc. Transparency doesn't work right either - which I see is the greatest attraction to aterm. Any ideas? I had this problem with an old Sony Vaio about a year ago. Xorg detected that the correct video driver to use was neomagic. Whatever automatic configuration method I was using at the time decided that I should be running with 32-bit color depth in xorg.conf. It turns out that the neomagic driver doesn't support color depth greater than 24-bits... This misconfiguration worked well enough until something like aterm tried to use transparency, then it would complain about the BadMatch, referring to the color depth that the driver supported versus the color depth that X was configured for, presumably. I found two solutions. One was to edit the Screen section of xorg.conf to remove the entry with DefaultDepth 32. The other was to edit xorg.conf to use the vesa driver instead of the neomagic driver. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Can't run WindowMaker. Fatal server error:
On Sun, 6 Jul 2008, Pedro Alves wrote: Hello I have a serious problem with windowmaker. Every time I configure .xinitrc to launch wmaker, the server aborts here what I get: Fatal server error: Caught signal 11. Server aborting xnit: connection to X server lost. wmaker warning: got signal1-exiting... So... What to do, as I don't find any information concerning this problem. I am running FreeBSD 7.0 One week old cvs ports upgrade. You may find clues in the Xorg log. /var/log/Xorg.0.log When I have this problem, or just about any X-related problem, it means I need to go rebuild my video driver. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Question about torrents via console
Hi all. Ok, I'm curious of something. I've done torrents before via the graphical interface before, but I want to setup a way to download isos and various FOSS apps via bittorrent, but I want to do it via the console so I can start a torrent in screen and then walk away, allowing my server to finish the work without having to leave my main workstation running to do the work like I normally do. I'm using bittornado right now, and if there's a way to do this, I'd love to know how. I run bittornado's btlaunchmany.py script through screen (sysutils/screen) for the very reason you describe. It's been working great for years. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Stop building all those kernel modules
I'm trying to use nanobsd to build a small kernel for an embedded system on FreeBSD 7. In previous versions, I could modify GENERIC by just commenting out all the devices and options that I wasn't interested in, and buildkernel KERNCONF=myconf would happily ignore all the things I wasn't interested in. This no longer works. Now everything gets built, even things that are broken/incompatible with other options I've chosen. I've tried setting NO_MODULES=YES, but the modules were still built. I see documentation for WITHOUT_MODULES, but it wants me to list every module I don't want to build? Come on! There must be a better way... ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Stop building all those kernel modules
On Sat, 17 May 2008, Luke Dean wrote: In previous versions, I could modify GENERIC by just commenting out all the devices and options that I wasn't interested in, and buildkernel KERNCONF=myconf would happily ignore all the things I wasn't interested in. This no longer works. Now everything gets built, even things that are broken/incompatible with other options I've chosen. Ah, I was confused. It built all of the modules in the past as well. It just wasn't a problem before because everything built correctly. Today it appears to be bombing while compiling the zyd device, which wouldn't be so annoying if it weren't at the end of the alphabet... I'm going to clean up and start over from the beginning. If it happens again and I can't figure it out, I'll post my kernel configuration and the error message. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Stop building all those kernel modules
On Sat, 17 May 2008, Luke Dean wrote: On Sat, 17 May 2008, Luke Dean wrote: In previous versions, I could modify GENERIC by just commenting out all the devices and options that I wasn't interested in, and buildkernel KERNCONF=myconf would happily ignore all the things I wasn't interested in. This no longer works. Now everything gets built, even things that are broken/incompatible with other options I've chosen. Ah, I was confused. It built all of the modules in the past as well. It just wasn't a problem before because everything built correctly. Today it appears to be bombing while compiling the zyd device, which wouldn't be so annoying if it weren't at the end of the alphabet... I'm going to clean up and start over from the beginning. If it happens again and I can't figure it out, I'll post my kernel configuration and the error message. There's nothing wrong with zyd. It just has the misfortune of being the last module to be built. I was building the kernel with nanobsd, which uses make -j3 by default. Running multiple make jobs at the same time like that made it impossible for me to see that the build was really crashing because I'd left something important out of my kernel configuration. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
DHCP server with no persistent storage
I'm running FreeBSD on a Soekris net4801. It boots from a read-only flash card, and has no permanent writable storage media - only memory disks. It runs several critical network services for me like DNS, and a firewall. One important service that it does not currently run is a DHCP server. My network has always been made up of a small number of machines with fixed IP addresses, but it's growing, and I'm feeling the need for DHCP. The handbook recommends the net/isc-dhcp3-server port, so that's what I'm looking at installing. I'm wondering what the implications of not having permanent writable storage will be for the DHCP service. Right now, without DHCP, if I pull the plug out of the wall then restart the box, the network comes right back up with no problems whatsoever. All I lost was some state tables and the DNS cache, which will be rebuilt automatically as needed. Will the DHCP server be this trouble-free if I switch my whole network to dynamic IPs? When the DHCP server goes offline, then comes back online, what happens? I'm hoping that the DHCP clients will renew their old leases based on the contents of their /var/db/dhclient.leases files, and that the server will comply with their wishes and repopulate DNS with their names when that happens. However I've read that the server keeps its own dhcpd.leases file. This file will disappear when I restart the server, because it will only exist on a memory disk. What will happen when a client says you gave me 192.168.1.5 but the server has no record of this in its dhcpd.leases file? I suppose a worse scenario would be if the DHCP clients did nothing until their leases expired. They'd be missing from the DNS table for awhile if that happened. Running around and rebooting every machine on the network just because the DHCP server went down for a minute is not something I want to have to do. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: time drift
On Thu, 15 May 2008, Christopher Cowart wrote: David Kelly wrote: Its PC commodity-grade. Not all that unusual even for stuff sold claiming to be a server. This is in no small part why ntpd exists. nptd calculates a correction coefficient and (under FreeBSD) stores it in /var/db/ntpd.drift for use on next start so as to more quickly establish a lock. So in short ntpd calibrates your clock in order to minimize the corrections required. Is The Right Thing To Do. We run a large number of FreeBSD servers under vmware. We've seen ntpd silently die, because the drift becomes insane. What do others do in this situation? (We've resorted to croning ntpdate for VMs.) kern.hz=100 in /boot/loader.conf solved this problem for me. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: netbeans 6.0.1 not run
On Sun, 11 May 2008, cuongvt wrote: Hi all! full explanation: FreeBSD 7.0-RELEASE (i386) uid=1001(mak) gid=0(wheel) groups=0(wheel),69(network) installed: jdk-1.6.0.3p4_1 diablo-jdk-1.5.0.07.01_10 javavmwrapper-2.3.2 I'm using zsh so I set JAVA_HOME and PATH of java into in my .zshrc. java -version: java 16 jdku3p1 etc etc Then I insralled netbeans from ports, firstly it warned: JAVA_HOME should not be defined. So i removed JAVA_HOME and PATH of java into in my .zshrc, reboot and reinstalled netbeans. Installation went fine. When I try run netbeans: [EMAIL PROTECTED]:~netbeans XIO: fatal IO error 0 (Unknown error: 0) on X server :0.0 after 0 requests (0 known processed) with 0 events remaining. This is exactly the same error message I get whenever I attempt to run any graphical java application on FreeBSD 7 using the diablo port. Between the diablo binaries and my Nvidia drivers, I'm relying on so many different compatibility layers and hacks that it's just too much. Installing jdk-1.6 and forcing my java applications to point to /usr/local/jdk1.6.0 instead of /usr/local/diablo* fixed the problem for me. So, I suspect that netbeans is still looking at the diablo java. Maybe there's a configuration option somewhere? If netbeans copies the java engine or contains its own, then I bet it's using diablo because that's the dependency that the port lists. I've not used netbeans on FreeBSD, so I don't know. Any ideas? P/S:All pkgs were installed via ports* -- View this message in context: http://www.nabble.com/netbeans-6.0.1-not-run-tp17179367p17179367.html Sent from the freebsd-questions mailing list archive at Nabble.com. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Firewalls
On Fri, 2 May 2008, Zane C.B. wrote: On Mon, 28 Apr 2008 20:50:06 +0100 Bruce Cran [EMAIL PROTECTED] wrote: Doug Hardie wrote: FreeBSD supports 3 firewalls: IPF, IPFW, and PF. Some time ago (perhaps years) I seem to recall some discussion that one or more of those was better maintained and higher quality than the others. I don't see any indications of this in the handbook. Several years ago I needed to do traffic shaping and used IPFW with dummynet. It worked but the need eventually went away. More recently I needed to incorporate spamd which defaults to PF so I used that. However, now I am back to needing traffic shaping again. I suspect trying to use both PF and IPFW simultaneously will not be a good approach. In addition, there now are instructions for using spamd with IPFW so it appears that either PF or IPFW will do what I need. Is there any additional information available to assist in selecting between those? Thanks. As I understand it pf is often found to be easiest to use and has lots of features like altq and os fingerprinting but is quite a bit slower than ipfw. There is one thing that IPFW has that PF does not that I have found to be very handy at times. It can be used to setup firewall rules that only affect a specific group or user. PF can do this too. There were threading/locking/crashing issues when last I tried to use that feature of PF back in FreeBSD 5.x, but that was a very long time ago. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: pf traffic shaping and perfomance
On Wed, 23 Apr 2008, Zbigniew Szalbot wrote: Hello, I would like to implement traffic shaping using pf. I know I need to recompile kernel to be able to achieve this but I have a more general question. I used to have pf with traffic shaping on a Pentium III 866 before and as soon as I activated it, the http response of the box was noticably slower. Here are the defs I used then: #altq on $ext_if cbq bandwidth 512Kb queue { def, smtp, udp, http, \ #ssh, icmp } #queue def bandwidth 13% cbq(default borrow red) #queue smtp bandwidth 25% cbq(borrow red) priority 7 #queue udp bandwidth 10% cbq(borrow red) #queue http bandwidth 40% cbq(borrow red) #queue ssh bandwidth 10% cbq(borrow red) ##{ ssh_interactive, ssh_bulk } ##queue ssh_interactive priority 7 #queue ssh_bulk priority 0 #queue icmp bandwidth 2% cbq It is quite possible that I misconfigured the shaping (as seen above). What would be suggested traffic shaping rules to allow smooth mail operation (smtp taking up to 40% of allowed bandwidth) and http responses? If that matters, uname -v FreeBSD 7.0-RELEASE #0 Many thanks in advance! I had the same problem with class-based queueing when I tried this. I suspect that the 512Kb in your initial queue definition is the limiting factor. I never did get it to work like I expected it to, however, so maybe I just don't understand it. Eventually I realized that I didn't actually want to chop up my bandwidth like this. What I really wanted to do was simply prioritize the traffic. The most important applications get first shot at the bandwidth, and the less important applications get choked when they need to be. I switched to priority queueing and I've been very happy with it. I'm sorry I can't help more with cbq, but unless you are able to make an accurate guess about how much bandwidth each class will really need to be using constantly, I think you'll find that you're reserving bandwidth unnecessarily. If your goal really is to cut down on your bandwidth usage, then please disregard this opinion. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: determining what's in the base system
On Tue, 15 Apr 2008, Chad Perrin wrote: I have two questions. First: Assume you have a FreeBSD system installed that has been running for at least a year, with a bunch of graphical desktop and productivity software installed, and have both installed and uninstalled a lot of software over that time. Now imagine that you want to know whether a given utility was something that came with the base system or was installed by some port or package later on. What's the easiest way to do that (preferably without installing the FreeBSD base system on a computer and checking whether the utility is present)? Second: Where can I get a list of all licenses on all software in the base system? I know there's at least the BSD License, the GPL, and the LGPL, but I'm a little hazy on what else is in there. I'm pretty sure there isn't any proprietary closed source software in there, but I wouldn't bet any substantial amount of money on it at this point, because I haven't really checked into it. For the first question, I'd first look at where the utility is. Base system utilities won't be in /usr/local. Add-on packages and ports should be. The pkg_info utility and all of its switches and options could be useful too, if your ports database is correct. For the second question, I've always assumed that /COPYRIGHT applied to everything in the base system. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Question regarding mail and dns server on Alix/Soekris?
On Mon, 7 Apr 2008, David Duong wrote: I'm planning to redoing my home network. I currently have one server (Opteron 170) that is currently a NAS, Email, and DNS server (btw, the main OS is FreeBSD). I was thinking of purchasing an Alix2c3/Soekris 5501 and use it as a Email + DNS server. Then dedicate my main server as a FreeBSD NFS server. My question is, has anyone installed a mail + DNS server on a ALIX/Soekris PC? If so, is it able to handle the load? I received a Soekris 4801 for Christmas 2005. I put FreeBSD 6 on it. It's my home network's gateway to the outside world, router, firewall (pf), dns server (bind), time server (ntpd), and socks proxy (nylon). I wanted this to be a highly reliable machine, so I opted not to install a hard drive. It boots from the compact flash card, mounted read-only so it won't wear out. I didn't want to trust my email or web content to a memory disk, so I've got those services running on another box. It's running sendmail just for nightly status reports, but that's probably not what you're interested in. It wasn't easy to set this up, but it was very rewarding. FreeBSD's diskless startup code was in a state of flux when I put this box together, but I expect it's a lot better now. I've been happy with it. I'm tempted to try upgrading it to FreeBSD 7 on some rainy weekend, and I may even install a DHCP server on it this time. I'm not sure what numbers you're interested in for determining if the box can handle the load. top registers no load, a mostly idle CPU, and mostly free memory. pfctl -s info registers between 800 and 1000 states and 255 searches per second when I'm saturating my connection with p2p traffic and using a bunch of complicated stateful firewall rules and priority queueing. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: FreeBSD Traffic Shaping
On Tue, 1 Apr 2008, [EMAIL PROTECTED] wrote: I am trying to limit the bandwidth available to some connections and I'm not sure FreeBSD can handle this. Maybe some of you can help. Here's what I need to have exactly. No matter what the number of connections, each connection should have at most/least 50kbps guaranteed outbound on port 80. I've tried dummynet but it doesn't do what I need because if I define a pipe with 1mbps and if I have 1000 connections, each connection will have less than 50kbps. Any way to do this in FreeBSD ? I can't think of any way to dynamically allocate a new pipe for each individual connection with any firewall software I've used. Have you considered getting your web server to do the limiting for you? I think mod_bandwidth for Apache is designed to do what you're asking, but I've never used it. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
How to replace libpthread correctly on 7.0?
This weekend I upgraded a desktop system from 6.2 to 7-STABLE using source. I removed and rebuilt all my ports using the ports system, and so far I haven't noticed any problems with them. When I attempted to rebuild some software outside the ports system, I ran into problems. I like to use the Code::Blocks IDE. It's in the FreeBSD ports system, but it's lagging about three years behind what's current, so I've been tracking the current code from the project's subversion system. It was working on FreeBSD 6.2. When I attempted to rebuild Code::Blocks, the build failed with an error relating to libpthread.so.2. I didn't get rid of old libraries when I upgraded to 7, so I still had this old libpthread.so.2 that hadn't been rebuilt. I know. Bad. I searched the web a bit and discovered that libpthread had fallen out of favor in 7 and was being replaced by something else. Ok, fine, so I got rid of my old broken libpthread.so.2. I tried to rebuild Code::Blocks again, and found that it was still attempting to link to libpthread. I threw away all the code and configurations, fetched the latest code, ran bootstrap, configure, and make and it's STILL running g++ with the -pthread switch and complaining about a missing libpthread.so.2. It finally dies with an undefined reference to [EMAIL PROTECTED]' So... Should I expect the new standard threading library (libthr I think?) to be a drop-in replacement for libpthread? If so, do you have any ideas where I need to look to convince this code to use it? For my personal projects, is the -pthread switch any good anymore, or is there a different switch I should be using now? Is libpthread even an option at all anymore? If so, how do I build it? I'm just not sure what direction I should start troubleshooting in now. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: How to replace libpthread correctly on 7.0?
On Sun, 30 Mar 2008, Luke Dean wrote: This weekend I upgraded a desktop system from 6.2 to 7-STABLE using source. I removed and rebuilt all my ports using the ports system, and so far I haven't noticed any problems with them. When I attempted to rebuild some software outside the ports system, I ran into problems. I like to use the Code::Blocks IDE. It's in the FreeBSD ports system, but it's lagging about three years behind what's current, so I've been tracking the current code from the project's subversion system. It was working on FreeBSD 6.2. When I attempted to rebuild Code::Blocks, the build failed with an error relating to libpthread.so.2. I didn't get rid of old libraries when I upgraded to 7, so I still had this old libpthread.so.2 that hadn't been rebuilt. I know. Bad. I searched the web a bit and discovered that libpthread had fallen out of favor in 7 and was being replaced by something else. Ok, fine, so I got rid of my old broken libpthread.so.2. I tried to rebuild Code::Blocks again, and found that it was still attempting to link to libpthread. I threw away all the code and configurations, fetched the latest code, ran bootstrap, configure, and make and it's STILL running g++ with the -pthread switch and complaining about a missing libpthread.so.2. It finally dies with an undefined reference to [EMAIL PROTECTED]' So... Should I expect the new standard threading library (libthr I think?) to be a drop-in replacement for libpthread? If so, do you have any ideas where I need to look to convince this code to use it? For my personal projects, is the -pthread switch any good anymore, or is there a different switch I should be using now? Is libpthread even an option at all anymore? If so, how do I build it? I'm just not sure what direction I should start troubleshooting in now. Never mind. It was my stupid mistake. I thought that make was making all new executables and libraries, but in fact it was pulling in an old library that was linked to libpthread. Silly me. If this had been in the ports system I would've run make deinstall first and avoided this problem. I removed the old library, which forced make to rebuild it and all was fine. The Porter's Handbook answered my question about the -pthread switch, I think. I have to go through this every now and then just to remind myself why I like the ports system so much. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
RAID mirror really worked
This isn't really a question, just a testimonial and a thank you to the people who make this work. I've got a Promise TX2300 SATA Raid controller that I use on my FreeBSD fileserver at home. It uses ataraid. I've set up a simple two-drive mirror. One of the drives has been sending me intermittent failure messages in the nightly emails for nearly a year - no more than one or two a month. Last night it finally croaked. The mirror broke, but it went into degraded mode and kept right on chugging along with no service interruption. Today I took down the system, replaced the bad drive, rebuilt the array using Promise's BIOS tools (since that's how I built it originally), and everything is back to normal. This was my first real drive failure with an inexpensive RAID card, and I'd always wondered if they could ease the pain of a hard drive failure. It looks like they can. I should add that before this, on another system, I had a RAID controller failure. A controller failure is a completely different scenario... ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: wxPython 2.8?
On Mon, 10 Sep 2007, Martin Tournoij wrote: On Sun 09 Sep 2007 21:09, Luke Dean wrote: On Mon, 10 Sep 2007, Martin Tournoij wrote: On Sun 09 Sep 2007 15:09, Luke Dean wrote: I've got a Python app that seems to require wxPython 2.8. The ports collection only has 2.6. Do you think it would it be worthwhile for me to download the source and attempt to compile it on my system or should I ask the port maintainer? wxgtk 2.8 is in the ports collection(x11/toolkits/wxgtk28), you're probably using an outdated snapshot of the ports collection. See the FreeBSD handbook on how to update your ports collection: http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/ports-using.html I guess I don't understand the relationship between wxPython and wxgtk. wxPython 2.6 builds with wxgtk 2.6 by default. If it works with wxgtk 2.8, maybe that's what I need. I'll give it a shot. Thanks. Ignore my previous post, it was late and I thought you couldn't find wxgtk2.8, I kind of missed the wxPython part ... :/ My fault, I'm sorry. I've cc-ed this email to the maintainer of py-wxPython26, I will look into creating a port for py-wxPython28 later today. Regards, Martin Tournoij Thank you! After I read a bit more about what wxPython really is, I realized that wxPython and wxgtk should be at the same version. In the spirit of open source, I attempted this myself. I installed wxgtk 2.8 from ports, then fetched the wxPython 2.8 source and attempted to port it following the example of the wxPython 2.6 makefile, but I got compillation errors early in the process in some gtk code. This made me think that perhaps the wxPython source I fetched might not be compatible with the wxgtk source I got from the ports system. I believe these versions are under development, so I suppose they change frequently and it may be difficult to get them to match up. Or maybe I really don't know what I'm doing. I've never attempted to port anything as complex as a toolkit before. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: wxPython 2.8?
On Mon, 10 Sep 2007, Neal Nelson wrote: On Mon, 2007-09-10 at 09:49 +0200, Martin Tournoij wrote: On Sun 09 Sep 2007 21:09, Luke Dean wrote: On Mon, 10 Sep 2007, Martin Tournoij wrote: On Sun 09 Sep 2007 15:09, Luke Dean wrote: I've got a Python app that seems to require wxPython 2.8. The ports collection only has 2.6. Do you think it would it be worthwhile for me to download the source and attempt to compile it on my system or should I ask the port maintainer? wxgtk 2.8 is in the ports collection(x11/toolkits/wxgtk28), you're probably using an outdated snapshot of the ports collection. See the FreeBSD handbook on how to update your ports collection: http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/ports-using.html I guess I don't understand the relationship between wxPython and wxgtk. wxPython 2.6 builds with wxgtk 2.6 by default. If it works with wxgtk 2.8, maybe that's what I need. I'll give it a shot. Thanks. Ignore my previous post, it was late and I thought you couldn't find wxgtk2.8, I kind of missed the wxPython part ... :/ My fault, I'm sorry. I've cc-ed this email to the maintainer of py-wxPython26, I will look into creating a port for py-wxPython28 later today. I have already submitted a port for wxPython 2.8 some weeks ago but for some reason it's languishing in the pr database. So if you want the port badly, hassle some committer to get it committed. For your reference the relevant PRs are: 115349 for the base port, 115350 for the common parts and 115351 for the unicode port. Regards, Neal. Ah, I didn't think to search for PRs. I should've looked there. I will try to check this out tonight and pass along any feedback I can give. Thank you for your work! ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
wxPython 2.8?
I've got a Python app that seems to require wxPython 2.8. The ports collection only has 2.6. Do you think it would it be worthwhile for me to download the source and attempt to compile it on my system or should I ask the port maintainer? ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: wxPython 2.8?
On Mon, 10 Sep 2007, Martin Tournoij wrote: On Sun 09 Sep 2007 15:09, Luke Dean wrote: I've got a Python app that seems to require wxPython 2.8. The ports collection only has 2.6. Do you think it would it be worthwhile for me to download the source and attempt to compile it on my system or should I ask the port maintainer? wxgtk 2.8 is in the ports collection(x11/toolkits/wxgtk28), you're probably using an outdated snapshot of the ports collection. See the FreeBSD handbook on how to update your ports collection: http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/ports-using.html I guess I don't understand the relationship between wxPython and wxgtk. wxPython 2.6 builds with wxgtk 2.6 by default. If it works with wxgtk 2.8, maybe that's what I need. I'll give it a shot. Thanks. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
kern.ipc.shm_allow_removed
The latest version of Opera claims to be faster by taking advantage of shared X memory if I set the sysctl kern.ipc.shm_allow_removed. I don't like to change sysctls from their default settings unless I understand the consequences. I've been unable to find a manpage that describes this setting. The description Enable/Disable attachment to attached segments marked for removal sounds a bit frightening to me. Would anyone care to comment on the effects and risks of changing this setting on a private desktop machine? ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: How to reset /dev/dsp ?
On Tue, 2 Jan 2007, Lowell Gilbert wrote: Luke Dean [EMAIL PROTECTED] writes: I'm running FreeBSD 6.1. My sound device shows up like this in my dmesg: pcm0: Intel ICH5 (82801EB) port 0xd800-0xd8ff,0xdc00-0xdc3f mem 0xfc001000-0xfc0011ff,0xfc002000-0xfc0020ff irq 17 at device 31.5 on pci0 pcm0: primary codec not ready! pcm0: Avance Logic ALC658 AC97 Codec My sound driver is compiled into the kernel: device sound device snd_ich I've got a java application that I run through diablo-jdk-1.5.0.07.01_1 that uses sound. It's a game. Partway through the game, the sound stops working. The people who make the game have been aware of the problem for many months, but don't understand what to do about it. Okay, I can accept that. What I can't accept is that this java application breaks the sound in such a way that NOTHING can play sound anymore until I reboot the machine! If I attempt to play a movie with mplayer after the game has broken the sound, it says: [AO OSS] audio_setup: Can't open audio device /dev/dsp: No such file or directory However, the dsp device still exists in /dev: [0:/dev ll dsp* crw-rw-rw- 1 root wheel - 0, 51 Dec 29 21:36 dsp0.0 crw-rw-rw- 1 root wheel - 0, 54 Dec 29 21:37 dsp0.1 crw-rw-rw- 1 root wheel - 0, 52 Dec 29 19:24 dspW0.0 crw-rw-rw- 1 root wheel - 0, 55 Dec 29 19:24 dspW0.1 crw-rw-rw- 1 root wheel - 0, 57 Dec 29 19:24 dspr0.1 The sndstat device doesn't show any problem, if I'm reading the output right: [0:/dev cat /dev/sndstat FreeBSD Audio Driver (newpcm) Installed devices: pcm0: Intel ICH5 (82801EB) at io 0xfc001000, 0xfc002000 irq 17 bufsz 16384 (1p/1r/0v channels duplex default) Is there anything I can do short of rebooting the machine to get my sound working when this happens? I thought maybe there was something I could do with devd or devctl to reset the device, but I can't figure out how to do that. I'm not even sure how to see the problem except to attempt to play a sound. Well, it's hard to say, because the hardware could be misbehaving, in which case the software may not know what's going on. It might be interesting to see whether fstat(1) sees anything holding the dsp devices. You could also try using vchans, which would (in theory) let you access the hardware from another device node after the first one hangs. fstat reveals that nothing is holding the dsp devices after the sound breaks. I read a bit about vchans, then set up a few with sysctl hw.snd.pcm0.vchans=4. I had to reboot first, because once the sound locks up, attempting to adjust this sysctl produces a device busy error. This produced some new dsp devices in /dev. I ran the java app again, broke the sound, and found that all of the vchans produce the same device busy error when I use mplayer switches to specify which vchan to use. (I was using commands like mplayer -ao oss:/dev/dsp0.1 blah.avi to test this.) Whatever this java app is doing to break the sound breaks it for all device nodes. Thanks for the idea anyway. I learned something. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
How to reset /dev/dsp ?
I'm running FreeBSD 6.1. My sound device shows up like this in my dmesg: pcm0: Intel ICH5 (82801EB) port 0xd800-0xd8ff,0xdc00-0xdc3f mem 0xfc001000-0xfc0011ff,0xfc002000-0xfc0020ff irq 17 at device 31.5 on pci0 pcm0: primary codec not ready! pcm0: Avance Logic ALC658 AC97 Codec My sound driver is compiled into the kernel: device sound device snd_ich I've got a java application that I run through diablo-jdk-1.5.0.07.01_1 that uses sound. It's a game. Partway through the game, the sound stops working. The people who make the game have been aware of the problem for many months, but don't understand what to do about it. Okay, I can accept that. What I can't accept is that this java application breaks the sound in such a way that NOTHING can play sound anymore until I reboot the machine! If I attempt to play a movie with mplayer after the game has broken the sound, it says: [AO OSS] audio_setup: Can't open audio device /dev/dsp: No such file or directory However, the dsp device still exists in /dev: [0:/dev ll dsp* crw-rw-rw- 1 root wheel - 0, 51 Dec 29 21:36 dsp0.0 crw-rw-rw- 1 root wheel - 0, 54 Dec 29 21:37 dsp0.1 crw-rw-rw- 1 root wheel - 0, 52 Dec 29 19:24 dspW0.0 crw-rw-rw- 1 root wheel - 0, 55 Dec 29 19:24 dspW0.1 crw-rw-rw- 1 root wheel - 0, 57 Dec 29 19:24 dspr0.1 The sndstat device doesn't show any problem, if I'm reading the output right: [0:/dev cat /dev/sndstat FreeBSD Audio Driver (newpcm) Installed devices: pcm0: Intel ICH5 (82801EB) at io 0xfc001000, 0xfc002000 irq 17 bufsz 16384 (1p/1r/0v channels duplex default) Is there anything I can do short of rebooting the machine to get my sound working when this happens? I thought maybe there was something I could do with devd or devctl to reset the device, but I can't figure out how to do that. I'm not even sure how to see the problem except to attempt to play a sound. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
detach/reattach remote GUI applications?
I run a headless server and I've become fond of using the screen utility with SSH to allow me to launch a text-based application, detach from it, and then reattach to it later to see how it's going. I'm wondering if there's a tool I can use that would allow me to do that with GUI applications. I know that I can tunnel X over SSH to have my remote headless server do the processing and my local machine do the display, but sometimes I have to switch off my local machine. This kills the display and thus kills the application. I might be able to figure out a way to set up a logical display on the remote machine, but then I wouldn't know how to see it. I'd have to tell the client application to switch Displays midstream wouldn't I? Can that be done? Windows Terminal Services (RDP) gives the capability I'm looking for by disconnecting and reconnecting a session. Surely there's some way to get this capability with X, and I just don't know how. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: detach/reattach remote GUI applications?
On Sun, 12 Nov 2006, Eric Schuele wrote: On 11/12/06 18:37, Luke Dean wrote: I run a headless server and I've become fond of using the screen utility with SSH to allow me to launch a text-based application, detach from it, and then reattach to it later to see how it's going. I'm wondering if there's a tool I can use that would allow me to do that with GUI applications. Could this be of use? http://www.tightvnc.com/ Yes! That's exactly the kind of thing I'm looking for. I used VNC with Windows years ago and it never occurred to me that it might be available on other platforms. Thanks for putting me on the right track. Luke ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: grub on FreeBSD
On Mon, 29 May 2006, Ask Bj?rn Hansen wrote: Hi, I am trying to use grub instead of the usual boot0 thing on a Compact Flash card I use in Soekris and PC Engines WRAP systems. I installed grub from ports/sysutils/grub and put the package on my nanobsd system on the CF card. Booting on a Soekris box and running grub, I get this: grub root (hd0,1) Filesystem type unknown, partition type 0xa5 grub root (hd0,1,a) Filesystem type unknown, partition type 0xa5 It seems like it can't read the ufs filesystem? Any ideas? Did you copy the stage1, stage2, and ufs2_stage1_5 files to /boot/grub on the CF card? As I understand it, grub needs these files to understand UFS2. Just a guess. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: 6.1_RELEASE Install Problem
On Wed, 10 May 2006, bc wrote: Ceri Davies wrote: On 10/5/06 05:03, bc [EMAIL PROTECTED] wrote: After make installworld I ran the final run thru of mergemaster and recived this error: ERROR CODE 64 FATAL ERROR: Cannot 'dc' into /usr/src/etc/ and install files to the temproot environment. Did you cut and paste that, or copy it in manually? Ceri Now that I am in front of the derver here is the full message (with a few lines before) pasted: install -o root -g wheel -m 640 /dev/null /var/tmp/temproot/var/log/sendmail.st install -o root -g wheel -m 644 freebsd.cf /var/tmp/temproot/etc/mail/sendmail.cf install -o root -g wheel -m 444 freebsd.submit.cf /var/tmp/temproot/etc/mail/submit.cf cd /usr/src/etc; install -o root -g wheel -m 644 /var/tmp/temproot/etc/ssh usage: install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode] [-o owner] file1 file2 install [-bCcpSsv] [-B suffix] [-f flags] [-g group] [-m mode] [-o owner] file1 ... fileN directory install -d [-v] [-g group] [-m mode] [-o owner] directory ... *** Error code 64 Stop in /usr/src/etc. *** FATAL ERROR: Cannot 'cd' to /usr/src/etc and install files to the temproot environment /usr/src Walnut SuperUser# install is being run with a blank space where the program name should be. That's what it's complaining about. It does not look like a directory problem or a permissions problem. I'd say that something is wrong with a makefile that builds ssh or with your make configuration. I'm afraid I don't know enough about the make process to say more than that, but perhaps if you can specify which platform you're building for and the contents of /etc/make.conf somebody more knowledgable might be able to help. Luke ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: find not finding file.
On Thu, 27 Apr 2006, RW wrote: Can anyone explain this: # find /etc/ -name 'named.conf' # ls -l /etc/namedb/named.conf -rw-r--r-- 1 root wheel 3715 Nov 13 17:35 /etc/namedb/named.conf named.conf is not found, but as a sanity check a similar search for ppp.conf succeeds # find /etc/ -name 'ppp.conf' /etc/ppp/ppp.conf there is nothing odd about the directories: # ls -dl /etc/namedb/ /etc/ppp/ drwxr-xr-x 5 root wheel 512 Nov 13 17:35 /etc/namedb/ drwxr-xr-x 2 root wheel 512 Feb 8 17:36 /etc/ppp/ Try this: $ ls -l /etc |grep namedb lrwxr-xr-x 1 root wheel 21 Dec 31 18:56 namedb - /var/named/etc/namedb namedb is actually a link, and find isn't following it. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Update from Ports removes dependency data
On Mon, 3 Apr 2006, [EMAIL PROTECTED] wrote: I recently noticed if I update a package from ports, the package manager looses the dependency data for the package. eg: updating old version of gettext # pkg_info -R gett* Information for gettext-0.14.5_1: Required by: libgpg-error-1.1 libgcrypt-1.2.2 gnutls-1.2.9 p5-gettext-1.03 gmake-3.80_2 [[[ update from ports - make ; make deinstall ; make reinstall ]]] # pkg_info -R gett* Information for gettext-0.14.5_2: Is my procedure incorrect, or is this normal behavior? -Thanks, Wayne This is normal. It's what makes upgrading those ports that are required by lots of other ports so difficult. Dependency information is version-specific. If you run pkg_info -r libgpg-error-1.1 you'll find that it still depends on gettext-0.14.5_1 until you rebuild libgpg-error. Last weekend I wrote some scripts to help me deal with stale dependencies like this. My code is such an ugly perl hack that I'm ashamed to post it, and my algorithm isn't really airtight, but I'll give you the algorithm I'm using. The script just uses pkg_info, pkg_version, and grep. 1) I build a dependency hash by parsing the output of pkg_info -ra. The keys to the hash are the names of the installed ports with the versions stripped off. The right-hand-side of the hash are the lists of dependencies with the version numbers stripped off. 2) I build what I call a full dependency hash from the dependency hash I built in step 1. I do this by recursively getting the dependencies for every dependency in the hash from step 1. The result is a hash just like the one from step one, only with the list of dependencies expanded to show all of the dependencies' dependencies, etc. all the way down the the ports that don't require anything. As I'm building this, I also count how many levels down I have to recurse to build each entry and remember this number for each port. I call this number the degree for the port. To avoid rebuilding the same ports over and over again, ports with high degrees should be built after ports with low degrees. 3) I get a list of everything I want to upgrade by parsing the output of pkg_version -qL =. That shows me which of my installed ports are not the same version as the port in the ports tree. Usually I'll want to upgrade those ports. 4) I add to that list all of the ports that have outdated dependencies. I get that list by parsing pkg_info -ra for each port and checking pkg_info -e for each dependency to see if it exists. If it doesn't exist, then I assume I need to upgrade the requiring port. Many times the only thing I gain by doing this is an update to the dependency information in the ports database, but sometimes (like with the recent libiconv upgrade) it's truly necessary. I don't know of any way to programmatically gauge the importance of an upgrade - all I can tell is whether the version of the port listed as a dependency still exists. 5) I take the list of ports that I built in step 3 and 4, look them up in the degrees hash I built at the end of step 2, sort the output by the degree I determined in step 2 and then print the list. Then I use that list to tell me which ports that I should consider rebuilding and in what order. I used this script to help me upgrade all the ports on a system that hadn't been upgraded in about six months. It worked pretty well. It's not always perfect, though. Dependencies can change dramatically between versions sometimes, and sometimes a port will depend on a port that doesn't exist in the database (like linux_base) but this approach is far far better than me trying to do a big upgrade of several ports by hand when I can't immediately see the dependencies. It also gives me the freedom/power to choose to build the ports I want to build, and to build them with the options I want to use. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: show port's deps
Hi, how can i know what the dependencies are for a particular port? try: # make pretty-print-run-depends-list - or - # make pretty-print-build-depends-list Where do you guys find these interesting and useful make targets? I've searched the manpages for make and make.conf and the Makefiles themselves, but I don't see anything like pretty-print-build-depends-list, yet it works. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Removable drives
On Tue, 28 Mar 2006, Christopher Sean Hilton wrote: I have a question to the community about removable drives, pendrives and usb and firewire attached hard drives. I'm just wondering how people are dealing with them in FreeBSD. I don't have any operational problems with them. I'm just wondering if I'm doing things the hard way. First Question: Which filesystem are people using on usb flash drives and removable hard drives? I'm using a mixture of ufs2, ext2, and msdos. I'm using ufs2 because I'm also using cfs to encrypt the contents and although I haven't tested this, I'm fairly certain cfs want's semantics that aren't in the msdos filesystem. I use msdosfs because I use my portable devices with MS Windows systems and digital cameras frequently, and I need compatibility more than anything else. Second Question: Are most people using vfs_usermount=1? I'm using the automounter. It's a little bit more work to setup but I'm using a laptop and since I've started to use the automounter the number of times that I've had to fsck my removable drive because I've suspended my laptop with a pendrive still attached and mounted has been reduced incredibly. I define the device to /etc/fstab with the noauto option, then explicitly mount and unmount the device as necessary. If I happen to need to mounst more than one of these devices at a time, I study the device numbers and read man pages until I remember how to mount something by its device name. So no, you're not doing things the hard way. :) ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Determining proper order to upgrade
I'm one of those masochists who insists on manually building all my pakages from the ports collection without using automating software like portmanager. Typically I upgrade my ports collection with cvsup -g -L 2 /etc/supfile-that-I-made-for-ports then run pkg_version -v -L '=' to get a list of what I need to upgrade. Then I pkg_info -o name-of-each-port to find out where the port lives in the ports collection, go there, and then use make with whatever custom options I need to upgrade it. The trouble I have is when one or more of those low-level ports that everything else depends on gets upgraded. I can spend DAYS building and rebuilding ports until everything gets built with the most current dependencies. I'm familiar with the -r and -R switches for pkg_info, and they are a huge help in this situation, but I'm wondering if anybody has written a script that will take a list of packages to be upgraded, examine their dependencies and what depends on them, and then return a sorted list of every installed package that should be rebuilt in order to keep the dependencies current. I don't want something to automate the whole process - just something to help me out with determining which ports are affected and what order they should be rebuilt in. Some of those port management packages must have something like this internally. I guess I could go look there. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Gripe about new dhclient
Ever since FreeBSD switched to OpenBSD's dhclient, I've had a serious problem. I'm running 6-STABLE as of earlier this week, but the problem has existed ever since we switched dhclients. Whenever I take the system offline long enough for the lease to expire, it will never get a new lease. On startup dhclient appears to broadcast requests for an IP address, but it never gets one. Once this situation occurs, the only way to get an IP address is to delete the leases file. I picked up this tip from someone on the list a few months ago, and it works. I even read that some people have gone so far as to create scripts that delete the leases file automatically somewhere in the boot process. I really don't want to do this. It's a hack and it defeats the purpose of having a leases file. This morning I spent an hour on the phone talking a nontechnical person through the process of deleting the leases file so we could get the server back online after it had been offline overnight, and I'm not going to do that again. I want to stick with the new dhclient program since it's the one that FreeBSD has adopted as its standard, but I can't live with this behavior. Does anyone have a good solution? my /etc/dhclient.conf is prepend domain-name-servers 127.0.0.1; my /etc/rc.conf contains ifconfig_vr0=DHCP and does not override any standard behavior for dhclient. This machine connects directly to a DSL modem, so there's no firewalls or NAT to interfere with getting an IP address. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Possible solution to lockups on FreeBSD 6?
In my ongoing quest to figure out why my machine keeps locking up since I upgraded from 5.4 to 6.0, I came across this interesting passage in man (5) pf.conf: BUGS Due to a lock order reversal (LOR) with the socket layer, the use of the group and user filter parameter in conjuction with a Giant-free netstack can result in a deadlock. If you have to use group or user you must set debug.mpsafenet to ``0'' from the loader(8), for the moment. This work- around will still produce the LOR, but Giant will protect from the dead- lock. In the release notes for FreeBSD 6, I found: Fine-grained locking has been applied to many of the data structures in the IPX/SPX protocol stack. While not fully MPSAFE at this point, it is generally safe to use IPX/SPX without the Giant lock (in other words, the debug.mpsafenet sysctl variable may be set to 1). While I don't fully understand what this passage is talking about, I take it to mean that the default value of debug.mpsafenet is now 1, and that this is a new feature of FreeBSD version 6. I use pf. I use the user filter parameter on some oft-evaluated rules. I discovered that my debug.mpsafenet is indeed 1 . My system is deadlocking occasionally under 6 where it never did so under 5.4. I think I'm going to attempt to set debug.mpsafenet to 0 in my loader.conf and see how many days I can go without a lockup this time. I've seen several threads lately from people having lockup troubles after upgrading to version 6, and at least one of them mentioned pf specifically, so I bet some of the rest of you may want to try this too. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Bacula vs. Amanda vs. whatever ...
On Thu, 29 Dec 2005, Kiffin Gish wrote: I have a home network with two FreeBSD servers (web- and file-server), a number of Windows desktops and a wireless FreeBSD laptop all connected to one another using Samba. What is the best tool to create automatic central backups? For now I just want to make backups on disk but later using an external tape drive. Some swear by Amanda, others insist Bacula works best with Samba, just curious is all. Amanda is the one that works with Samba. Bacula has its own client/server software. I have a similar home network. I used Bacula for over a year. I used the SQLite database for the catalog and backed up to a Raid array. I chose bacula because it allows me to have a single backup solution for both Windows and BSD machines, and the only thing I need to set up for communication is TCP/IP. I had some serious deadlocking problems with bacula recently, and was led to believe that those problems were corrected in the version that's currently in the ports collection. However I'm now having hardware problems (why did I buy a $20 RAID controller?) and am unable to make backups of any kind. I will probably give bacula another try after I fix my hardware problems. I have never tried Amanda. These are the only two solutions I know of for making automated backups of both Windows and BSD machines. I've read that bacula can make a backup that's complete enough to do a bare metal recovery on BSD. Thankfully I've never had the opportunity to test that myself. I don't believe either bacula or amanda can make a complete enough backup to do a bare metal recovery of a Windows machine because Windows locks so many critical system files while it's in operation. I wonder if a combination of Microsoft's backup utilities and Samba might accomplish this. There are several simple and good backup solutions for FreeBSD. It's those darn Windows machines that are so hard to back up. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Troubleshooting a lockup
I run a headless machine that has sporadic lockup problems, and I need some advice on what I can do to gather enough information to give me some idea of what's causing it. The machine acts as a router, DNS, web server, mail server, nfs server, firewall, and lots of other services. These problems started occurring after I upgraded from FreeBSD 5.4 to 6.0 and installed a secondary hard drive controller. Before that it ran perfectly for months at a time. That's a lot of variables to rule out. When the lockups occur, both network interfaces just plain die. Also, if I bring over a monitor and plug it in, I can't get a video signal, even if I tap the keyboard to wake it up. The lights on the keyboard still work, so I don't believe the box is completely frozen. The only option I have is to hit the reset button. Inspection of /var/log/messages never gives me any clues, except for one time I saw one message about my rl0 interface getting a watchdog timeout, but that was only one time and I can't imagine why a failure on one network interface would cause both network interfaces to stop responding. Inspection of the httpd logs just gives me an idea of about what time the lockup occurred, since there's no activity after that point. I don't know of any other log files that might be of assistance. I thought about trying to configure a dump device, but I don't believe the machine is panicing, except perhaps when I hit the reset button. I may try to figure out some way to disable the power management on the video, hook up a monitor, and leave top running on it to see if that gives me any clues. I plan on googling for serial terminal this afternoon. Any other suggestions? ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
kernel: rl0: watchdog timeout
Since I upgraded to version 6, my server has frozen under heavy network traffic three times. It ran happily for months on various versions of 5. Same hardware, same configuration. There was never anything in /var/log/messages until this last time. This time I found: kernel: rl0: watchdog timeout I don't know what it's trying to tell me, and I don't know if it's a cause or an effect. This is 6.0-STABLE as of late Sunday night. i386. rl0 is my WAN-side interface. vr0 is my internal interface. This machine runs DNS, email, web, NFS host, and several other services. It uses pf with altq. I don't think the system is panicking. It doesn't reboot. It may be freezing. It becomes unresponsive on both network interfaces. When the incident occurs, there's always a lot of simple TCP/IP traffic that's just passing through the box (not NFS, email, web, or anything the box provides on its own) and my first response is to hit the reset button to get it going again. What should I look for? The handbook says to check cables, and they seem fine. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: dhclient.leases grows forever?
On Mon, 14 Nov 2005, Luke Dean wrote: I've recently upgraded to version 6, and I'm using the new dhclient. It seems to be working happily, but I've noticed that /var/db/dhclient.leases.rl0 just keeps growing and growing, with new leases being added to the end. Leases that expired several hours ago remain in the file. According to the manpage for dhclient.leases, this database is supposed to contain only leases that are still valid, yet the expiration time on many of them has long since passed. Is there a bug, a misconfiguration, or am I responsible for setting up my own script for purging this thing? I can't find any cron-driven script that cleans up this file, so I'm wondering if dhclient is supposed to handle this itself, and if so, when? I haven't gotten any replies to this, but now that more time has passed I see that dhclient is cycling the contents of the leases file. There's about twenty leases in there all the time. Some of them have expired and some are planned many hours out into the future. I don't know when it cycles the leases or what the meaning of the far-future leases are, but it's working and cleaning up after itself somehow, so I'll shut up now. If somebody would like to explain how dhclient cycles through these leases though, I'd appreciate it. The manpages aren't enough to make me understand the process. I do have one practical reason for wanting to understand this. When I need to take the server down for maintenance, I like to know when my IP lease is due to expire. This lets me know how much time I can be offline before I lose my IP address. With the old dhclient that only had one lease in the file, the answer was obvious. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
dhclient.leases grows forever?
I've recently upgraded to version 6, and I'm using the new dhclient. It seems to be working happily, but I've noticed that /var/db/dhclient.leases.rl0 just keeps growing and growing, with new leases being added to the end. Leases that expired several hours ago remain in the file. According to the manpage for dhclient.leases, this database is supposed to contain only leases that are still valid, yet the expiration time on many of them has long since passed. Is there a bug, a misconfiguration, or am I responsible for setting up my own script for purging this thing? I can't find any cron-driven script that cleans up this file, so I'm wondering if dhclient is supposed to handle this itself, and if so, when? ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
illegal option in mergemaster
I synchronized my sources to RELENG_6 built the world, built the kernel, and installed the kernel. When I run mergemaster, it always dies with the following message. It's trying to use a -l option with cap_mkdb and that's not a legal switch. I can't find any documentation on that switch either. What is it trying to do, and why? cd /usr/src/etc; install -o root -g wheel -m 644 amd.map apmd.conf auth.conf crontab csh.cshrc csh.login csh.logout devd.conf devfs.conf dhclient.conf disktab fbtab ftpusers gettytab group hosts hosts.allow hosts.equiv hosts.lpd inetd.conf login.access login.conf mac.conf motd netconfig network.subr networks newsyslog.conf portsnap.conf pf.conf pf.os phones profile protocols rc rc.bsdextended rc.firewall rc.firewall6 rc.initdiskless rc.sendmail rc.shutdown rc.subr remote rpc services shells snmpd.config sysctl.conf syslog.conf usbd.conf etc.i386/ttys /usr/src/etc/../gnu/usr.bin/man/manpath/manpath.config /usr/src/etc/../usr.bin/mail/misc/mail.rc /usr/src/etc/../usr.bin/locate/locate/locate.rc printcap /var/tmp/temproot/etc; cap_mkdb -l /var/tmp/temproot/etc/login.conf; install -o root -g wheel -m 755 netstart pccard_ether rc.suspend rc.resume /var/tmp/temproot/etc; install -o root -g wheel -m 600 master.passwd nsmb.conf opieaccess /var/tmp/temproot/etc; pwd_mkdb -L -p -d /var/tmp/temproot/etc /var/tmp/temproot/etc/master.passwd cap_mkdb: illegal option -- l usage: cap_mkdb [-v] [-f outfile] file [file ...] *** Error code 1 Stop in /usr/src/etc. *** FATAL ERROR: Cannot 'cd' to /usr/src/etc and install files to the temproot environment ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: illegal option in mergemaster
On Thu, 10 Nov 2005, Luke Dean wrote: I synchronized my sources to RELENG_6 built the world, built the kernel, and installed the kernel. When I run mergemaster, it always dies with the following message. It's trying to use a -l option with cap_mkdb and that's not a legal switch. I can't find any documentation on that switch either. What is it trying to do, and why? cd /usr/src/etc; install -o root -g wheel -m 644 amd.map apmd.conf auth.conf crontab csh.cshrc csh.login csh.logout devd.conf devfs.conf dhclient.conf disktab fbtab ftpusers gettytab group hosts hosts.allow hosts.equiv hosts.lpd inetd.conf login.access login.conf mac.conf motd netconfig network.subr networks newsyslog.conf portsnap.conf pf.conf pf.os phones profile protocols rc rc.bsdextended rc.firewall rc.firewall6 rc.initdiskless rc.sendmail rc.shutdown rc.subr remote rpc services shells snmpd.config sysctl.conf syslog.conf usbd.conf etc.i386/ttys /usr/src/etc/../gnu/usr.bin/man/manpath/manpath.config /usr/src/etc/../usr.bin/mail/misc/mail.rc /usr/src/etc/../usr.bin/locate/locate/locate.rc printcap /var/tmp/temproot/etc; cap_mkdb -l /var/tmp/temproot/etc/login.conf; install -o root -g wheel -m 755 netstart pccard_ether rc.suspend rc.resume /var/tmp/temproot/etc; install -o root -g wheel -m 600 master.passwd nsmb.conf opieaccess /var/tmp/temproot/etc; pwd_mkdb -L -p -d /var/tmp/temproot/etc /var/tmp/temproot/etc/master.passwd cap_mkdb: illegal option -- l usage: cap_mkdb [-v] [-f outfile] file [file ...] *** Error code 1 Stop in /usr/src/etc. *** FATAL ERROR: Cannot 'cd' to /usr/src/etc and install files to the temproot environment Ah, nevermind. It's a new option and I couldn't use it until I installed the world. I should've guessed. Sorry for the noise. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: higher resolution console screen?
On Wed, 21 Sep 2005, bob self wrote: 5.4-STABLE FreeBSD 5.4-STABLE #0: Tue Sep 20 15:33:58 EDT 2005 I have a 19 LCD monitor and usually run in console mode. My screen resolution is 80x25, but I'd like to get something like 120 x 60 if possible. I've played with vidcontrol, but so far can't get it to do anything useful. Also my console screen is now shifted to the left one character, but when I run startx my gui screen is shifted to the right about 1 character. My card is a nvidia card with 128 megs or ram and I have enabled vesa in the kernel and have vesa_load=YES in /boot/loader.conf. However, I get this error in dmesg: module_register: module vesa already exists! Module vesa failed to register: 17 How can I fix these problems? About the left/right shifting.. I also have a 19 LCD monitor, Nvidia card (a GeForce FX 5500), vesa enabled in the kernel, and I used to get a shifting problem like you describe. I switched from Xorg's nv driver to Nvidia's linux driver (ports/x11/nvidia-driver) and the shifting problem went away. Before I switched, resyncing the monitor would correct the problem. Mine is a dual-boot machine that runs WinXP and FreeBSD 5.4, and since switching to Nvidia's driver made the problem go away, I reasoned that the nv driver and the Nvidia driver for Windows were confusing the monitor by using slightly different refresh rates for the same screen mode or something, and that's why I had to resync it every time I rebooted from one OS to the other. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: /boot on a separate partition
On Mon, 18 Jul 2005, Ross Kendall Axe wrote: I am currently trying to get to grips with FreeBSD and am trying it out on an old Pentium machine. However, the machine's BIOS can't seem to read past 504MB, so I want to place the /boot directory in a small 25MB partition at the start of the drive. Setting up the partition with sysinstall is easy enough, but does anyone have any suggestions of how to diddle the bootloader to accept this configuration? I don't particularly want to go for the standard 'small / partition and separate partitions for /usr, /var, /home...' since I only have a 1GB drive to play with and judging the partition sizes down the nearest KB would be... tricky. I have performed this procedure before (many, many times) on Linux using both LILO and GRUB, but I can't seem to get my head around the FreeBSD bootloader. All I would expect you have to do is use FDISK to make two partitions, remembering to mark the first one as bootable. Then use disklabel to create your slices. Make a /boot slice on the first partition, then make a / slice and a swap slice on the second partition. That should be all that's required for what you're trying to do. A little over a year ago, I had to split up a drive to solve the same problem you're having, but I went the small / route instead, so you might be running into a problem I didn't have. Luke Dean ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: FreeBSD 5.4 / linux-opera
On Thu, 23 Jun 2005, Thomas Hill wrote: However, I was previously using the web browser Opera under Linux emulation because it supports lots of the browser plugins But now when installing linux-opera from the ports collection it complains about libX11.so.6, which I know is a common X11 library. Oh yes, and it doesn't run at all now. Just wondering, is anyone using linux-opera 8.02 on FreeBSD 5.4 and experiencing the same problem. 8.02? The latest thing I can find in the ports collection is linux-opera-8.01.20050615_2 It runs just great for me; much better than 8.0 did. libX11.so.6 is a standard X library, but linux-opera needs the linux version of the library. It sounds like you may need to work on getting your linux compatibility working. That's the trickiest part of the process. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Enabling acpi_thermal
After having a router literally melt during a period of record-high temperatures last week, I've become interested in remotely monitoring the temperature of some of my systems. I've got two FreeBSD systems running 5.4. Both use acpi, and several acpi sysctl variables are present, but the hw.acpi.thermal family is missing. At least one of these machines has an internal thermometer because the BIOS setup screen can show a temperature. I don't suppose that means this thermometer has to be visible to the operating system, but I'm hoping that it is. I've read the acpi and acpi_thermal manpages. /boot/loader.conf is empty, so I'm just getting the defaults from /boot/defaults/loader.conf. Is there some simple switch that I can flip to make the hw.acip.thermal sysctls appear or should I start down the path suggested in http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/acpi-debug.html for Fixing your ASL? Also, if these acpi sysctls won't appear, is there any hope that any of the temperature-monitoring applications in the ports collection will work? ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Enabling acpi_thermal
On Tue, 31 May 2005, Dan Nelson wrote: In the last episode (May 31), Luke Dean said: After having a router literally melt during a period of record-high temperatures last week, I've become interested in remotely monitoring the temperature of some of my systems. I've got two FreeBSD systems running 5.4. Both use acpi, and several acpi sysctl variables are present, but the hw.acpi.thermal family is missing. You'll probably have better luck by installing xmbmon and either setting up an exec line in snmpd.conf that runs mbmon -T1 -c1, or writing a little cgi that runs mbmon, and polling that remotely. Thanks. That seems to work. Now I just have to read the manual to make sure I understand what it's telling me. I had assumed that acpi was the only way to get to the thermometer, but I see now that it isn't. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Hardware RAID Cards..
On Tue, 24 May 2005, Kirk Strauser wrote: On Tuesday 24 May 2005 09:57, Aaron C. Meadows wrote: My question is, since that chipset is unsupported for hardware RAID, would I be better off to software RAID them, or get a different RAID card? What RAID level do you plan on using? Mirroring shouldn't use much CPU, for example, but parity might put a bit of a load on a hard-working system. That's a good question, though. Several cards are listed in the hardware compatibility notes, but they stop short of saying this card is completely supported or stay away from this one. What cards have people had good luck with in practice? I've been using a Promise FastTrak S150 TX2/plus for close to a couple of years now. It supports two parallel and two serial ATA drives. I bought it to support my parallel ATA drives and then I thought I'd migrate to SATA, but I haven't done so yet. I've got two parallel drives in a RAID1 (mirrored) array. This configuration is discouraged by the manufacturer because the drives have to share a cable and failure on one drive will very likely lock up the system, but that's not really important to me. I'm more worried about hardware failure than uninterrupted uptime. I've been using this setup since FreeBSD version 5.2, and I'm currently running 5.4. The dmesg looks like: atapci0: Promise PDC20371 SATA150 controller port 0x9800-0x987f,0x9400-0x940f,0x9000-0x903f mem 0xfb00-0xfb01,0xfb027000-0xfb027fff i q 22 at device 2.0 on pci2 atapci0: failed: rid 0x20 is memory, requested 4 That little failure at the end has always been there in one form or another. It doesn't seem to hinder operation as far as I can tell though. I've only had to use the built-in maintenance utilities once to fix something, and that was after a really bad kernel upgrading accident. It worked fine. Overall I'm happy with this card. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: CVS tag for specific release candidate
On Fri, 6 May 2005, Kris Kennaway wrote: On Fri, May 06, 2005 at 08:55:56PM -0700, Luke Dean wrote: I was using 5.4-RC1 until today. Today I attempted to update to 5.4, but I can't get into single-user mode without a panic. I'm just about certain that this is because of the work in progress in that ata-raid driver right now. I'm using a Promise FastTrack S150 TX2Plus, and it's not happy with the recent changes. I'd be glad to help test the changes, but right now I'm more concerned with getting my system in a stable shape. I want to go back to 5.4-RC1, but I can't seem to get the CVS tag right. Release candidates are not tagged. You can use a specific date with the -D option to roll back to before they were committed. This is also more helpful because you can isolate the specific change that caused you problems. Unfortunately by now it might be too late to get this fixed for 5.4-RELEASE, but at least it can be fixed soon afterwards if there is a problem. It's not a problem with the ata stuff after all. It's definitely something else. I've managed to get 5.4-RELEASE installed, but I absolutely cannot get into single-user mode. I've tried selecting option 4 from the boot menu and just running a shutdown now from multiuser mode. When it gets to the point of asking what shell I want to run, usually the system just freezes. I can hit scroll lock to go back and read the dmesg, but there's no way I'm going to get a shell prompt. I can't type. Alternately, sometimes I'll get a page fault panic as soon as I hit a key - this is what usually happens if I'm trying to get into single-user mode via shutdown. If it doesn't panic, I can press ctrl-alt-delete (this is on i386 by the way) and I'll get a message about how it can't write to /var/db/mixer0-state or write other things to /var... which isn't surprising since I couldn't get to a shell prompt to mount anything in the first place. I don't understand why it's trying to mess with /var at all. I think something about my configuration must be fouled up. I don't think it's 5.4. Sorry for the confusion. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
CVS tag for specific release candidate
I was using 5.4-RC1 until today. Today I attempted to update to 5.4, but I can't get into single-user mode without a panic. I'm just about certain that this is because of the work in progress in that ata-raid driver right now. I'm using a Promise FastTrack S150 TX2Plus, and it's not happy with the recent changes. I'd be glad to help test the changes, but right now I'm more concerned with getting my system in a stable shape. I want to go back to 5.4-RC1, but I can't seem to get the CVS tag right. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: CVS tag for specific release candidate
On Fri, 6 May 2005, Luke Dean wrote: I was using 5.4-RC1 until today. Today I attempted to update to 5.4, but I can't get into single-user mode without a panic. I'm just about certain that this is because of the work in progress in that ata-raid driver right now. I'm using a Promise FastTrack S150 TX2Plus, and it's not happy with the recent changes. I'd be glad to help test the changes, but right now I'm more concerned with getting my system in a stable shape. I want to go back to 5.4-RC1, but I can't seem to get the CVS tag right. Ah, nevermind. I just found out I can combine a tag and a date together, like: *default release=cvs tag=RELENG_5_4 date=2005.03.28.00.00.00 ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: FreeBSD Installation Horror
I downloaded the three floppy images for 5.3-RELEASE and dd'ed them on the disks. Then I booted the installation and tried to partition my hard drive. To my surprise, the partition table shown by the installation was complete nonsense. I figured it probably had something to do with the fact that my BIOS doesn't support the disk size. I'm using the OnTrack disk manager to fix the problem for Windows. So I booted from the disk, and used the OnTrack feature to boot from a floppy after OnTrack has been loaded. The partition table was exactly the same junk, though. I also tried different geometries (reported by LILO, BIOS, FreeBSD installation, etc.), but this didn't change the view of the partition table either. I was able to install FreeBSD 5.2 on a machine of this generation that didn't support large partitions either, but it wasn't easy. Windows-based workarounds like OnTrack won't work. The trick I used was to make a primary bootable partition on the hard drive that fit within the size limitations that the BIOS could natively understand. (I don't have that number in front of me right now, I'm sorry.) Put the parts of FreeBSD that you need to boot in this partition and boot from it. Once FreeBSD boots, it's able to support large partitions that your old BIOS can't understand, so you can mount the rest of your hard drive, no matter how large it is. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Clock running fast
On Wed, 4 May 2005, Ryan Winograd wrote: Charles Swiger wrote: On May 4, 2005, at 3:49 PM, Ryan Winograd wrote: I recently noticed that the system clock on a machine i recently set up is running very quickly, about 2x realtime by my measuring. What can i do to solve/investigate this problem? What information would be helpful? Try changing the kern.timecounter.hardware sysctl; you can look at the available choices via: sysctl kern.timecounter.choice Thanks for all the advice everyone. The solution was changing the kern.timecounter.hardware sysctl to i8254 (was ACPI-safe). I was using NTP, but when the clock is at 2x even having cron run ntp every minute is too innacurate. Thx again! Ryan I wish I'd known about that sysctl when I had this problem on my last system! I tried ntp, but having ntp constantly resetting the clock just added new problems. Thanks for sharing the outcome with us. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Prevent DHCP from changing resolv.conf
I have a minor problem regarding my network configuration, specifically that the external interface on my router gets it's IP via DHCP from the ISP, so in rc.conf ifconfig_xl0=DHCP is set. This leads to the single entry in resolv.conf that I want to be there, namely nameserver 127.0.0.1 being replaced with my ISP's nameservers, which in return makes resolving of LAN IPs or even localhost via the installed BIND difficult for the machine. I don't want dhclient to change the resolv.conf. I checked the man pages for resolv.conf, rc.conf and dhclient but couldn't find anything there relating to my problem. Thanks, Fabian In dhclient.conf, put: prepend domain-name-servers 127.0.0.1; This keeps the DNS that DHCP gives you, but it puts your local one first so you'll be able to find your local domains. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: problems with pf
I have read the manpage on pf but I am still stumped. I get some error messages when starting up that say something like 'rule expands to no possible valid combination' or something to that effect. If someone can tell me how I can find out what the error messages are when I boot I will post them. Parse your rules without actually loading them: pfctl -n -f /etc/pf.conf Flush the old rules and load the new ones without rebooting: pfctl -F -f /etc/pf.conf Both are must-haves for tinkering with pf. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: problems with pf
Parse your rules without actually loading them: pfctl -n -f /etc/pf.conf Flush the old rules and load the new ones without rebooting: pfctl -F -f /etc/pf.conf Gah! I knew that didn't look right... pfctl -F all -f /etc/pf.conf Or check out man pfctl for other options. Both are must-haves for tinkering with pf. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Reusing a port after a crash
From time to time, my torrent filesharing application will crash or need to be killed. The application is configured to listen on a specific port. If I try to restart the application after improper termination, I receive a fatal error message stating that the port is already in use. Before restarting the application, I verify that the application did indeed shut down (via 'ps -aux') and 'sockstat -l' shows that no application is using the port in question, so I don't understand how the port could still be 'in use'. The only way I've found to restart the application after improper termination is to reboot the whole system. Is there something else I could try? Does this kind of thing tend to happen to network applications when they crash or could this be specific to my application? What little tinkering I've done in the socket world reminds me of this: -- % man setsockopt ... SO_REUSEADDR enables local address reuse SO_REUSEPORT enables duplicate address and port bindings ... SO_REUSEADDR indicates that the rules used in validating addresses sup- plied in a bind(2) call should allow reuse of local addresses. SO_REUSEPORT allows completely duplicate bindings by multiple processes if they all set SO_REUSEPORT before binding the port. This option per- mits multiple instances of a program to each receive UDP/IP multicast or broadcast datagrams destined for the bound port. -- Basically, what I remember about this is that if you supply those options when creating the socket and your app crashes, you can start it back up using the same host:port otherwise you get the message you're getting... I'm sure a good networking would be able to explain it better... Thanks! That's what I needed to know. I didn't know such options existed. I poked around in the code for the app, and even though I know next to nothing about python I found that there's a reuse parameter on the call that binds the listening port, and it's defaulting to false. I bet if I explicitly set it to true that I'll get the behavior I want (or break the whole thing...) Open Source is fun. :) Luke Dean ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]