Re: Error Compile Kernel
On Sat, Mar 17, 2007 at 11:35:47AM +0700, Toan. Bach Quang Bao wrote: ip_input.o(.text+0x200): In function `ip_init': ../../../netinet/ip_input.c:312: undefined reference to `nf_sockopt_init' You forgot to mention what version of FreeBSD you are trying to compile, but I can't find any remotely similar function call in that file (or in the entire kernel) in either 6.x or 7.x. Are you sure this isn't a local modification you made? Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: FreeBSD 2.2.9 / Installation problem
On Thu, Mar 15, 2007 at 12:28:12PM +0100, Christian Walther wrote: On 15/03/07, Nino Ivanov [EMAIL PROTECTED] wrote: Dear Sir or Madam, [...] The target machine is a HP Omnibook 600c laptop with 8MB RAM, a 486 DX4 processor, 300 MB disk space. No CD drive. No network available. External floppy drive without DMA. (I tried NetBSD, but because of the lack of DMA it did not work properly.) The functioning of the floppy drive is critical, being the machine's only practical means of communicating with the outer world. Due to cost and time considerations, no upgrades are possible. If the target machine is not suitable for an installation of FreeBSD, please let me know so I stop further attempts. I guess you're without luck in this case. AFAIK FreeBSD needs at least 64 MB RAM to work happily. I tried installing it on an P1/133MHz Laptop with 16MB RAM, and it freezes after a few minutes. And it's dead slow. Well it is only true of more modern versions that they do not function well on systems with e.g. 8MB. FreeBSD 2.x was happy with as little as 4MB. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: FreeBSD 2.2.9 / Installation problem
On Thu, Mar 15, 2007 at 08:43:32PM +0100, Nino Ivanov wrote: Dear Chistian, Dear Kris, I also think the RAM will not be the issue, as it is a text-only install, and indeed, I am not planning to get fancy. I completely don't need X. Midnight Commander is perfectly fine as a working environment. 4.11 seemed OK. But I am having a different problem right now, which I am still researching: It does not recognize the device from where to mount root correctly. I mean the following: When I put FreeBSD into the Compaq for installation, the harddrive is ad4 or ad8. But in the system where I want to run it, the HP Omnibook, it is ad0. Now, when I start it back in the HP Omnibook, it says that swap is not configured correctly on ad8s-something. Which is true, it should look for it on ad0... I have only once been able till now to mount root. (And this is my basis for assuming that even 4.11 CAN potentially run.) I said as command ufs:/dev/ad0 when it asked me where to mount root from. This worked, however, e.g. ufs:/dev/ad0s1 did not work. I am thinking that I might have made a mistake, and should have said ad0s1a. Yet, the principal new problem persists: FreeBSD does not realize that it should now look at ad0 instead of ad4 or ad8. (However, in the booting process, it correctly sees ad0 as having 325 MB etc.) Is there a way to solve this? Probably the /etc/fstab is wrong and refers to the ad4 or ad8 devices. The root should indeed typically be ufs:/dev/ad0s1a. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Should I Upgrade 5.4 - 6.2?
On Thu, Mar 15, 2007 at 04:47:06PM -0400, [EMAIL PROTECTED] wrote: First off, I want to thank the people who responded to my thread Stability Issues on a 5.4-RELEASE box a couple of weeks ago; after disabling hyperthreading, getting a clean run of Memtest back, and doing some serious fsck'ing of the disks, the box appears to now be completely stable. I'm still not sure which of the above fixed the problem...but I'll take a stable system at this point. :-) That said, in that thread I had asked about the advisability of upgrading to 6.2, and it was intelligently pointed out that doing so in pursuit of stability was a bad idea. Now that the box is stable, though, I'm back to the same question: should I make the upgrade, and if so, how should I do it? My primary driver for doing so would be to keep current enough that I'm still getting security and other patches on a regular basis, and that I can upgrade my applications from ports as necessary. If this is not an issue, then my only remaining concern would be that it's usually easier to get support on lists like this if you're running a modern version of the OS (that's certainly the case with the OpenBSD folks). My primary concern with upgrading is that the box is in Portland, OR, and I'm in Arlington, VA...and while the ISP is friendly, I doubt that I could count on them for major system recovery if I botch something during the upgrade. My other worry is that I don't want to break existing apps if possible (the main one I'm concerned about is Zope/Plone). This is a production box with moderate traffic, and it would be a problem if there was extensive downtime. Is it worth upgrading? If so, what's the best way to do so -- CVSup, or some other way? Are there any major caveats if I do choose to upgrade (or choose to stay with the existing OS)? On general grounds it is well worth running 6.2 over 5.x - depending on your workload you should see performance improvements, and support for 6.2 is much better than for the legacy 5.x branch. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Should I Upgrade 5.4 - 6.2?
On Thu, Mar 15, 2007 at 05:09:57PM -0400, Jerry McAllister wrote: On Thu, Mar 15, 2007 at 04:47:06PM -0400, [EMAIL PROTECTED] wrote: First off, I want to thank the people who responded to my thread Stability Issues on a 5.4-RELEASE box a couple of weeks ago; after disabling hyperthreading, getting a clean run of Memtest back, and doing some serious fsck'ing of the disks, the box appears to now be completely stable. I'm still not sure which of the above fixed the problem...but I'll take a stable system at this point. :-) That said, in that thread I had asked about the advisability of upgrading to 6.2, and it was intelligently pointed out that doing so in pursuit of stability was a bad idea. Now that the box is stable, though, I'm back to the same question: should I make the upgrade, and if so, how should I do it? My primary driver for doing so would be to keep current enough that I'm still getting security and other patches on a regular basis, and that I can upgrade my applications from ports as necessary. If this is not an issue, then my only remaining concern would be that it's usually easier to get support on lists like this if you're running a modern version of the OS (that's certainly the case with the OpenBSD folks). My primary concern with upgrading is that the box is in Portland, OR, and I'm in Arlington, VA...and while the ISP is friendly, I doubt that I could count on them for major system recovery if I botch something during the upgrade. My other worry is that I don't want to break existing apps if possible (the main one I'm concerned about is Zope/Plone). This is a production box with moderate traffic, and it would be a problem if there was extensive downtime. Is it worth upgrading? If so, what's the best way to do so -- CVSup, or some other way? Are there any major caveats if I do choose to upgrade (or choose to stay with the existing OS)? You should if you can reasonably do it, for the reasons you give plus improvements in performance and in some utilities. My sentiment is usually to do a clean install over major version numbers. It tends to leave less dross laying around. but I do not have to worry about down times very much, a couple of hours at night is not terribly noticable in my stuff. It does require more time down to do a clean from scratch install. But, I think you can get away with a cvsup upgrade from 5.4 to 6.2. Then your downtime is just the reboot and stuff at single user (mergemaster), plus probably some for upgrading various ports. Yes, a source upgrade from 5.x to 6.x (followed by portupgrade -fa) isn't too bad. As with any upgrade you do need a recovery strategy though. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Should I Upgrade 5.4 - 6.2?
On Thu, Mar 15, 2007 at 08:46:45PM -0700, Garrett Cooper wrote: Kris Kennaway wrote: On Thu, Mar 15, 2007 at 05:09:57PM -0400, Jerry McAllister wrote: On Thu, Mar 15, 2007 at 04:47:06PM -0400, [EMAIL PROTECTED] wrote: First off, I want to thank the people who responded to my thread Stability Issues on a 5.4-RELEASE box a couple of weeks ago; after disabling hyperthreading, getting a clean run of Memtest back, and doing some serious fsck'ing of the disks, the box appears to now be completely stable. I'm still not sure which of the above fixed the problem...but I'll take a stable system at this point. :-) That said, in that thread I had asked about the advisability of upgrading to 6.2, and it was intelligently pointed out that doing so in pursuit of stability was a bad idea. Now that the box is stable, though, I'm back to the same question: should I make the upgrade, and if so, how should I do it? My primary driver for doing so would be to keep current enough that I'm still getting security and other patches on a regular basis, and that I can upgrade my applications from ports as necessary. If this is not an issue, then my only remaining concern would be that it's usually easier to get support on lists like this if you're running a modern version of the OS (that's certainly the case with the OpenBSD folks). My primary concern with upgrading is that the box is in Portland, OR, and I'm in Arlington, VA...and while the ISP is friendly, I doubt that I could count on them for major system recovery if I botch something during the upgrade. My other worry is that I don't want to break existing apps if possible (the main one I'm concerned about is Zope/Plone). This is a production box with moderate traffic, and it would be a problem if there was extensive downtime. Is it worth upgrading? If so, what's the best way to do so -- CVSup, or some other way? Are there any major caveats if I do choose to upgrade (or choose to stay with the existing OS)? You should if you can reasonably do it, for the reasons you give plus improvements in performance and in some utilities. My sentiment is usually to do a clean install over major version numbers. It tends to leave less dross laying around. but I do not have to worry about down times very much, a couple of hours at night is not terribly noticable in my stuff. It does require more time down to do a clean from scratch install. But, I think you can get away with a cvsup upgrade from 5.4 to 6.2. Then your downtime is just the reboot and stuff at single user (mergemaster), plus probably some for upgrading various ports. Yes, a source upgrade from 5.x to 6.x (followed by portupgrade -fa) isn't too bad. As with any upgrade you do need a recovery strategy though. Kris I agree with both Kris and Jerry. Besides, if you run 6.2 you're running a supported version of FreeBSD whereas 5.4 isn't supported anymore (5.5 is the last supported version in the legacy 5.x branch). Plus there are slight improvements from 5.x to 6.x. s/slight/major/ ;) Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Memory leak and deep swap upon the restart?
On Thu, Mar 15, 2007 at 06:55:24AM +0300, Abdullah Ibn Hamad Al-Marri wrote: Hello, I have a webmail server, has apache 2.2.4, mysql 5.0.33, php 5.2.1, clamav, mailscanner ..etc. The weird issue it goes into deep swap when it starts or I restart it. *sigh* This happened since like 6 months I don't know why? it was okay before that. Your running processes are trying to use more memory than is present in your system, so swapping is the only possibility. Look at the VSZ column (virtual size) to see how overloaded your system is. USER PID %CPU %MEM VSZ RSS TT STAT STARTED TIME COMMAND vscan 449 0.0 0.7 45840 3520 ?? Ss3:45AM 0:00.48 amavisd (master) (perl5.8.8) root 474 0.0 10.9 61788 56864 ?? Ss3:45AM 0:02.95 /usr/local/bin/spamd -c -Q -d -r /var/run/spamd/spamd.pid (perl5.8 vscan 482 0.0 0.0 46472 0 ?? IW - 0:00.00 amavisd (virgin child) (perl5.8.8) vscan 483 0.0 0.0 46472 0 ?? IW - 0:00.00 amavisd (virgin child) (perl5.8.8) vscan 484 0.0 0.0 46472 0 ?? IW - 0:00.00 amavisd (virgin child) (perl5.8.8) vscan 485 0.0 0.0 46472 0 ?? IW - 0:00.00 amavisd (virgin child) (perl5.8.8) vscan 486 0.0 0.0 46472 0 ?? IW - 0:00.00 amavisd (virgin child) (perl5.8.8) vscan 487 0.0 0.0 46472 0 ?? IW - 0:00.00 amavisd (virgin child) (perl5.8.8) vscan 488 0.0 0.0 46472 0 ?? IW - 0:00.00 amavisd (virgin child) (perl5.8.8) vscan 489 0.0 0.0 46472 0 ?? IW - 0:00.00 amavisd (virgin child) (perl5.8.8) root 646 0.0 0.7 61788 3428 ?? I 3:45AM 0:00.01 spamd child (perl5.8.8) root 647 0.0 0.3 61788 1424 ?? I 3:45AM 0:00.00 spamd child (perl5.8.8) postfix649 0.0 0.2 18916 1180 ?? Is3:45AM 0:00.01 MailScanner: master waiting for children, sleeping (perl5.8.8) postfix650 0.0 0.5 79972 2464 ?? S 3:45AM 0:03.10 MailScanner: waiting for messages (perl5.8.8) root 667 0.0 1.3 49788 6732 ?? Ss3:45AM 0:00.18 /usr/local/sbin/httpd www742 0.0 0.0 49864 0 ?? IW - 0:00.00 /usr/local/sbin/httpd www763 0.0 1.3 50704 6980 ?? I 3:45AM 0:00.02 /usr/local/sbin/httpd www765 0.0 2.0 52184 10140 ?? I 3:45AM 0:00.31 /usr/local/sbin/httpd www766 0.0 0.7 49856 3476 ?? I 3:45AM 0:00.01 /usr/local/sbin/httpd www767 0.0 0.7 49856 3476 ?? I 3:45AM 0:00.00 /usr/local/sbin/httpd www768 0.0 0.7 49856 3476 ?? I 3:45AM 0:00.01 /usr/local/sbin/httpd www769 0.0 2.1 50836 10788 ?? I 3:45AM 0:00.07 /usr/local/sbin/httpd www770 0.0 2.3 50936 12172 ?? I 3:45AM 0:00.21 /usr/local/sbin/httpd www771 0.0 0.0 49816 0 ?? IW - 0:00.00 /usr/local/sbin/httpd www772 0.0 0.0 49816 0 ?? IW - 0:00.00 /usr/local/sbin/httpd www773 0.0 0.0 49816 0 ?? IW - 0:00.00 /usr/local/sbin/httpd postfix774 0.0 0.3 79972 1600 ?? S 3:45AM 0:03.05 MailScanner: waiting for messages (perl5.8.8) postfix775 0.0 0.3 79972 1792 ?? S 3:45AM 0:03.05 MailScanner: waiting for messages (perl5.8.8) postfix776 0.0 0.3 79972 1376 ?? S 3:45AM 0:03.05 MailScanner: waiting for messages (perl5.8.8) postfix777 0.0 14.4 79972 74812 ?? S 3:45AM 0:03.15 MailScanner: waiting for messages (perl5.8.8) mysql 518 0.0 3.7 51508 19184 con- I 3:45AM 0:00.23 /usr/local/libexec/mysqld --defaults-extra-file=/var/db/mysql/my.c Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: dmesg and GIANT-LOCK
On Tue, Mar 13, 2007 at 01:06:12AM -0700, James Long wrote: With regard to the recent thread about looking for GIANT-LOCKs in dmesg, why would one system say: ns : 00:56:29 /home/james uname -v ;dmesg | grep fxp FreeBSD 6.2-STABLE #0: Tue Feb 20 15:47:09 PST 2007 fxp0: Intel 82559 Pro/100 Ethernet port 0x2400-0x243f mem 0xc4fff000-0xc4ff,0xc4e0-0xc4ef irq 10 at device 2.0 on pci0 miibus0: MII bus on fxp0 fxp0: Ethernet address: 00:02:a5:0a:57:73 while a more recent build says: t30 : 00:56:19 /home/james uname -v ;dmesg | grep fxp FreeBSD 6.2-STABLE #2: Thu Mar 8 08:23:11 PST 2007 fxp0: Intel 82801CAM (ICH3) Pro/100 VE Ethernet port 0x7400-0x743f mem 0xd020-0xd0200fff irq 11 at device 8.0 on pci2 miibus0: MII bus on fxp0 fxp0: Ethernet address: 00:09:6b:86:82:a6 fxp0: [GIANT-LOCKED] fxp is not giant locked, you can check the source for the INTR_MPSAFE flag in sys/dev/fxp/if_fxp.c. I'm not sure how you are seeing this, please describe the configuration of this system further (kernel config, loader.conf). Kris pgpkzr1YXlNlZ.pgp Description: PGP signature
Re: 6.2 running 4.x binaries (missing libc.so.3)
On Tue, Mar 13, 2007 at 10:27:27AM -0500, Robert Huff wrote: Sergio Lenzi writes: However, the 4.x binaries want to find libc.so.3, which doesn't seem to be anywhere around. Anyone know the trick to getting this to work? /usr/ports/misc/compat4x ? or as a workaround add in the /etc/libmap.conf libc.so.3libc.so libm.so.3 libm.so .. that is every shared library it complains add it to the /etc/libmap.conf That can work. On the other hand, there's no guarantee the ABI - never mind the internal operation - of any function will remain constant across a major version bump. (As far as I know.) In fact in FreeBSD a version bump is almost always a guarantee that they are incompatible and you will break some applications by doing this. Kris pgpP5h4rGhyeY.pgp Description: PGP signature
Re: 6.2 running 4.x binaries (missing libc.so.3)
On Tue, Mar 13, 2007 at 08:32:49AM -0400, Thomas David Rivers wrote: I've been trying to run 4.x binaries on FreeBSD 6.2; with the compat4x package (and/or port) installed. However, the 4.x binaries want to find libc.so.3, which doesn't seem to be anywhere around. Anyone know the trick to getting this to work? libc.so.3 was FreeBSD 3.x, not 4.x. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: system does not come up after reboot due to devices in /dev missing freebsd v3.2
On Tue, Mar 13, 2007 at 01:07:37PM -0400, David Glassman wrote: We have a freebsd 3.2 system and had to reboot it. Afterwards it could not fine the following: swapon: /dev/da0s1b: No such file or directory swapon: /dev/da2s1b: No such file or directory Automatic reboot in progress... Can't stat /dev/da0s1a: No such file or directory Can't stat /dev/da0s1a: No such file or directory /dev/da0s1a: CAN'T CHECK FILE SYSTEM. /dev/da0s1a: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY. it seems that the above devices disappeared in /dev (See fstab file below)and goes into maintenance mode. _How do I recover the devices from my fstab file?? Why would they suddenly disappear if there was no disk failure?? Only due to disk corruption or explicit removal. In old versions of FreeBSD you create devices by hand using the /dev/MAKEDEV script. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: 6.2 running 4.x binaries (missing libc.so.3)
On Tue, Mar 13, 2007 at 11:59:21AM -0700, Eric P. Scott wrote: [Kris Kennaway [EMAIL PROTECTED]] libc.so.3 was FreeBSD 3.x, not 4.x. And misc/compat3x is marked FORBIDDEN. Yep, that's easily overridden, but something the OP will have to evaluate for himself. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: dmesg and GIANT-LOCK
On Tue, Mar 13, 2007 at 12:12:26PM -0700, James Long wrote: On Tue, Mar 13, 2007 at 01:50:34PM -0400, Kris Kennaway wrote: On Tue, Mar 13, 2007 at 01:06:12AM -0700, James Long wrote: With regard to the recent thread about looking for GIANT-LOCKs in dmesg, why would one system say: ns : 00:56:29 /home/james uname -v ;dmesg | grep fxp FreeBSD 6.2-STABLE #0: Tue Feb 20 15:47:09 PST 2007 fxp0: Intel 82559 Pro/100 Ethernet port 0x2400-0x243f mem 0xc4fff000-0xc4ff,0xc4e0-0xc4ef irq 10 at device 2.0 on pci0 miibus0: MII bus on fxp0 fxp0: Ethernet address: 00:02:a5:0a:57:73 while a more recent build says: t30 : 00:56:19 /home/james uname -v ;dmesg | grep fxp FreeBSD 6.2-STABLE #2: Thu Mar 8 08:23:11 PST 2007 fxp0: Intel 82801CAM (ICH3) Pro/100 VE Ethernet port 0x7400-0x743f mem 0xd020-0xd0200fff irq 11 at device 8.0 on pci2 miibus0: MII bus on fxp0 fxp0: Ethernet address: 00:09:6b:86:82:a6 fxp0: [GIANT-LOCKED] fxp is not giant locked, you can check the source for the INTR_MPSAFE flag in sys/dev/fxp/if_fxp.c. I'm not sure how you are seeing this, please describe the configuration of this system further (kernel config, loader.conf). Kris It just dawned on me when you said INTR_MPSAFE, would having options IPSEC in the kernel config cause fxp to use GIANT? dmesg says in part: FreeBSD 6.2-STABLE #2: Thu Mar 8 08:23:11 PST 2007 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/T30 WARNING: debug.mpsafenet forced to 0 as ipsec requires Giant WARNING: MPSAFE network stack disabled, expect reduced performance. Yes. Use FAST_IPSEC instead, it's also faster in other ways than just having better SMP scaling properties. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Exim 4.66 Causing Kernel Panics?
On Tue, Mar 13, 2007 at 03:39:24PM -0700, Don O'Neil wrote: Anyone aware of a reason why a fresh build/install of exim 4.66 would cause kernel panics and reboots on my FreeBSD 6.1 machine? It shouldn't, of course. Please follow up with the panic in the usual way (developers handbook, PR, etc) Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Tool for validating sender address as spam-fighting technique?
On Sun, Mar 11, 2007 at 12:41:48PM -0600, Chad Leigh -- Shire.Net LLC wrote: On Mar 11, 2007, at 6:31 AM, Justin Mason wrote: for what it's worth, I would suggest *not* adopting this as an anti-spam technique. Sender-address verification is _bad_ as an anti-spam technique, in my opinion. Basically, there's one obvious response for spammers looking to evade it -- use real sender addresses. Where's an easy place to find real addresses? On the list of target addresses they're spamming! This is a red-herring. They already do that. They have been doing that for a long time. And it has nothing to do with sender verification. Sender verification works and works well. I hate sender verification because it forces me (the sender) to jump through hoops just for the privilege of sending email to you. I send a lot of courtesy emails to e.g. port maintainers who have problems with their ports, and when I encounter someone with such a system I usually don't bother following up (their port just gets marked broken in the usual way, and they can follow up on it on their own if they want to). Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Tool for validating sender address as spam-fighting technique?
On Sun, Mar 11, 2007 at 01:43:22PM -0600, Chad Leigh -- Shire.Net LLC wrote: On Mar 11, 2007, at 1:36 PM, Kris Kennaway wrote: On Sun, Mar 11, 2007 at 12:41:48PM -0600, Chad Leigh -- Shire.Net LLC wrote: On Mar 11, 2007, at 6:31 AM, Justin Mason wrote: for what it's worth, I would suggest *not* adopting this as an anti-spam technique. Sender-address verification is _bad_ as an anti-spam technique, in my opinion. Basically, there's one obvious response for spammers looking to evade it -- use real sender addresses. Where's an easy place to find real addresses? On the list of target addresses they're spamming! This is a red-herring. They already do that. They have been doing that for a long time. And it has nothing to do with sender verification. Sender verification works and works well. I hate sender verification because it forces me (the sender) to jump through hoops just for the privilege of sending email to you. No, it forces you to set up a correct RFC abiding system I send a lot of courtesy emails to e.g. port maintainers who have problems with their ports, and when I encounter someone with such a system I usually don't bother following up (their port just gets marked broken in the usual way, and they can follow up on it on their own if they want to). If your system is following the RFCs then you should have no problems. YOU should fix your broken system. Sending emails without a valid from address is disconsiderate. Why should I accept a mail from an account that violates the RFCs about accepting DSN back? Perhaps we are talking about different things, I am talking about systems which send me an email back requiring me to do steps a, b or c in order to complete delivery of the email. kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Howmany CPU Does FreeBSD Support ?
On Sun, Mar 11, 2007 at 01:18:12AM +0800, David Schulz wrote: I have heard it does not scale well above 4 It all depends on your workload. FreeBSD 7.0 will have good scaling on 8 or more CPUs on common workloads, see e.g.: http://people.freebsd.org/~kris/scaling/mysql.html Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Howmany CPU Does FreeBSD Support ?
On Sat, Mar 10, 2007 at 10:40:47PM +0100, Wojciech Puchar wrote: I have heard it does not scale well above 4 to be clear. kernel task (disk I/O, network etc.) is always on first processor, everything else on any CPU. This is incorrect for approximately the last 7 years (it is only true for FreeBSD 4.x and below). Kris pgpWafo8UQ1NP.pgp Description: PGP signature
Re: Max CPUs in SMP
On Sun, Mar 11, 2007 at 01:01:31PM +0530, Susanth K wrote: Which UNIX flavour supports the MAX CPS in SMP ? This isn't really a meaningful question. Do you really have a system with e.g. 1024 CPUs that you need to run an OS on? If not, what hardware are you really asking about? Kris pgpBLNCJqfLL8.pgp Description: PGP signature
Re: Package missing
On Fri, Mar 09, 2007 at 11:09:04AM +0100, Olivier Vimont wrote: The pdflib binary package which is needed by gnuplot-4.0 is missing on the AMD64 port of the 6.2 release in th ftp site Thanks if you can restore it This software may not be freely redestributed according to the license imposed by the software authors. Please talk to them about your concerns. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Restore /libexec/ld-elf.so.1
On Mon, Mar 05, 2007 at 08:33:33AM -0500, Huy Ton That wrote: I renamed this file and now I cannot boot up anymore... How can I restore this file? Help See the archives, I answered this question a few days ago ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED] ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Stability Issues on 5.4-RELEASE Box
On Wed, Feb 28, 2007 at 05:39:48PM -0500, [EMAIL PROTECTED] wrote: I'm not running your application mix, but I've never seen random reboots unless there were hardware issues. With 5.X these included having hyperthreading turned on, which I know caused problems with my dual XEON system. Hmmm...two answers so far, two people saying hyperthreading can be an issue. I'll definitely have that turned off ASAP. HTT isn't expected to cause problems at least on modern versions of FreeBSD (I dont remember if old versions like 5.4 have a bug), but it could be if you are running older hardware with broken BIOS support. * Issues with files that are not found on startup sometimes, but are other times. Prime example: the Zope CMS system that's been installed failed to find libmysqlclient.so after a planned soft reboot, but found it with no trouble on a subsequent boot a few minutes later, with no config changes in between. Haven't seen that; are there any messages indicating you're having filesystem problems? Thanks for asking; I see some new nasties in /var/log/messages: Feb 27 09:05:37 www fsck: /dev/ad4s1f: PARTIALLY TRUNCATED INODE I=9397392 Feb 27 09:05:37 www fsck: /dev/ad4s1f: UNEXPECTED SOFT UPDATE INCONSISTENCY; RUN fsck MANUALLY. You definitely need to drop to single-user mode and fsck -f: filesystem corruption can cause many problems. * Given my dmesg below, do you see any specific problems? The interrupt storm on uhci1+ is not a good thing. Any thoughts on how to fix it? You can disable USB support if you do not need it, otherwise a fix will probably involve an upgrade to a newer version as a first step. Kris pgpBT0Fmvzg3S.pgp Description: PGP signature
Re: /libexec/ld-elf.so.1
On Wed, Feb 28, 2007 at 01:37:56PM -0800, Noah wrote: so something strange has happened and my machine no longer boots and I am not clear why. Here is what happens during boot: Trying to mount root from ufs:/dev/ad0s1a WARNING: / was not properly dismounted ELF interpreter /libexec/ld-elf.so.1 not found Enter full pathname of shell or RETURN for /bin/sh: ELF interpreter /libexec/ld-elf.so.1 not found Enter full pathname of shell or RETURN for /bin/sh: sh FEnter full pathname of shell or RETURN for /bin/sh: /bin/sh ELF interpreter /libexec/ld-elf.so.1 not found Enter full pathname of shell or RETURN for /bin/sh: ELF interpreter /libexec/ld-elf.so.1 not found Enter full pathname of shell or RETURN for /bin/sh: ELF interpreter /libexec/ld-elf.so.1 not found Enter full pathname of shell or RETURN for /bin/sh: ELF interpreter /libexec/ld-elf.so.1 not found Enter full pathname of shell or RETURN for /bin/sh: what is the best suggestion for troubleshooting this? That's a pretty serious error, it indicates your system has lost the ability to run any dynamically linked binaries (i.e. almost all of them, by default) because the dynamic linker was removed somehow. The cause of this could either be accidental misuse of rm or similar, filesystem corruption, disk failure, etc. Try to work out what you or the other admins were doing prior to this failure. To repair, you can boot -s and use the statically linked tools in /rescue to try and investigate the cause and possible fix. One thing that might work is that if you have done an installworld on this machine in the past then you might have a useable backup /libexec/ld-elf.so.1.old which you could copy into place. If not, and you can't find a way to get a copy of this file onto the machine, then your remaining alternative would be a reinstall. Kris pgpKTDXCP15yO.pgp Description: PGP signature
Re: mysql50-server on FreeBSD 6.2 w/ LINUX_THREADS?
On Tue, Feb 27, 2007 at 10:47:07AM -0800, patrick wrote: Is it still advisable to build the mysql50-server on FreeBSD 6.2 using the LINUX_THREADS option? I'm using the SMP kernel on an older dual 1.0GHz Pentium III. This page http://wiki.freebsd.org//MySQL suggests that the libthr library in FreeBSD 6.x is optimized for MySQL and perhaps better than using linuxthreads. Any thoughts? I think lunixthreads is no longer needed, but try it and see, if you're interested. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: kernel debug messages
On Sun, Feb 25, 2007 at 04:23:10PM +0100, Andrea Venturoli wrote: In the meanwhile I have seen this in the log: Feb 24 17:51:53 soth kernel: lock order reversal: Feb 24 17:51:53 soth kernel: 1st 0xc6a37090 inp (divinp) @ /usr/src/sys/netinet/ip_divert.c:336 Feb 24 17:51:53 soth kernel: 2nd 0xc070a18c tcp (tcp) @ /usr/src/sys/netinet/ip_fw2.c:1982 Feb 24 17:51:53 soth kernel: KDB: stack backtrace: Any comments? Divert sockets are known to have lock order reversals which can cause deadlocks under certain situations. I think this particular one has not been reported, so you might like to send it to [EMAIL PROTECTED] for inclusion in his list (http://sources.zabbadoz.net/freebsd/lor.html). You might like to ping the net@ list to see if anyone is working on fixing these problems. Kris pgpZUBXBFB8AG.pgp Description: PGP signature
Re: 64-bits platform question
On Wed, Feb 21, 2007 at 02:35:58PM -0500, Aard Nerd wrote: Hi list, as far as I know Intel 64 architecture (formerly known as Intel Extended Memory 64 Technology, or Intel EM64T) enables 64-bit computing on desktop when combined with supporting software. If I am right, 64-bit computing (on Intel architecture) requires a computer system with a processor, chipset, BIOS, operating system, device drivers and applications enabled for Intel EM64T architecture. So I bought an ASUS P4P800-VM with a 3.0GHz processor that supports Intel EM64T and 1Gb of Infineon PC3200 RAM memory. The system is ok...so why I can't install BSD 64 bits with my system ??? Because you inserted your windows 95 CD instead??? Seriously, give us a hint here about what you did and what went wrong :) Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Is it possible to build sun-jdk-1.5 from the source?
On Mon, Feb 19, 2007 at 08:09:41PM +0100, Daniel Tourde wrote: Hello, On Mon, Feb 19, 2007 at 07:20:45PM +0100, Daniel Tourde wrote: Hello, Is it possible to build sun-jdk-1.5 from the source code? If yes, how to do that? Yes, use the port. To be honest, I would like to avoid to install a binary that requires Linux support packages. It's not that I do not like Linux (I have 3 Gentoo Linux at my disposal), it's simply that my machine is a FreeBSD machine and I would like to keep it as simple as possible. Unfortunately java is written in java so you will have to install a binary compiler (either linux or freebsd) first to bootstrap. OK. Where could I find the FreeBSD binary? Are you referring to: http://www.freebsdfoundation.org/downloads/java.shtml ? Or the equivalent diablo port. Then, would the binary be replaced by the one compiled or would they be in parallel on my system? They'd both be there. Daniel ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED] ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: make in ports not working
On Tue, Feb 20, 2007 at 11:57:19AM +1100, Brad Kowalczyk wrote:
Kris Kennaway wrote:
On Tue, Feb 20, 2007 at 10:50:35AM +1100, Brad Kowalczyk wrote:
Hi
I get the following error when trying to make from the ports tree:
/usr/ports/Mk/bsd.port.mk, line 2292: warning: String comparison
operator should be either == or !=
/usr/ports/Mk/bsd.port.mk, line 2292: warning: String comparison
operator should be either == or !=
/usr/ports/Mk/bsd.port.mk, line 2292: Malformed conditional
(((${OSVERSION} 504105 || (${OSVERSION} = 60 ${OSVERSION}
600103) || (${OSVERSION} = 70 ${OSVERSION} 700012))
${PKGORIGIN} != ports-mgmt/pkg_install) ||
exists(${LOCALBASE}/sbin/pkg_info))
/usr/ports/Mk/bsd.port.mk, line 2293: warning: String comparison
operator should be either == or !=
/usr/ports/Mk/bsd.port.mk, line 2293: warning: String comparison
operator should be either == or !=
/usr/ports/Mk/bsd.port.mk, line 2293: Malformed conditional
((${OSVERSION} 504105 || (${OSVERSION} = 60 ${OSVERSION}
600103) || (${OSVERSION} = 70 ${OSVERSION} 700012))
${PKGORIGIN} != ports-mgmt/pkg_install)
/usr/ports/Mk/bsd.port.mk, line 2308: if-less else
/usr/ports/Mk/bsd.port.mk, line 2308: Need an operator
/usr/ports/Mk/bsd.port.mk, line 2322: if-less endif
/usr/ports/Mk/bsd.port.mk, line 2322: Need an operator
/usr/ports/Mk/bsd.port.mk, line 5987: if-less endif
/usr/ports/Mk/bsd.port.mk, line 5987: Need an operator
make: fatal errors encountered -- cannot continue
Any help appreciated
Also I am not subscribed to the list so reply-all.
You forgot to mention which version of FreeBSD you're running, but
it's probably an ancient one that is no longer supported.
Kris
yup ok its an ancient one:
su-2.05a# uname -v
FreeBSD 4.8-RELEASE : Tue Apr 8 11:35:48 GMT 2003
I am stuck with it (its a virtual server) so is there any way i can get
ports working? I just love 'portinstall xyz'!
FreeBSD 4 is no longer supported in any way by ports, update to 6.2 or
make do with out-of-date ports that will never again be updated (you
can use the RELEASE_4_EOL tag until you can upgrade the box).
Kris
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Is it possible to build sun-jdk-1.5 from the source?
On Mon, Feb 19, 2007 at 07:20:45PM +0100, Daniel Tourde wrote: Hello, Is it possible to build sun-jdk-1.5 from the source code? If yes, how to do that? Yes, use the port. To be honest, I would like to avoid to install a binary that requires Linux support packages. It's not that I do not like Linux (I have 3 Gentoo Linux at my disposal), it's simply that my machine is a FreeBSD machine and I would like to keep it as simple as possible. Unfortunately java is written in java so you will have to install a binary compiler (either linux or freebsd) first to bootstrap. kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Compiler Flags for SPARC64
On Mon, Feb 19, 2007 at 02:14:04PM +0100, Christian Baer wrote: Hello everybody out there! Please excuse my posting this question again on this list, but the last post on the freebsd-sparc64 didn't help much. There isn't really much traffic on that list. Assuming that gcc when run on sparc64 produces v7 code (for sun4/4c) by default, I went about trying to improve that as v7 code is known to be a fair bit slower as v9 (sun4u) code. The improvement can be as much as 100% for some apps like OpenSSL or OpenSSH. I went about trying some Compiler flags. -mcpu=ultrasparc and -mcpu=v9 both came into mind. However this lead to several problems of programs not compiling anymore. Most notably was the failure of 'make buildworld'. When gcc is told to produce v9 code, it doesn't produce 64bit code (you have to set -m64 for that), it just uses a few additional commands the CPU knows, which should make the resulting code faster but no longer compatible with older CPUs (non-UltraSPARC). This means that there shouldn't be any problem with pointers that are now strange to the code. But even if I explicitly set the -m32 flag, I still can't make the world. I discussed this in a German newsgroup, where someone told me that the CPU is set to v9 by default on FreeBSD, as it only supports SPARC64 and not SPARC32. Although this assumption makes sense, I couldn't find any evidence to back it up. While some compiler flags are set by default on some platforms for optimization for that particular CPU, there doesn't seem to be anything set for sparc64. Additionaly, if the mcpu were really set to ultrasparc or v9, then setting it again shouldn't cause buildworld to stop with the error I don't know what platform this is. Has anyone got any ideas on how to go on with this? You'll have to look at the compiler spec and how it is bootstrapped. FWIW, I don't think there are any secret flags you can set to improve the compiler targetting, as the defaults are already appropriate. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: make in ports not working
On Tue, Feb 20, 2007 at 10:50:35AM +1100, Brad Kowalczyk wrote:
Hi
I get the following error when trying to make from the ports tree:
/usr/ports/Mk/bsd.port.mk, line 2292: warning: String comparison
operator should be either == or !=
/usr/ports/Mk/bsd.port.mk, line 2292: warning: String comparison
operator should be either == or !=
/usr/ports/Mk/bsd.port.mk, line 2292: Malformed conditional
(((${OSVERSION} 504105 || (${OSVERSION} = 60 ${OSVERSION}
600103) || (${OSVERSION} = 70 ${OSVERSION} 700012))
${PKGORIGIN} != ports-mgmt/pkg_install) ||
exists(${LOCALBASE}/sbin/pkg_info))
/usr/ports/Mk/bsd.port.mk, line 2293: warning: String comparison
operator should be either == or !=
/usr/ports/Mk/bsd.port.mk, line 2293: warning: String comparison
operator should be either == or !=
/usr/ports/Mk/bsd.port.mk, line 2293: Malformed conditional
((${OSVERSION} 504105 || (${OSVERSION} = 60 ${OSVERSION}
600103) || (${OSVERSION} = 70 ${OSVERSION} 700012))
${PKGORIGIN} != ports-mgmt/pkg_install)
/usr/ports/Mk/bsd.port.mk, line 2308: if-less else
/usr/ports/Mk/bsd.port.mk, line 2308: Need an operator
/usr/ports/Mk/bsd.port.mk, line 2322: if-less endif
/usr/ports/Mk/bsd.port.mk, line 2322: Need an operator
/usr/ports/Mk/bsd.port.mk, line 5987: if-less endif
/usr/ports/Mk/bsd.port.mk, line 5987: Need an operator
make: fatal errors encountered -- cannot continue
Any help appreciated
Also I am not subscribed to the list so reply-all.
You forgot to mention which version of FreeBSD you're running, but
it's probably an ancient one that is no longer supported.
Kris
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: pkg_version: not found
On Sun, Feb 18, 2007 at 09:27:38AM -0600, Jeffrey Goldberg wrote: On Feb 18, 2007, at 4:49 AM, Richard Collyer wrote: I'm guessing a sym link from /usr/local/sbin/ to the correct location would be ok? I don't see any reason why it wouldn't be ok. But keep in mind that I've been using FreeBSD for less than two weeks. It wouldn't be OK because it's trying to use the /usr/local/sbin version for a reason: the base system version as shipped in old releases is inadequate. Kris pgpDgR6G6ck8b.pgp Description: PGP signature
Re: pkg_version: not found
On Sun, Feb 18, 2007 at 10:49:11AM +, Richard Collyer wrote: Jeffrey Goldberg wrote: On Feb 17, 2007, at 6:16 PM, Richard Collyer wrote: there is no pkg_version in that dir but pkg_version -v works. Odd no? Try which pkg_version to find out the path of the one that is working. Also try whereis pkg_version Thanks, [EMAIL PROTECTED]:~] $ which pkg_version /usr/sbin/pkg_version [EMAIL PROTECTED]:~] $ whereis pkg_version pkg_version: /usr/sbin/pkg_version /usr/share/man/man1/pkg_version.1.gz I'm guessing a sym link from /usr/local/sbin/ to the correct location would be ok? Didn't realise that 5.4 was eol just yet. Might look at doing a upgrade to 6 this week. It's not yet entirely EOL, but support for anything but the latest release on the stable branch is never guaranteed so in practise, older releases do not work as well. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: LKM Trojan?
On Sun, Feb 18, 2007 at 11:04:18PM +0900, FreeBSD MailingLists wrote: When I run chkrootkit I get the following lines. Checking `lkm'... You have 107 process hidden for readdir command chkproc: Warning: Possible LKM Trojan installed rkhunter doesn't seem to find anything. I suspect that my machine might be compromised. running ls in the /proc directory returns an empty list. I have recompiled the kernel and world but the problem persists. Any suggestions on how to fix this without having to reinstall from scratch? When using any tool you need to understand the limitations of that tool. One of the major limitations of this kind of pattern recognition security tool is that they just aren't very accurate, and have lots of false positives. So you may have a LKM trojan (even though FreeBSD doesn't use LKMs, it uses KLDs ;), or (more likely) you might have just encountered a poorly specified search pattern in the tool. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Error compiling/upgrading Azureus
On Mon, Feb 19, 2007 at 04:29:41PM +1000, Warren Liddell wrote: Running FreeBSD 6.2-STABLE .. ran CVSUP Today so all my ports/src is all up to date. No matter what i try i cant get the new version of Azureus to compile .. any ideas/suggestions would be appreciated. Yes, it seems to be broken; talk to the maintainer. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: pkg_version: not found
On Sun, Feb 18, 2007 at 12:16:38AM +, Richard Collyer wrote: Hello, Having a small issue with a 5.4 box. When I do a make on a port I get a list of /usr/local/sbin/pkg_version: not found come up. Varies by the port. Surely enough there is no pkg_version in that dir but pkg_version -v works. Odd no? There are other pkg_info etc in there so I am assuming that the file has become deleted / corrupt? Anyone know this package installs these tools there so I can try a re-install to get them back. Been ages but I think it was port-utils or something similar as I don't remember them being the base install. Since you're using an old unsupported version of FreeBSD you're triggering some code which attempts to work around some missing features of old releases by using the sysutils/pkg_install port. It seems you have only a partial installation of this port, i.e. /usr/local/sbin/pkg_info exists but pkg_version does not. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Release 6.2
On Fri, Feb 16, 2007 at 01:49:35PM +0100, York Rapp wrote: Hello Guys. I am looking for a DVD Image of Release 6.2 to download, but unfortunately (stupid as I am ;-)) I cannot find it. The FreeBSD project only provides CD iso images. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: **questions** Re: serious performance problems with 6.2 Release
On Thu, Feb 15, 2007 at 10:50:18AM -0700, Steven H. Baeighkley wrote: If bugs is the correct list then that's where we'll send it. However we were not initially thinking it was a bug. We were thinking it was a configuration error on our part. We certainly weren't expecting kernel patches, just advice on where next to proceed. Thanks for the send-pr suggestion. We have verbose dmesg logs for all of our testing, I didn't want to send them initially because they are large and we have 12 of them. bugs isn't correct either, that's only for automated mailing of problem reports. I'd recommend either freebsd-stable or freebsd-performance, those are technical lists read by developers. Kris P.S. I second the recommendation to ignore Ted :-) ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: diagnosing a reacurring system freeze
On Thu, Feb 15, 2007 at 02:25:22PM -0700, Ross Penner wrote: Hi mailing list. I've been running FreeBSD 6.2 on a computer as a gateway for my (very small) local area network. Every two weeks since I've got it up and running, it will completely lock up, seemingly randomly. I would love to investigate the root of the problem, but I'm not exactly sure where to start. I'm sure there are many files I should be looking at but I don't know what they are. Can someone guide me in the right direction? See the chapter on kernel debugging in the developers handbook. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: diagnosing a reacurring system freeze
On Thu, Feb 15, 2007 at 11:30:25PM +0100, Zbigniew Szalbot wrote: dc0: watchdog timeout Either your dc hardware or the driver is malfunctioning, so this is what you need to address. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: diagnosing a reacurring system freeze
On Thu, Feb 15, 2007 at 05:02:27PM -0600, Derek Ragona wrote: Hard to tell if it is your dc0 ethernet adapter or a swap issue. I would try a different ethernet controller and see what happens as that is a cheap experiment. Yeah, I missed the swap message - when your system is swapping then performance will definitely be terrible. In fact now I see that I've already given this advice to the original poster on two previous occasions (November and January). I guess he chooses not to believe it. Kris ___ 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 on a Pentium 4
On Thu, Feb 08, 2007 at 11:55:24AM -0500, John Nielsen wrote: On Thursday 08 February 2007 11:16, Philip Radford wrote: Hi all, Just wondering if anyone could help me out with an issue on FreeBSD which has been puzzling me for a while and only now do I have the time to go and figure it out. We currently have version 5.4 installed but understand that the architecture it is set up for is a generic i386. How do I go about optimising my base system and/or installed ports to recognise my CPU as an i686 and therefore make use of this type of CPU. I have enclosed the first part of my dmesg output to identify the CPU. CPU: Intel(R) Pentium(R) 4 CPU 3.00GHz (3000.12-MHz 686-class CPU) Origin = GenuineIntel Id = 0xf41 Stepping = 1 Features=0xbfebfbffFPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MC A,C MOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,PBE Hyperthreading: 2 logical CPUs Add CPUTYPE?=pentium4 to /etc/make.conf. Remove the cpu I486_CPU and cpu I586_CPU lines from your kernel config (if present), leaving only cpu I686_CPU. Rebuild your kernel, world, and ports. Correct as far as it goes, but the OP will see a much bigger performance gain by updating to 6.x than the negligible gains from recompiling with different compiler optimizations. Kris ___ 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 get make buildworld to work with recent cvsup.
On Thu, Feb 08, 2007 at 04:26:37PM -0800, Nicole Harrington wrote: /usr/src/lib/libncurses/../../contrib/ncurses/ncurses/tinfo/make_keys.c /usr/obj/usr/src/tmp/usr/bin/ld: cannot find -lc Usually means that your system clock is wrong. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: what's up with ftp-archive.freebsd.org ?
On Fri, Feb 09, 2007 at 12:06:34AM -0600, Len Conrad wrote: Sometimes it works, but very often: pkg_add -r db41-4.1.25_2 Error: FTP Unable to get ftp://ftp-archive.freebsd.org/pub/FreeBSD-Archive/old-releases/i386/4.10-RELEASE/packages/Latest/db41-4.1.25_2.tgz: Not logged in pkg_add: unable to fetch 'ftp://ftp-archive.freebsd.org/pub/FreeBSD-Archive/old-releases/i386/4.10-RELEASE/packages/Latest/db41-4.1.25_2.tgz' by URL Is not open for business all day? what are the hours? It's probably just full. Kris pgpv9t7JIrd7w.pgp Description: PGP signature
Re: Intel Core Duo. SMP kernel but still only 50% load while using make on ports...
On Wed, Feb 07, 2007 at 04:44:42PM -0500, Bill Moran wrote: In response to Daniel Tourde [EMAIL PROTECTED]: Hello, I have at my disposal an Inspiron 9400 with an Intel CoreFreeBSD 6.2 is installed and rebuilt to fit the processor. The kerne l is in SMP mode. I noticed something strange: When I compile using ma in the ports tree, I only have 50% load. CPU1 is used at CPU0 is idle... How do you know this? I tried make -j2 but it did not work, Any idea? -j2 does not guarantee that you'll use both CPUs. It's entirely possible that the IO is slow enough that both of the processes are waiting on disk and only able to push the overall system usage to 50%. Try make -j99. make -j in the ports tree is not going to compile the source in parallel, it is going to try and run the port targets in parallel (but they cannot be parallelized so nothing special will happen). In theory it might work on some ports to pass in MAKE_ARGS=-j2, but a huge number of ports cannot be safely be compiled in parallel (i.e. the build will fail) because their developers have not added support for this. Kris pgpA6IMa1JtLH.pgp Description: PGP signature
Re: What Happens When /proc is not Mounted in FreeBSD5.4?
On Thu, Feb 01, 2007 at 11:17:50AM -0600, Martin McCormick wrote: I made a startling discovery when using strace to trouble-shoot a different problem on a freeBSD5.4 system that has been running since last October. Both it and another new 5.4 system had a /proc mount point but no process files. The mount point had the May 5 date from 2005 as do most files from that distribution. I mounted /proc the way it is done in fstab for 4.x systems proc /proc procfs rw 0 0 and there were all the process directories. The only reason I had done this was because strace won't work without /proc. Nothing else had seemed wrong and there hadn't been any compelling reason to look at /proc until now. Would an unmounted /proc make the system run slower since proc files allow for examination of the operation of the running processes? So basically, I have fixed the problem if it really was one in the first place. As you have found, proc is almost entirely unused in FreeBSD apart from one or two debugging facilities, and in fact not recommended on multi-user systems because the long history of security vulnerabilities. Kris pgp9hId0qEHIQ.pgp Description: PGP signature
Re: What Happens When /proc is not Mounted in FreeBSD5.4?
On Thu, Feb 01, 2007 at 02:26:00PM -0600, Martin McCormick wrote: Kris Kennaway writes: As you have found, proc is almost entirely unused in FreeBSD apart from one or two debugging facilities, and in fact not recommended on multi-user systems because the long history of security vulnerabilities. Thanks to you and Fabian Keil for your succinct answers. I took it back off and commented out the line I added to /etc/fstab so it can be brought back temporarily when needed but isn't just sitting there waiting for lightning to strike. You could also leave it in fstab with the noauto option so it can be easily mounted with mount /proc if needed. Kris pgpbqd5Jqe7Rl.pgp Description: PGP signature
Re: mknod, devfs and FreeBSD
On Tue, Jan 30, 2007 at 09:25:24AM +, Freminlins wrote: Kris, On 29/01/07, Kris Kennaway [EMAIL PROTECTED] wrote: To put it bluntly, it's something you're just going to have to get over :-) That's unhelpful. It is, in my opinion, a bad idea to have to mount up 1400 instances of devfs just to get a few device nodes. It just doesn't seem right. It's a kludge. What I will do instread is migrate the box to Solaris where I can do what I want to do. OK. It's a poor argument to say basically that's the way it is. I have always found FreeBSD to be flexible, not restrictive. That's the way it is is a statement of truth, not an argument. If devfs is the only way to go, why does mknod still exist? Why does it allow me to create device nodes that don't work? Compatibility with other OSes when used as an NFS server. Kris pgpnT0brPNzkz.pgp Description: PGP signature
Re: FreeBSD 6.1 6.2 hanging and/or spontaneous rebooting
On Tue, Jan 30, 2007 at 07:20:47PM -0600, Joe Vender wrote: I think I've fixed the spontaneous reboot and system hang problem I've been having with FBSD 6.1 6.2 as described in earlier threads of the same subject. I'm now using FBSD 6.2 to send this message. What I did was to set a hint in /boot/device.hints and also /boot/loader.conf to disable acpi hint.acpi.0.disabled=1 and then I set a hint to enable apm hint.apm.0.disabled=0 hint.apm.0.flags=0x20 plus I set apm and apmd to start at boot in /etc/rc.conf apm_enable=YES apmd_enable=YES and in /boot/loader.conf, I put apm_load=YES The messages log shows that APM is being detected, and the computer now fully powers down via shutdown -p now. I've a suggestion. It would be nice if freebsd could try to detect if acpi is supported on the computer in which it is being installed, and if none is found, it should completely disable acpi and then check if apm is supported. It should then set the appropriate settings to enable the correct power management support for the computer. This would prevent this particular spontaneous rebooting and hanging on old computers like mine which are pre-acpi but have apm support. It does try to detect it; the problem is apparently that your BIOS lies and says yep, ACPI capable!, but is in fact too buggy to use. Look for a BIOS update if possible. kris pgpr3BrAB0T6o.pgp Description: PGP signature
Re: FreeBSD 6.1 6.2 hanging and/or spontaneous rebooting
On Tue, Jan 30, 2007 at 08:01:18PM -0600, Joe Vender wrote: On Tuesday 30 January 2007 19:39, Kris Kennaway wrote: snipped It does try to detect it; the problem is apparently that your BIOS lies and says yep, ACPI capable!, but is in fact too buggy to use. Look for a BIOS update if possible. kris I hate to say it, but I spoke to soon or jinxed myself! Right after I sent the email to the list, my computer locked up tighter than a drum while loading a webpage. Argh! Here's my messages log. Hope someone can help. OK, I see you're using pppd - unfortunately this is known to have serious problems and is essentially unmaintained in FreeBSD. Use ppp(8) instead, or if you really don't want to change over then you'll have to configure debugging as in the developers' handbook chapter on kernel debugging, and try to determine whether or not pppd is really to blame. Kris pgp70FgHtovqy.pgp Description: PGP signature
Re: mknod, devfs and FreeBSD
On Mon, Jan 29, 2007 at 01:07:25PM +, Freminlins wrote: Kris, On 28/01/07, Kris Kennaway [EMAIL PROTECTED] wrote: I not understand this no sentence :) Sorry, I didn't read what I typed. I meant to type Was the effect of this considered at all? Yes it was. The benefits of dynamic devices were considered to outweight the downsides of having to mount a devfs instance. What reasons, other than cosmetic, do you have for not wanting to do this? Well, I am sure you would agree it is simpler to mknod for a small subset of /dev than to mount a devfs. Also, it means I have to migrate my existing set up which works perfectly as it is. Actually I disagree. Once you write the simple devfs ruleset it is a single command to instantiate a new /dev. You don't have to worry about making each individual device node N times and possibly making a mistake. Of course you probably have a script to do this now, but that just means you need to adjust your script as part of your migration strategy. It isn't just cosmetic, it really is more awkward than running mknod. I take your point that there's no technical reason not to do this, but it isn't pretty. To put it bluntly, it's something you're just going to have to get over :-) Kris pgpnzCOgZwbcz.pgp Description: PGP signature
Re: 6.1 hard freeze after 2 hrs cp from ntfs to ufs
On Mon, Jan 29, 2007 at 09:24:48AM -0700, Steve Franks wrote: So, I have no ideas here. I am migrating to 6.1 from win2k and I am a bit leery as my system is freezeing. Basically, I mount my ntfs drive, and I have a fresh blank disk that I want to move my hundred-odd gigs of data off of my old ntfs volume. It goes for on average about 2 hours, then the system locks up - no coredump, no ssh, no shell, dead and gone. It never freezes if I don't copy. The os is on a seperate disk on a seperate controller. I am slowly getting the copy completed, by using cp -n, but still, bsd is supposed to not crash, period. Why else replace my crappy win2k machine? (it decided it didn't want to log onto the net 75% of the time anymore, and liked to reboot every 20 mins wether you were copying or not). As I said, i've been running the bsd system daily for over 2 weeks, and it never crashes or freezes until you copy. I was even using these disks over smb before, and moved my entire cvs repository (only several hundred megs) off this same disk onto the root disk sucessfully last week. Any ideas? Out of swap? I installed on a 160GB drive with the sysinstall/bsdlabel -A (auto) option, so i would think I have a reasonable amount of swap (about 1GB as I recall). Gotta go to work. Thanks for any advice. Try with 6.2, there were some bug fixes to ntfs which might be relevant. If it persists, follow the directions in the developers handbook chapetr on kernel debugging and file a PR. Kris pgpflVoPnXCLs.pgp Description: PGP signature
Re: Release of updated 'XORG' port
On Mon, Jan 29, 2007 at 03:36:44PM -0500, Gerard Seibert wrote: Does anyone have any information regarding when the 'xorg' port will be updated from 6.9.0 to what I believe now is version X11R7.1 being release by 'x.org'? This is a FAQ, please see the archives or the wiki. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: mknod, devfs and FreeBSD
On Sun, Jan 28, 2007 at 03:56:29PM +, Freminlins wrote: Kris, On 26/01/07, Kris Kennaway [EMAIL PROTECTED] wrote: Sorry, it's the only way. Was the considered at all? I not understand this no sentence :) There's simply no way that I would mount up 1400 devfs. It is a backward step. What reasons, other than cosmetic, do you have for not wanting to do this? Kris pgpGqiZWyLboe.pgp Description: PGP signature
Re: mknod, devfs and FreeBSD
On Fri, Jan 26, 2007 at 04:40:24PM +, Freminlins wrote: Hello, I have a web server still running FreeBSD 4.7 which I want to update to FreeBSD 6.2. There are quite a few sites on this machine, and each of them has a chroot containing their own /dev. In their /dev are things like null, zero, random and so on. I don't really want to set up or mount numerous devfs file systems. I tried creating the the relevent files using mknod but they don't work. What is the best way to proceed? Set up and mount numerous devfs file systems ;) Really it's not hard, you just specify the devices you want with a simple devfs(8) ruleset. Kris pgpkI4ApWH4oZ.pgp Description: PGP signature
Re: mknod, devfs and FreeBSD
On Fri, Jan 26, 2007 at 11:05:37PM +, Freminlins wrote: Kris, On 26/01/07, Kris Kennaway [EMAIL PROTECTED] wrote: Set up and mount numerous devfs file systems ;) That is exactly what I am trying to avoid. One of the servers has 1400 sites on it, and I really don't want 1400 devfs mounts. If the only way to do this now is by having so many devfs mounts I am better off not upgrading, and it is very arugable that FreeBSD has lost some functionality by forcing such a scheme. Really it's not hard, you just specify the devices you want with a simple devfs(8) ruleset. It's not how hard it is, it's how untidy it is. Sorry, it's the only way. Kris ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Can not compile kernel.
On Fri, Jan 26, 2007 at 04:40:13PM -0800, Grant Wagner wrote: Hi, although have used various forms of unix for quite a while, I still consider myself a rather novice user. I have reciently reconfigured my machine to dual boot FreeBSD 6.2 and Windows (damn gaming addiction). I have installed a basic system (only base, games, man and src distros) and modified my /etc/make.conf to look like the following... CFLAGS= -O3 -pipe -funroll-loops -ffast-math COPTFLAGS= -O2 -pipe -funroll-loops -ffast-math Step 0) Note the warning about changing these settings in /usr/share/examples/etc/make.conf and in the make.conf manpage. Step 1) Revert those silly optimizations back to the default Step 2) Rebuild everything to undo the damage Kris pgpJr49xFYpZX.pgp Description: PGP signature
Re: Can not compile kernel.
On Fri, Jan 26, 2007 at 07:33:44PM -0800, Grant Wagner wrote: Kris Kennaway [EMAIL PROTECTED] wrote: On Fri, Jan 26, 2007 at 04:40:13PM -0800, Grant Wagner wrote: Hi, although have used various forms of unix for quite a while, I still consider myself a rather novice user. I have reciently reconfigured my machine to dual boot FreeBSD 6.2 and Windows (damn gaming addiction). I have installed a basic system (only base, games, man and src distros) and modified my /etc/make.conf to look like the following... CFLAGS= -O3 -pipe -funroll-loops -ffast-math COPTFLAGS= -O2 -pipe -funroll-loops -ffast-math Step 0) Note the warning about changing these settings in /usr/share/examples/etc/make.conf and in the make.conf manpage. Step 1) Revert those silly optimizations back to the default Step 2) Rebuild everything to undo the damage Kris Well, in short, that worked. I have now build the kernel. I'm a little confused though and could use a bit of an explination. I thought only the COPTFLAGS options where used during kernel compilation and I had attempted to build with those commented out completely before. I can only guess that the CFLAGS are still in effect too. CFLAGS are used for module builds. Now I have a custom kernel which is failing to build. I've attached the config file for it, and it fails trying to build with references about ieee80211. The odd thing is I have no wireless in my box and have commented out all the wireless references. What else is dependant on them and should be commented out as well? The last bit of output is below. Go back to GENERIC (you stripped out too much) or check the comments more carefully...or note the error message and check whether you have anything related still in your kernel. if_ural.o(.text+0x66): In function `ural_free_tx_list': : undefined reference to `ieee80211_free_node' if_ural.o(.text+0x2d3): In function `ural_rxeof': : undefined reference to `ieee80211_find_rxnode' Kris P.S. Please wrap your lines at 70 characters so that your emails may be easily read. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Page Faulting Box?
On Mon, Jan 22, 2007 at 10:25:31AM -0800, Jay Chandler wrote: One of our servers is restarting at random. Not entirely sure what causes it-- hopefully someone here can help me track it down Often it's bad hardware, especially if only manifesting onone server out of several with the same hardware and workload. (I suspect hardware at some point, potentially the Broadcom NIC). This is what's in the messages log-- what else can I provide y'all with? The backtrace from the crashdump, or at the very least look up the IP in the kernel.debug. See the developers handbook chapter on kernel debugging. Kris Jan 22 10:16:55 montreal kernel: kernel trap 12 with interrupts disabled Jan 22 10:16:55 montreal kernel: Jan 22 10:16:55 montreal kernel: Jan 22 10:16:55 montreal kernel: Fatal trap 12: page fault while in kernel mode Jan 22 10:16:55 montreal kernel: cpuid = 2; apic id = 06 Jan 22 10:16:55 montreal kernel: fault virtual address = 0x104 Jan 22 10:16:55 montreal kernel: fault code = supervisor read, page not present Jan 22 10:16:55 montreal kernel: instruction pointer= 0x20:0xc066c731 Jan 22 10:16:55 montreal kernel: stack pointer = 0x28:0xe4f99c90 Jan 22 10:16:55 montreal kernel: frame pointer = 0x28:0xe4 Jan 22 10:16:55 montreal kernel: f99c9c Jan 22 10:16:55 montreal kernel: Jan 22 10:16:55 montreal kernel: code segment = base 0x0, limit 0xf, type 0x1b Jan 22 10:16:55 montreal kernel: = DPL 0, pres 1, def32 1, gran 1 Jan 22 10:16:55 montreal kernel: Jan 22 10:16:55 montreal kernel: processor eflags = resume, IOPL = 0 Jan 22 10:20:58 montreal syslogd: kernel boot file is /boot/kernel/kernel -- Jay Chandler Network Administrator, Chapman University 714.628.7249 / [EMAIL PROTECTED] Today's Excuse: excess surge protection ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED] pgp5s389l5Ln2.pgp Description: PGP signature
Re: virtual memory management
On Sat, Jan 20, 2007 at 08:57:27AM +0100, Zbigniew Szalbot wrote: Dear all, Is there a FBSD command to manage virtual memory? I think my swap size is now a bit too much used: last pid: 19824; load averages: 0.06, 0.05, 0.02 up 50+10:00:17 08:54:00 230 processes: 1 running, 227 sleeping, 2 zombie CPU states: 0.0% user, 0.0% nice, 0.4% system, 0.8% interrupt, 98.8% idle Mem: 232M Active, 27M Inact, 91M Wired, 212K Cache, 60M Buf, 142M Free Swap: 512M Total, 482M Used, 29M Free, 94% Inuse The swap size usage grow so big probably because I started wget to download an iso image and then WinSCP to grab it from the FBSD machine to my laptop. When I started wget, the swap usage was around 19% and had been like that for many days. That should not cause such a thing (wget does not try to fit the whole file in memory, so it won't be pushing stuff out to swap). Look at the process sizes in top to see what is using the swap space - something(s) that is still running is using that 482MB. Probably you have one or more processes that are using a large amount of virtual memory, which is too big to fit in RAM. That's the situation you need to address. Is there any way to handle swap size usage other than restarting the box? kill(1). Kris pgpuL9xkWePKe.pgp Description: PGP signature
Re: virtual memory management
On Sat, Jan 20, 2007 at 09:13:48AM +0100, Zbigniew Szalbot wrote: Hello again, The swap size usage grow so big probably because I started wget to download an iso image and then WinSCP to grab it from the FBSD machine to my laptop. When I started wget, the swap usage was around 19% and had been like that for many days. That should not cause such a thing (wget does not try to fit the whole file in memory, so it won't be pushing stuff out to swap). Look at the process sizes in top to see what is using the swap space - something(s) that is still running is using that 482MB. I do not see such a process: PID USERNAME THR PRI NICE SIZERES STATETIME WCPU COMMAND 21442 root 2 200 224M 26128K kserel 0:06 0.00% java 896 mysql 16 200 70756K 14764K kserel 255:35 0.00% mysqld 98693 bind 1 960 32812K 32172K select 0:22 0.00% dnscache 5035 www 1 40 28372K 15660K accept 5:05 1.86% httpd 5026 www 1 40 28240K 15104K accept 4:46 0.00% httpd 5065 www 1 40 28128K 15196K accept 4:29 0.00% httpd 5030 www 1 40 27892K 15144K accept 4:21 0.00% httpd 5126 www 1 40 27784K 14864K accept 4:20 0.00% httpd 5029 www 1 40 27760K 14644K accept 4:22 0.00% httpd 5027 www 1 40 27740K 15140K accept 4:30 0.00% httpd 5028 www 1 40 27516K 14812K accept 4:03 0.00% httpd 95977 www 1 40 27216K 14532K accept 2:21 0.00% httpd 703 root 1 960 16412K 2296K select 4:35 0.00% httpd 91014 mailman 1 80 11492K 1600K nanslp 6:00 0.00% python 91012 mailman 1 80 11024K 1560K nanslp 5:32 0.00% python 91010 mailman 1 80 11008K 1568K nanslp 5:23 0.00% python 91009 mailman 1 80 11008K 1552K nanslp 5:20 0.00% python I see lots of them; every one in that list is contributinig. If you add up all those process sizes you'll see where the space is going. Basically you are just overloading your system by trying to run too much at once. Reduce the load or add more RAM. Kris pgpA2qhyW02hO.pgp Description: PGP signature
Re: virtual memory management
On Sat, Jan 20, 2007 at 03:51:38AM -0500, Kris Kennaway wrote: On Sat, Jan 20, 2007 at 09:13:48AM +0100, Zbigniew Szalbot wrote: Hello again, The swap size usage grow so big probably because I started wget to download an iso image and then WinSCP to grab it from the FBSD machine to my laptop. When I started wget, the swap usage was around 19% and had been like that for many days. That should not cause such a thing (wget does not try to fit the whole file in memory, so it won't be pushing stuff out to swap). Look at the process sizes in top to see what is using the swap space - something(s) that is still running is using that 482MB. I do not see such a process: PID USERNAME THR PRI NICE SIZERES STATETIME WCPU COMMAND 21442 root 2 200 224M 26128K kserel 0:06 0.00% java 896 mysql 16 200 70756K 14764K kserel 255:35 0.00% mysqld 98693 bind 1 960 32812K 32172K select 0:22 0.00% dnscache 5035 www 1 40 28372K 15660K accept 5:05 1.86% httpd 5026 www 1 40 28240K 15104K accept 4:46 0.00% httpd 5065 www 1 40 28128K 15196K accept 4:29 0.00% httpd 5030 www 1 40 27892K 15144K accept 4:21 0.00% httpd 5126 www 1 40 27784K 14864K accept 4:20 0.00% httpd 5029 www 1 40 27760K 14644K accept 4:22 0.00% httpd 5027 www 1 40 27740K 15140K accept 4:30 0.00% httpd 5028 www 1 40 27516K 14812K accept 4:03 0.00% httpd 95977 www 1 40 27216K 14532K accept 2:21 0.00% httpd 703 root 1 960 16412K 2296K select 4:35 0.00% httpd 91014 mailman 1 80 11492K 1600K nanslp 6:00 0.00% python 91012 mailman 1 80 11024K 1560K nanslp 5:32 0.00% python 91010 mailman 1 80 11008K 1568K nanslp 5:23 0.00% python 91009 mailman 1 80 11008K 1552K nanslp 5:20 0.00% python I see lots of them; every one in that list is contributinig. If you add up all those process sizes you'll see where the space is going. By which I mean the difference between size and res, which indicates the amount of process memory allocated but not currently resident in RAM. This isn't a foolproof method (see e.g. the FAQ entry on rpc.statd), but it's true in your case. Basically you are just overloading your system by trying to run too much at once. Reduce the load or add more RAM. Kris pgpTxEofIrvtn.pgp Description: PGP signature
Re: Why does my machine give a panic?
On Sat, Jan 20, 2007 at 02:36:47PM +0200, Halid Faith wrote: I used to use freebsd4.X. I didn't have any problem about it. After I reinstalled freebsd6.1 my server has started to given a panic. I am sure that the server's hardware is good. I ran kgdb on the server as below; # kgdb kernel.debug /var/crash/vmcore.0 [GDB will not be able to debug user-mode threads: /usr/lib/libthread_db.so: Undefined symbol ps_pglobal_lookup] GNU gdb 6.1.1 [FreeBSD] Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type show copying to see the conditions. There is absolutely no warranty for GDB. Type show warranty for details. This GDB was configured as i386-marcel-freebsd. Unread portion of the kernel message buffer: dev = amrd0s1e, block = 1008936, fs = /var panic: ffs_blkfree: freeing free block cpuid = 1 Uptime: 1d18h28m28s Dumping 1023 MB (2 chunks) chunk 0: 1MB (159 pages) ... ok chunk 1: 1023MB (261886 pages) 1008 992 976 960 944 928 912 896 880 864 848 832 816 800 784 768 752 736 720 704 688 672 656 640 624 608 592 576 560 544 528 512 496 480 464 448 432 416 400 384 368 352 336 320 304 288 272 256 240 224 208 192 176 160 144 128 112 96 80 64 48 32 16 #0 doadump () at pcpu.h:165 165 __asm __volatile(movl %%fs:0,%0 : =r (td)); How to get rid off that ? 1) fsck the filesystem in case you have lingering corruption that was not detected before. 2) update to 6.2 3) if it still persists, then follow the directions in the developers handbook chapter on kernel debugging, and file a PR. Kris pgpOx0YhnMCy0.pgp Description: PGP signature
Re: Mail server intermittent freeze
On Sat, Jan 20, 2007 at 07:56:08PM -0600, Rich Winkel wrote: I'm pulling out what's left of my hair trying to figure out this one. It's not a pretty sight. Save the people who have to look at me! It's a 1ghz intel P3 with 512MB ram, running 4.11-p26 with sendmail, imapd-uw, qpopper, stunnel-4.14_2 and p5-Mail-SpamAssassin-3.1.1, as well as tty logins. It's an nis and nfs client (nfs=home dirs). It has an fxp network card which is in polling mode (although the problem started before it was put in polling mode) The system will freeze maybe every 5 minutes, sometimes for up to a minute. Almost completely: low level terminal io on the console still works. I can switch tty's (ALT-Fn) and carriage returns are echoed (and discarded) while showing netstat -w 1 output. But interactive prompts are frozen. No user-level processing is apparent. Network traffic is heavy, we're getting massive amounts of spam, and spamd's load on the system has jumped considerably in the past few weeks. Average system load used to hover around .8, now it averages over 2, mostly due to spamd. Usually this is because a transient load is causing it to swap. 512MB isn't really a lot of RAM on a heavily loaded server. Kris pgpJAgtxeasK7.pgp Description: PGP signature
Re: handbook section 25.9.2.3 appears dated (Samba3)
On Fri, Jan 19, 2007 at 08:23:23PM -0700, Steve Franks wrote: Just installed latest ports/net/samba3 and the smb.conf file disagrees with section 25.9.2.3 Security Settings (Samba) of the handbook with respect to the default authentication mode so far as I can tell - look like smbpasswd has been replaced with something newer? Check the samba documentation (and then consider submitting an update once you've found the answer). kris pgpQYR8fbZvj0.pgp Description: PGP signature
Re: md5sum is missing, but not entirely
On Thu, Jan 18, 2007 at 01:29:44AM -0800, [EMAIL PROTECTED] wrote: The command to calculate md5 sums is 'md5', not 'md5sum', and it is part of the base. ... Not sure why apropos md5 shows md5sum and gmd5sum (g for gnu?), maybe it's from a port. I also suspect that gmd5sum is a link to md5sum or vice versa. gmd5sum is indeed from a port, but md5sum does not exist at all on my system, either as an executable or as a manpage. md5 does exist, as both an executable and a manpage. Either md5 or gmd5sum can be used to verify downloads; as someone else mentioned they produce the same results (although formatted differently). The coreutils package installs the GNU utilities that form the basis of Linux distributions - grep, chmod etc. Since most of these names clash with the FreeBSD base system, the binaries all get renamed with a g prefix. The GNU docs still internally refer to them with their original names e.g. man gmd5sum will refer to md5sum It still seems like a bug that apropos includes md5sum(1) in its output, when no such manpage exists. Bug in that port's manpage. Talk to the port maintainer or fix it yourself and submit the patch. Kris pgpfZiT0OHhDp.pgp Description: PGP signature
Re: ports - make index fail
On Thu, Jan 18, 2007 at 10:25:58AM -0800, ann kok wrote: Hi all I install 6.2 and update the port by cvsup but i fail to make index Thank you Add delta 1.26 2006.11.27.16.49.49 oliver Shutting down connection to server Finished successfully f62# cd /usr/ports f62# make index Generating INDEX-6 - please wait..perl: not found Install perl. Kris pgpiCIX8gISbS.pgp Description: PGP signature
Re: Does Firefox run on the SPARC64 port of FreeBSD?
On Thu, Jan 18, 2007 at 04:59:17PM +0100, Christian Baer wrote: Michael Johnson wrote: Firefox only runs on = 601101 sparc64. I am guessing that means a special revision of the UltraSPARC II processor, but I don't really know, because google gets a lot of hits, mainly explaining all sorts of soft that seems to have the same problem, but none of these hits really explain the meaning behind this. So even though this is getting a little OT: In English, please! That's a number indicating a version of FreeBSD, see http://www.freebsd.org/doc/en_US.ISO8859-1/books/porters-handbook/book.html#AEN5722 Basically, it does not work on 6.1-RELEASE, so you should consider updating to 6.2-RELEASE. Kris pgpJmirFqR5mK.pgp Description: PGP signature
Re: md5sum is missing, but not entirely
On Tue, Jan 16, 2007 at 10:29:57PM -0800, [EMAIL PROTECTED] wrote: Both the executable and the manpage for md5sum seem to be missing from this 6.1-RELEASE system, making it difficult to check the downloaded 6.2-RELEASE and FreeSBIE ISOs. Isn't md5sum supposed to be part of the base? Meanwhile, apropos md5 yields this line, among others gmd5sum(1), md5sum(1) - compute and check MD5 message digest but man md5sum says No manual entry for md5sum How did that line get into the apropos, when the corresponding manpage is (apparently) not installed? I dunno, it doesn't seem to be part of my FreeBSD system (nor has it ever been in FreeBSD, afaik). Are you sure you didnt run that command on a Linux system by mistake? The md5 utility is called md5 in FreeBSD, and apropos md5 correctly returns for me: md5(1), sha1(1), sha256(1), rmd160(1) - calculate a message-digest fingerprint (checksum) for a file Kris pgpuIRmVwCsyk.pgp Description: PGP signature
Re: What have you done for me lately !!!
On Tue, Jan 16, 2007 at 08:06:58AM +1100, Joe Arcaro wrote: Hi, Maybe this is just a rant, But I'll vent anyway. I've been watching with some skepticism, the whole apple circus freak fanboy show ... I was just curious, does it not bother any of the BSD developers that Apple inc (TM) has based its entire business model on open source software, and yet seems to have given little if any thing back in return. I have on occasion looked at the apple web site, and never has apple even given credit to any form of BSD ! Is this all just acceptable, have I missed the whole point of open source (Give something back when you can) or have I not read the fine print. anyhow... Cheers to all the developers at Freebsd, I am an admirer of BSD and the open source community Yeah, you did miss the point. The BSD license does not require this, and we BSD developers work on BSD software because we acknowledge and even like that aspect of it: we are developing quality software that can be used for any purposes with effectively no strings attached. Besides, apple has and does contribute code back to FreeBSD. Kris pgpX23Ch6Zii4.pgp Description: PGP signature
Re: FreeBSD 6.2 stable crasches when running dump on mounted snapshot.
On Tue, Jan 16, 2007 at 09:47:47PM +0100, Mattias Bj?rk wrote: Hi there, When I run dump on a mounted snapshot, my machine panics with the error that says the following: Fatal double fault Panic: double fault You forgot to mention/obtain the important bits of the error ;) See the chapter on kernel debugging in the developers handbook, then follow up to stable@ and/or file a PR. Kris pgpxYfOXvA18I.pgp Description: PGP signature
Re: uname question after update
On Mon, Jan 15, 2007 at 10:37:19AM -0800, Jay Chandler wrote: I have two boxes I've updated so far to 6.2. uname -a returns two different strings: FreeBSD box1.mydomain.com 6.2-RELEASE FreeBSD 6.2-RELEASE #0: Fri Jan 12 20:01:29 PST 2007 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/SMP i386 FreeBSD box2.mydomain.com 6.2-RELEASE FreeBSD 6.2-RELEASE #4: Sat Jan 13 15:40:40 PST 2007 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/SMP i386 What does the #0 / #4 mean? The number of times you have recompiled your kernel. Kris pgplfEQ9ZsDJ6.pgp Description: PGP signature
Re: FreeBSD 6.2, rebuild. 'i386-undermydesk-freebsd'?
On Mon, Jan 15, 2007 at 08:45:31PM +0100, Daniel Tourde wrote: Hello, I have been rebuilding FreeBSD6.2 yesterday and I noticed a strange flag: -DTARGET='i386-undermydesk-freebsd' or something like this. What is this? I have set CPUTYPE=pentium4. I had expected something like -DTARGET='i386-pc-freebsd' or something like this. Any idea? It's meaningless so you can just ignore it. Kris pgpt1Ip52STzu.pgp Description: PGP signature
Re: FreeBSD 6.2, rebuild. NO-MMX, NO-SSE?
On Mon, Jan 15, 2007 at 08:52:04PM +0100, Daniel Tourde wrote: Hello, Yesterday I rebuit FreeBSD 6.2 using CPUTYPE=pentium4 and the classic procedure described in the manual. The machine (Inspiron 9400) is fast but I saw at certain moments something like NO-MMX, NO-SSE (some flags or variables) during the compilation process. I thought then How come? What a pity not to use these instructions. Can someone tell me what it was and if it is really supposed to be like this? My roots are in Gentoo Linux where it is possible to get the maximum out of a processor when building a system from scratch by using properly certain C and C++ flags. I am pretty sure the same is possible with FreeBSD, however these aforementionned flags raised some questions in my mind... ;) Yes, it's supposed to be that way. Certain parts of the FreeBSD system cannot use MMX or SSE instructions (e.g. the boot loader) but it's okay since they are absolutely not performance critical. Kris pgp02zTFfWhMG.pgp Description: PGP signature
Re: general question about packages and ports working together
On Thu, Jan 11, 2007 at 07:58:08AM -0600, Kirk Strauser wrote: On Wednesday 10 January 2007 8:38 pm, Kris Kennaway wrote: On Thu, Jan 11, 2007 at 03:30:18AM +0100, Stevan Tiefert wrote: Fetching ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages-6.1-release/All/k delibs-3.5.1_1.tbz... Done. pkg_add: package 'kdelibs-3.5.1_1' conflicts with kdelibs-nocups-3.5.5 pkg_add: please use pkg_delete first to remove conflicting package(s) or -f to f orce installation pkg_add: pkg_add of dependency 'kdelibs-3.5.1_1' failed! vagabund# That's the problem, you have a conflicting package (kdelibs built with WITHOUT_CUPS set) installed. If you really want to use packages you'll need to revert that to the standard setting of including cups support. Well, that and he's trying to install -RELEASE packages on a -RELEASE system with -STABLE ports. How does one tell ports to install -STABLE packages - is that uname-dependent? Yeah, if uname = *-STABLE it gets the stable packages. You can just set PACKAGESITE if you want to use the poorly-tested but updated stable packages on a release system. Kris pgpmW6hlFdayt.pgp Description: PGP signature
Re: /pub/FreeBSD/ports/i386/packages-5-stable/All
On Thu, Jan 11, 2007 at 02:00:08AM -0500, Kris Kennaway wrote: On Thu, Jan 11, 2007 at 05:57:56PM +1100, Ian Smith wrote: Hi Kris, I know things must be pretty busy with 6.2, but is there any chance that the 5.5-STABLE packages can be updated soon? I just checked again, and at least apache and phpyadmin are still stale, going on two months now. Mark, what is the status of the upload of these packages? OK, I've uploaded the packages now and they'll begin propagating out to the mirrors. Kris pgpshUaurwYPz.pgp Description: PGP signature
Re: general question about packages and ports working together
On Wed, Jan 10, 2007 at 11:49:04PM +0100, Stevan Tiefert wrote: Hello list, portaudit suggested me to update kdelibs. Ok I've done it via ports. Two days later I wanted to add the package de-koffice-i18n. The package de-koffice-i18n tried to install also my unsecure kdelibs again, if I hadn't stopped it I would now have two kdelibs on my harddrive... May it be that the packages are not accepting the newer versions from the ports? No, this should not be it. Post the exact output of the commands you tried so we can try to help. Kris pgpBoKI9Ij3fZ.pgp Description: PGP signature
Re: general question about packages and ports working together
On Thu, Jan 11, 2007 at 03:30:18AM +0100, Stevan Tiefert wrote: Am Donnerstag, 11. Januar 2007 02:47 schrieb Kris Kennaway: On Wed, Jan 10, 2007 at 11:49:04PM +0100, Stevan Tiefert wrote: Hello list, portaudit suggested me to update kdelibs. Ok I've done it via ports. Two days later I wanted to add the package de-koffice-i18n. The package de-koffice-i18n tried to install also my unsecure kdelibs again, if I hadn't stopped it I would now have two kdelibs on my harddrive... May it be that the packages are not accepting the newer versions from the ports? No, this should not be it. Post the exact output of the commands you tried so we can try to help. Kris Excuse me! It seems that I need sleep :-( It is 3:30 am in my country... That is the right log: vagabund# pkg_add -r de-koffice-i18n Fetching ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages-6.1-release/Lates t/de-koffice-i18n.tbz... Done. Fetching ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages-6.1-release/All/k delibs-3.5.1_1.tbz... Done. pkg_add: package 'kdelibs-3.5.1_1' conflicts with kdelibs-nocups-3.5.5 pkg_add: please use pkg_delete first to remove conflicting package(s) or -f to f orce installation pkg_add: pkg_add of dependency 'kdelibs-3.5.1_1' failed! vagabund# That's the problem, you have a conflicting package (kdelibs built with WITHOUT_CUPS set) installed. If you really want to use packages you'll need to revert that to the standard setting of including cups support. Kris pgpn1Egi9izw5.pgp Description: PGP signature
Re: general question about packages and ports working together
On Thu, Jan 11, 2007 at 03:47:34AM +0100, Stevan Tiefert wrote: Am Donnerstag, 11. Januar 2007 03:38 schrieb Kris Kennaway: On Thu, Jan 11, 2007 at 03:30:18AM +0100, Stevan Tiefert wrote: Am Donnerstag, 11. Januar 2007 02:47 schrieb Kris Kennaway: On Wed, Jan 10, 2007 at 11:49:04PM +0100, Stevan Tiefert wrote: Hello list, portaudit suggested me to update kdelibs. Ok I've done it via ports. Two days later I wanted to add the package de-koffice-i18n. The package de-koffice-i18n tried to install also my unsecure kdelibs again, if I hadn't stopped it I would now have two kdelibs on my harddrive... May it be that the packages are not accepting the newer versions from the ports? No, this should not be it. Post the exact output of the commands you tried so we can try to help. Kris Excuse me! It seems that I need sleep :-( It is 3:30 am in my country... That is the right log: vagabund# pkg_add -r de-koffice-i18n Fetching ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages-6.1-release/Lates t/de-koffice-i18n.tbz... Done. Fetching ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages-6.1-release/All/k delibs-3.5.1_1.tbz... Done. pkg_add: package 'kdelibs-3.5.1_1' conflicts with kdelibs-nocups-3.5.5 pkg_add: please use pkg_delete first to remove conflicting package(s) or -f to f orce installation pkg_add: pkg_add of dependency 'kdelibs-3.5.1_1' failed! vagabund# That's the problem, you have a conflicting package (kdelibs built with WITHOUT_CUPS set) installed. If you really want to use packages you'll need to revert that to the standard setting of including cups support. Kris Hello Kris, that is the problem... The other mail I sent before handles abaout the kdelibs-upgrade I did two days ago. And you will see, that the building stopped where cups should be integrated. You see also, that it failed and that was the reason I installed the kdelibs-nocups port. The kdelibs don't want to be installed. I don't understand why... OK, that's what you need to solve. You can either post your errors here so we can try to solve them, or just delete the nocups port and go with the package. Kris pgpFcMdcFDLdT.pgp Description: PGP signature
Re: general question about packages and ports working together
On Thu, Jan 11, 2007 at 04:10:59AM +0100, Stevan Tiefert wrote: Am Donnerstag, 11. Januar 2007 04:05 schrieb Kris Kennaway: On Thu, Jan 11, 2007 at 03:47:34AM +0100, Stevan Tiefert wrote: Am Donnerstag, 11. Januar 2007 03:38 schrieb Kris Kennaway: On Thu, Jan 11, 2007 at 03:30:18AM +0100, Stevan Tiefert wrote: Am Donnerstag, 11. Januar 2007 02:47 schrieb Kris Kennaway: On Wed, Jan 10, 2007 at 11:49:04PM +0100, Stevan Tiefert wrote: Hello list, portaudit suggested me to update kdelibs. Ok I've done it via ports. Two days later I wanted to add the package de-koffice-i18n. The package de-koffice-i18n tried to install also my unsecure kdelibs again, if I hadn't stopped it I would now have two kdelibs on my harddrive... May it be that the packages are not accepting the newer versions from the ports? No, this should not be it. Post the exact output of the commands you tried so we can try to help. Kris Excuse me! It seems that I need sleep :-( It is 3:30 am in my country... That is the right log: vagabund# pkg_add -r de-koffice-i18n Fetching ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages-6.1-release/Lat es t/de-koffice-i18n.tbz... Done. Fetching ftp://ftp.freebsd.org/pub/FreeBSD/ports/i386/packages-6.1-release/All /k delibs-3.5.1_1.tbz... Done. pkg_add: package 'kdelibs-3.5.1_1' conflicts with kdelibs-nocups-3.5.5 pkg_add: please use pkg_delete first to remove conflicting package(s) or -f to f orce installation pkg_add: pkg_add of dependency 'kdelibs-3.5.1_1' failed! vagabund# That's the problem, you have a conflicting package (kdelibs built with WITHOUT_CUPS set) installed. If you really want to use packages you'll need to revert that to the standard setting of including cups support. Kris Hello Kris, that is the problem... The other mail I sent before handles abaout the kdelibs-upgrade I did two days ago. And you will see, that the building stopped where cups should be integrated. You see also, that it failed and that was the reason I installed the kdelibs-nocups port. The kdelibs don't want to be installed. I don't understand why... OK, that's what you need to solve. You can either post your errors here so we can try to solve them, or just delete the nocups port and go with the package. Kris Hello again, this is the log of kdelibs3: vagabund# cd /usr/ports/x11/kdelibs3 vagabund# make install ~/kdelibs3-install.log vagabund# And here are the last few raws of kdelibs3-install.log: OK, make sure everything required by kdelibs is up-to-date (with portupgrade -R kdelibs or similar). It is buildable on a clean 4.x system, although since 4.x is EOL in a couple of weeks you might prefer to spend your time on an upgrade to a supported version like 6.2. Kris pgptup6hjw0fe.pgp Description: PGP signature
Re: /pub/FreeBSD/ports/i386/packages-5-stable/All
On Thu, Jan 11, 2007 at 05:57:56PM +1100, Ian Smith wrote: Hi Kris, I know things must be pretty busy with 6.2, but is there any chance that the 5.5-STABLE packages can be updated soon? I just checked again, and at least apache and phpyadmin are still stale, going on two months now. Mark, what is the status of the upload of these packages? Kris pgpUFYxpmYL53.pgp Description: PGP signature
Re: diablo-jdk 1.5.0 problem
On Tue, Jan 09, 2007 at 07:54:34AM +0100, scaligeracarni wrote: I made : pkg_add diablo-jdk-freebsd6.i386.1.5.0.07.01.tbz on a freebsd 6.1 but when I try to run java the computer say that I need libz.so.2! How can resolve my problem? Are you absolutely sure you are running the freebsd 6 binary and not the freebsd 5? libz.so.2 is the version present in FreeBSD 5. Kris pgpSWWlsdGxYN.pgp Description: PGP signature
Re: is THIS why the 6.2 release seems stalled ?
On Mon, Jan 08, 2007 at 09:33:18AM -0800, Jim Pazarena wrote: http://farragut.flameeyes.is-a-geek.org/articles/2007/01/08/a-shadow-lies-upon-all-bsd-distributions - Gentoo/FreeBSD: license problems require a development pause http://farragut.flameeyes.is-a-geek.org/articles/2007/01/07/gentoo-freebsd-license-problems-requires-a-development-pause The big license mess, part 2 http://farragut.flameeyes.is-a-geek.org/articles/2007/01/07/the-big-license-mess-part-2 -- Gentoo/FreeBSD On Hold Due To Licensing Issues No. Kris pgpDgGN952vdm.pgp Description: PGP signature
Re: filesystem size
On Tue, Jan 09, 2007 at 02:01:47PM -0500, Andy Greenwood wrote: On 1/9/07, Simon Gao [EMAIL PROTECTED] wrote: Hi, What's largest filesystem size supported by FreeBSD 5.2.1 i386? You might want to read this: http://www.freebsd.org/projects/bigdisk/index.html And update to a modern release of FreeBSD, if large filesystem support is important to you. Kris pgpyGrRXbEvp5.pgp Description: PGP signature
Re: PMAP_SHPGPERPROC
On Sat, Dec 30, 2006 at 08:43:14PM +1030, Ian Moore wrote: Hi, Since I upgraded to KDE 3.5.5 the other day, I've been getting the following messages after KDE starts: kernel: collecting pv entries -- suggest increasing PMAP_SHPGPERPROC I've noticed the system seems a bit unstable, it often locks up when KDE starts and I guess that is the reason for the instability. I doubt it. I did a bit of hunting on the mailing list archives, but could only come up with one question that didn't shed much light on the matter. Does anyone know what I should do to increase PMAP_SHPGPERPROC? Change the value in your kernel configuration file. I'm running 5.5-RELEASEp9 on i386. Try updating to 6.2 if you are having instability problems, and then proceed from there. Kris pgpHphZOXErll.pgp Description: PGP signature
Re: processes not getting fair share of available disk I/O
On Thu, Dec 14, 2006 at 09:07:35AM +, Dieter wrote: Mutex profiling would show if there is a mutex somehow getting in the way of your I/O (e.g. if Giant is somehow being forced). I dont think it would show anything though. You can try to study interrupt issues (e.g. look for an interrupt storm during I/O) with vmstat -i. Other than that you'd probably have to get your hands dirtier in the code. =20 maxtotal count avg cnt_hold cnt_lock name 1158725 11853301596 74200 /usr/src= /sys/amd64/amd64/pmap.c:1563 (pmap) 1158721 11665931596 7301 17 /usr/src= /sys/amd64/amd64/pmap.c:1562 (vm page queue mutex) 90598 578551 199304 234 /usr/src/= sys/kern/kern_sx.c:157 (lockbuilder mtxpool) 83234 967612 124000 700 /usr/src/= sys/vm/vm_fault.c:906 (vm object) If I'm reading the man page right, pmap holds a lock for over 1 second? In total, over 1600 operations. It's not an issue. The man page says: max The longest continuous hold time in microseconds. Which together with the total number, would imply 1 time it took 1.158725 seconds, and the other 1595 times averaged 16.7 usec, which seems unlikely. Am I misinterpreting the man page? Sorry, yes. Nothing else contended for it though, so it doesn't appear to be a source of performance problems - it is probably a secondary effect from something else. I guess you're running some old version of FreeBSD since those line numbers don't correspond to anything reasonable in the current 6.x source, so I dunno what exactly. The rest looks fine at a quick glance too. What should I be looking for? Do I need to collect stats for a long period of time, or is a few seconds enough? Dd can kill the transfer in about 3 seconds. You need to make sure your sampling while the system is in the bad state. A mutex that has a lot of acquisitions and a lot of contention for those acquisitions is a performance bottleneck. Nothing below falls into that class - in particular it's definitely not Giant causing performance loss to the filesystem. Still looks like it's a driver and/or hardware problem, but you'd need specialized knowledge to proceed with debugging it. Here are stats with the Ethernet-disk program competing with dd from /dev/zero to a file, sorted by cnt_hold and by cnt_lock. maxtotal count avg cnt_hold cnt_lock name 89 6095988 1089530 5 536 979 /usr/src/sys/kern/kern_intr.c:546 (Giant) 2649 1950594 360310 5 4490 /usr/src/sys/netinet/tcp_usrreq.c:598 (tcp) 13553 2348809 2361099 3502 /usr/src/sys/kern/kern_sysctl.c:1280 (Giant) 36 230008 85274 2 141 241 /usr/src/sys/kern/kern_timeout.c:258 (Giant) 453 300251 180186 1 890 /usr/src/sys/kern/uipc_socket.c:1011 (so_rcv) 199878 23516414706 499 884 /usr/src/sys/kern/kern_synch.c:236 (Giant) 2138 337086 180186 1 850 /usr/src/sys/kern/uipc_socket.c:1255 (so_rcv) 67 1120044 722561 1 830 /usr/src/sys/kern/subr_taskqueue.c:219 (taskqueue) 57 219660 134214 1 70 102 /usr/src/sys/netinet/tcp_subr.c:1404 (tcp) 73 1830084 18034410 650 /usr/src/sys/dev/bge/if_bge.c:3321 (bge0) 472 272697 180155 1 630 /usr/src/sys/kern/uipc_socket.c:1401 (so_rcv) 77 541532 361447 1 480 /usr/src/sys/kern/subr_taskqueue.c:204 (taskqueue) 519826977 244 110 261 /usr/src/sys/kern/kern_exit.c:233 (Giant) 50 545822 361062 1 260 /usr/src/sys/net/netisr.c:233 (ip_inq) 1649 287241 102509 2 24 30 /usr/src/sys/kern/vfs_bio.c:1397 (buf queue lock) 20981 212022931722 192 /usr/src/sys/vm/vm_fault.c:683 (vm object) 65261623656 7 184 /usr/src/sys/kern/kern_conf.c:324 (Giant) 7546 218853 412 531 171 /usr/src/sys/amd64/amd64/pmap.c:2481 (vm page queue mutex) 11013982 24457 170 /usr/src/sys/kern/kern_exit.c:385 (Giant) 491 278319 170512 1 162 /usr/src/sys/kern/kern_synch.c:238 (process lock) 5227355227912 169 /usr/src/sys/net/netisr.c:339 (Giant) 57627530137919
Re: Upgrade a binary package
On Thu, Dec 14, 2006 at 12:46:05PM -0700, Eric Brunson wrote: I've been searching the manual and the man pages, but I can't seem to find a command to update an installed package to a newer version. I always get a message like this: pkg_add: package 'expect-5.43.0_1' or its older version already installed I only find info on pkg_add and pkg_delete for binary packages, the only references to updating a package give directions on how to use the ports tree. Is there a way to update an installed package with a newer version? sysutils/portupgrade is the usual tool. I think portmaster can do this too. Kris pgpE4zZ93tjb8.pgp Description: PGP signature
Re: processes not getting fair share of available disk I/O
On Thu, Dec 14, 2006 at 02:37:31PM +, Dieter wrote: Sorry, yes. Nothing else contended for it though, so it doesn't appear to be a source of performance problems - it is probably a secondary effect from something else. I guess you're running some old version of FreeBSD since those line numbers don't correspond to anything reasonable in the current 6.x source, so I dunno what exactly. FreeBSD 6.0 Erk. How about retrying with something modern ;-) We do fix lots of bugs over time you know! The rest looks fine at a quick glance too. =20 What should I be looking for? Do I need to collect stats for a long period of time, or is a few seconds enough? Dd can kill the transfer in about 3 seconds. You need to make sure your sampling while the system is in the bad state. A mutex that has a lot of acquisitions and a lot of contention for those acquisitions is a performance bottleneck. Nothing below falls into that class - in particular it's definitely not Giant causing performance loss to the filesystem. Aren't the numbers (other than max and avg) going to depend a lot on how long I collect data? Are you looking for one or two locks that have contention a couple orders of magnitude higher than everything else? Yes, and with a large number of acquisitions. i.e. it's not usually an issue if a mutex is contended but is only acquired a few thousand times out of billions of mutex operations, but it is an issue if it's heavily used and also heavily contended. Still looks like it's a driver and/or hardware problem, but you'd need specialized knowledge to proceed with debugging it. Maybe I didn't beat on it hard enough. Data below is with two processes reading data from Ethernet and writing to disk. (common Ethernet, different disks) and a loop with 3 copies of dd writing from /dev/zero to disks, and then 3 copies of dd reading the files back and writing to /dev/null. This ground away for a few minutes. Interrupt CPU usage might be high, but the first thing you should do is retry with 6.2-rc1 and work from there. Kris pgprbmivFmi6T.pgp Description: PGP signature
Re: processes not getting fair share of available disk I/O
On Thu, Dec 14, 2006 at 04:26:18PM +, Dieter wrote: FreeBSD 6.0 Erk. How about retrying with something modern ;-) We do fix lots of bugs over time you know! In my defense, 6.0 is only one revision down. (until 6.2 comes out real soon now) Or to put it another way, your system is missing 13 1/2 months of continuous bug fixes :-D 32.3%Sys 31.5%Intr 0.0%User 0.0%Nice 36.2%Idl 710892 inact32 3: sio1 74.8%Sys 25.2%Intr 0.0%User 0.0%Nice 0.0%Idl 828676 inact36 3: sio1 Interrupt CPU usage might be high, but the first thing you should do is retry with 6.2-rc1 and work from there. Whoops, the systat -vmstat and vmstat -i were with mutex profiling enabled. Sorry about that. CPU usage is much lower with it off. 85-95% idle when writing. OK, that's better. Still, the only thing that fits is some kind of driver or hardware problem, so check 6.2-rc1 and see if it's still there. Kris pgpOBIdaTHE44.pgp Description: PGP signature
Re: cvsup and amd64
On Thu, Dec 14, 2006 at 08:18:11PM -0700, Z. Wade Hampton wrote: Greetings to all, I'm running a dell 1501 laptop with amd64x2 processor. I got a disk from a guru specifically for this architecture, booted it, and installed 6.1 via FTP. Yesterday I ran cvsup successfully. Today I did make buildworld successfully. Now, I have a little paranoia about buildkernel. Please tell me, did I need to specify anything specifically about the amd64 architecture when running cvsup? Did I possibly get a GENERIC kernel file through cvsup that is not compatible with amd? No, unless you tried hard to avoid it, the source tree you downloaded includes the amd64 code, and you'll be building an amd64 kernel from it. Kris pgpGxFfAWh0UG.pgp Description: PGP signature
Re: shmmax tops out at 2G?
On Wed, Dec 13, 2006 at 12:27:22AM -0800, Michael K. Smith - Adhost wrote: Hello Bill: -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Bill Moran Sent: Monday, December 11, 2006 2:15 PM To: [EMAIL PROTECTED] Subject: shmmax tops out at 2G? uname -a FreeBSD db00.lab00 6.2-BETA3 FreeBSD 6.2-BETA3 #1: Fri Dec 8 09:27:37 EST 2006 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/DB-2850-amd64 amd64 sysctl kern.ipc.shmmax=22 kern.ipc.shmmax: 21 - -2094967296 Looks like an unsigned 32-bit int. That doesn't seem to scale as well as would be expected on 64-bit arch. Is this a mistake, or intentional? I'm working with some big memory systems, and I sure would like to allocate more than 2G for PostgreSQL to use ... -- Bill Moran Collaborative Fusion Inc. --- This may be a silly question but, have you compiled a PAE-enabled kernel? If not, check out /sys/i386/conf/PAE. Yeah, it is ;-) PAE is a hack for legacy i386 systems which cannot run in full 64-bit (amd64) mode - it's not relevant to this problem. Bill's guess is probably right, so someone needs to go over the sysv ipc code and make it 64-bit capable. Kris pgpRyOzVYEy9s.pgp Description: PGP signature
Re: /pub/FreeBSD/ports/i386/packages-5-stable/All/
On Thu, Dec 14, 2006 at 03:44:13AM +1100, Ian Smith wrote: Hi all, FreeBSD paqi.smithi.id.au 5.5-STABLE FreeBSD 5.5-STABLE #0: Sun Nov 19 20:22:12 EST 2006 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/PAQI5S_2 i386 On 4th December, after a recent portsnap fetch/update, I ran portupgrade -anPP to prefetch all available packages for a well overdue upgrade of all ports on this box, most dating from 5.4-RELEASE CDs blush Apart from taking ~7 hours to fetch ~550MB for ~220 packages, and except for a few non-packageable ports, that went fine. Then on 10th December, after much study of UPDATING and adopting the procedures there for KDE, I ran portupgrade -aPP on those packages, which apart from updating PHP4 then installing PHP5 on top of it (which I'll take up later) went better than I'd dared to dream, taking ~8 hours. Awesome work guys! However after then running portsnap fetch/update to pick up anything new since the 4th, and after upgrading portupgrade, ran another portupgrade -anPP to pick up available packages for the ~35 ports newly out of date, intending to finish off by building any remaining ports from sources. I was glad I'd specified -PP .. every fetch from $subject directory failed. Checking manually, then and again tonight, I see that indeed only the versions of files that were (correctly) current at 4th December are still there now. The latest file date there says 17th November. Is this likely a temporary glitch, or do -stable packages only get updated to match the current ports tree after some expectable delay? There's always a lag, of course (computers aren't yet infinitely fast ;-). It's usually only a lag of a couple of days for 6.x, longer for 5.x since it's a legacy branch and not our main focus of activity. However the main FTP distribution server has been offline with hardware failure for the past week or two, so I can't push out any of the subsequent updates. Hopefully this will be resolved soon (it's also holding up the 6.2 release cycle). Kris pgpMT8zVTzbiv.pgp Description: PGP signature
Re: /pub/FreeBSD/ports/i386/packages-5-stable/All/
On Thu, Dec 14, 2006 at 07:26:42AM +1100, Ian Smith wrote: On Wed, 13 Dec 2006, Kris Kennaway wrote: On Thu, Dec 14, 2006 at 03:44:13AM +1100, Ian Smith wrote: [..] I was glad I'd specified -PP .. every fetch from $subject directory failed. Checking manually, then and again tonight, I see that indeed only the versions of files that were (correctly) current at 4th December are still there now. The latest file date there says 17th November. Is this likely a temporary glitch, or do -stable packages only get updated to match the current ports tree after some expectable delay? There's always a lag, of course (computers aren't yet infinitely fast ;-). It's usually only a lag of a couple of days for 6.x, longer for 5.x since it's a legacy branch and not our main focus of activity. As we're often enough reminded :) Thought I'd get it all up to date, then cvsup to 6.2 once released. However the main FTP distribution server has been offline with hardware failure for the past week or two, so I can't push out any of the subsequent updates. Hopefully this will be resolved soon (it's also holding up the 6.2 release cycle). Thanks Kris, may it Get Well Soon. BTW, just to try, I'd installed 6.1-R on another box over the net from the boot-only CD, and enjoyed being able to install heaps of packages from sysinstall that way, but was a bit dismayed to find it hadn't kept the fetched packages .. is there a way to ask sysinstall to do that? I dont think so, sysinstall isn't really intended as a post-install package management tool. Kris pgpX5si95BLLb.pgp Description: PGP signature
Re: processes not getting fair share of available disk I/O (was: Re: TCP parameters and interpreting tcpdump output )
On Mon, Dec 11, 2006 at 10:32:53AM +, Dieter wrote: Did this problem start before you made port2file run with rtprio? Yes. I only added rtprio because it wasn't working. Can you please include a copy of your kernel configuration file and dmesg? I think you asked that before: :-) OK, sorry - I lost track. Is Giant the only mutex/lock that could be a bottleneck across disks? The only one I can think of that is generic. One would have to do more extensive profiling and diagnosis to try and figure out what is wrong with your system. The only explanation that seems to fit is that it's something to do with your particular hardware (i.e. driver issue), since it's certainly not a problem on general configurations. I know that many people have bad things to say about nforce chipsets, although I dont know if your particular problem has been reported before. Kris pgpUCPWs2njwn.pgp Description: PGP signature
Re: processes not getting fair share of available disk I/O
On Wed, Dec 13, 2006 at 04:37:42PM +, Dieter wrote: Is Giant the only mutex/lock that could be a bottleneck across disks? The only one I can think of that is generic. One would have to do more extensive profiling and diagnosis to try and figure out what is wrong with your system. Suggestions of what to look at would be welcome. Mutex profiling would show if there is a mutex somehow getting in the way of your I/O (e.g. if Giant is somehow being forced). I dont think it would show anything though. You can try to study interrupt issues (e.g. look for an interrupt storm during I/O) with vmstat -i. Other than that you'd probably have to get your hands dirtier in the code. The only explanation that seems to fit is that it's something to do with your particular hardware (i.e. driver issue), since it's certainly not a problem on general configurations. I know that many people have bad things to say about nforce chipsets, although I dont know if your particular problem has been reported before. Could APIC have anything to do with this? It is currently turned off in firmware. Problems with interrupt delivery could certainly be relevant. Today I experimented with vfs.hirunningspace. If I crank it up, I get better total write speed with multiple drives doing dd from /dev/zero to files on disks. But it doesn't help my real applications, and in fact appears to hurt them. Yes, I don't expect there are any viable high-level workarounds for this issue at a lower layer. Kris pgpQTqXsyQfA0.pgp Description: PGP signature
Re: processes not getting fair share of available disk I/O
On Wed, Dec 13, 2006 at 10:42:03PM +, Dieter wrote: Mutex profiling would show if there is a mutex somehow getting in the way of your I/O (e.g. if Giant is somehow being forced). I dont think it would show anything though. You can try to study interrupt issues (e.g. look for an interrupt storm during I/O) with vmstat -i. Other than that you'd probably have to get your hands dirtier in the code. maxtotal count avg cnt_hold cnt_lock name 1158725 11853301596 74200 /usr/src/sys/amd64/amd64/pmap.c:1563 (pmap) 1158721 11665931596 7301 17 /usr/src/sys/amd64/amd64/pmap.c:1562 (vm page queue mutex) 90598 578551 199304 234 /usr/src/sys/kern/kern_sx.c:157 (lockbuilder mtxpool) 83234 967612 124000 700 /usr/src/sys/vm/vm_fault.c:906 (vm object) 83102 2515439 450378 500 /usr/src/sys/kern/subr_sleepqueue.c:369 (process lock) 82878 20495403215 637 1962 /usr/src/sys/kern/kern_synch.c:236 (Giant) 82632 947545 124000 704 /usr/src/sys/vm/vm_fault.c:295 (vm object) 82550 285981 124000 240 /usr/src/sys/vm/vm_fault.c:929 (process lock) 4674546789 11 425300 /usr/src/sys/kern/vfs_subr.c:1041 (vm object) 4674152927 6468110 /usr/src/sys/vm/vm_object.c:1775 (vm page queue mutex) 30068 10504612308520 /usr/src/sys/vm/vm_map.c:1380 (vm object) 24083 300793 136380 211 /usr/src/sys/vm/vm_object.c:454 (vm object) 24076329601020 /usr/src/sys/vm/vm_object.c:625 (vm page queue mutex) 19419701137295 900 /usr/src/sys/vm/vm_fault.c:787 (vm object) 160246538854941112 /usr/src/sys/vm/vnode_pager.c:1181 (vnode interlock) 16018516088791 579 /usr/src/sys/vm/vnode_pager.c:1169 (vm object) 14398 1084811 2519843 1083 /usr/src/sys/kern/kern_sysctl.c:1280 (Giant) 11940 274443 37582 701 /usr/src/sys/kern/vfs_bio.c:3082 (vm object) 11567 625811 312742 202 /usr/src/sys/kern/kern_lock.c:168 (lockbuilder mtxpool) 11096456665241 814 /usr/src/sys/vm/vm_map.c:2404 (vm object) If I'm reading the man page right, pmap holds a lock for over 1 second? In total, over 1600 operations. It's not an issue. The rest looks fine at a quick glance too. Kris pgpTmYq3vfOjT.pgp Description: PGP signature
Re: Failure to compile
On Tue, Dec 12, 2006 at 09:38:19PM -, Brian Levie wrote: I recently installed FreeBSD 6.1, and copied many Unix files with no problems at all. However when I try to run the C compiler which worked fine with Unix, I get the error message '/usr/bin/cc Exec format error Binary file not executable'. I tried changing permissions and owner with no change. Any suggestions or won't the unix C compiler not work with FreeBSD? What do you mean by Unix? Kris pgpggBWpKgyl4.pgp Description: PGP signature
Re: List Protocol (was: Major Version Upgrade 4.11 to 5.x)
On Wed, Dec 13, 2006 at 12:31:39PM +1030, Greg 'groggy' Lehey wrote: On Tuesday, 12 December 2006 at 16:49:39 -0800, Ted Mittelstaedt wrote: Until then STFU you ungrateful bastards. All you once were dumb newbies who didn't know FreeBSD from free beer, and I'll bet more than a few of you sent e-mail to questions, thinking it was an actual person who gave a damn. Boy were you surprised! Ted, there are other aspects of the list protocol. One has to do with message format. You seem to have great difficulty with this one, requiring other people to manually reformat, and often to guess what you're talking about. Another has to do with politeness. You seem to abuse this one again and again; it's one of the reasons why I seldom read this mailing list any more. You've probably driven off a number of people who would be able to give *helpful* answers. Please stop. Yeah, I've been procmailing him to /dev/null for a couple of years now. Kris pgpOBSvfEV8Ao.pgp Description: PGP signature
Re: gmake upgrade
On Sat, Dec 09, 2006 at 10:44:29PM +0100, Grzegorz Danecki wrote: Hello list! it may be a stupid question, but i'm a little bit scared before gmake upgrade. I'm not using portupgrade, everything is build from the ports tree, so, when some package is old, and there is newer version in ports, then make all make deinstall make install make clean. It works pretty fine for now. But what about gmake? Will I be able to compile new gmake sources when I'll do `make deinstall in /usr/ports/devel/gmake` first? gmake does not require gmake to build. If it did, how could you build it for the first time? Kris pgpBXouskof21.pgp Description: PGP signature
Re: processes not getting fair share of available disk I/O (was: Re: TCP parameters and interpreting tcpdump output )
On Thu, Dec 07, 2006 at 05:41:51PM +, Dieter wrote: However, I don't know what you mean by data is lost. Data should never be lost from the filesystem regardless of how slow the I/O is happening, unless there's something else going wrong (e.g. driver bug). Also, rtprio should not be used in general - see the manpage. Were you using rtprio in your original scenario? It can easily cause resource starvation. I have data arriving on Ethernet. The data rate is 2.5 MB/s max, but the other end only has a small buffer. If the BSD box doesn't read the port fast enough, the data is lost. I have a C program (port2file) reading from the port into a *large* circular buffer, currently 431,226,880 bytes. This should be enough to buffer over 2 minutes of data. It does non-blocking 64KB writes to stdout. Shell script calls this program and redirects stdout to a disk file. Very little if any other i/o to this disk. Even with disk cache in write-through mode, I can write at about 6-7 MB/s. The process needs very little CPU. Sounds like this should be no problem. And it seems to work okay if the system is otherwise idle. The problem is that if some other process is writing to some other disk, it somehow slows down writes to ALL disks. Enough that, dispite the non-blocking writes (?), the TCP receive window shrinks and shrinks and finally is smaller than a packet. The src machine obediantly stops sending packets, its small buffer fills up, and data is lost. Things I have done so far: BIG buffer (over 2 minutes worth). The port2file process cranks up the TCP receive window from 65700 to 197100. It also cranks up rtprio from 20 to 5. sysctl net.inet.tcp.delayed_ack=0 The only process running rtprio is port2file. All other processes are either default priority or niced down with the classic nice(1). Thanks for explaining the problem in more detail. Did this problem start before you made port2file run with rtprio? Can you please include a copy of your kernel configuration file and dmesg? Kris pgpbpfD9fod7H.pgp Description: PGP signature
