Re: Complete hangs while extracting source
David Kirchner wrote on Sun, 02 Oct 2005 16:50: On 10/2/05, Jakob Breivik Grimstveit [EMAIL PROTECTED] wrote: How can I tell? I doubt this to be the problem, since I experience same problem on two other machines as well. Only thing they share (except same lousy administrator :-) is that they all run on AMD CPUs (1700+, 1800+ and 3500+). dmesg will show you irqs, but you can also find it using systat's vmstat mode. It'll also allow you to monitor interrupt activity. It's best used from the console of the machine. I would guess you'd want to watch for two drivers, sharing an interrupt, and both incrementing at the same rate. (I've seen that before, and it was associated with severe performance problems, but unfortunately I do not recall how it was fixed.) My machine at work (AMD1700+) has same problems - here is it's dmesg IRQs: [EMAIL PROTECTED] ~]$ dmesg | grep -i irq ioapic0 Version 0.3 irqs 0-23 on motherboard rl0: Accton MPX 5030/5038 10/100BaseTX port 0xec00-0xecff mem 0xdf00-0xdfff irq 19 at device 8.0 on pci0 uhci0: VIA 83C572 USB controller port 0xe000-0xe01f irq 21 at device 16.0 on pci0 uhci1: VIA 83C572 USB controller port 0xe400-0xe41f irq 21 at device 16.1 on pci0 uhci2: VIA 83C572 USB controller port 0xe800-0xe81f irq 21 at device 16.2 on pci0 pcm0: VIA VT8235 port 0xdc00-0xdcff irq 22 at device 17.5 on pci0 atkbd0: AT Keyboard irq 1 on atkbdc0 psm0: PS/2 Mouse irq 12 on atkbdc0 fdc0: Enhanced floppy controller at port 0x3f0-0x3f5 irq 6 drq 2 on isa0 unknown: PNP0f13 can't assign resources (irq) drm0: Matrox G550 (AGP) mem 0xdf00-0xdf7f,0xdfefc000-0xdfef,0xdc00-0xddff irq 16 at device 0.0 on pci1 While doing this: $ cd /tmp tar jxf /usr/ports/distfiles/openoffice.org2.0/OOo_1.9m130_source.tar.bz2 During it's entire run I ran this as well (my computer did nothing else during that time): while(true); do date vmstat.log vmstat -i vmstat.log sleep 1; done The logfile can be found at http://www.grimstveit.no/jakob/files/temp/vmstat.log. Are you able to parse anything usable from it? Thanks in advance. If you need anything else, please just ask - I am very eager to find the cause of these problems. -- Jakob Breivik Grimstveit, http://www.grimstveit.no/jakob/, 48298152 Besøk Newsergalleriet: http://www.newsergalleriet.no/ No violence, gentlemen -- no violence, I beg of you! Consider the furniture! -- Sherlock Holmes ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Complete hangs while extracting source
On 10/3/05, Jakob Breivik Grimstveit [EMAIL PROTECTED] wrote: The logfile can be found at http://www.grimstveit.no/jakob/files/temp/vmstat.log. Are you able to parse anything usable from it? I don't see anything wrong with the figures from that log. The only other thing I can think of to check is to ensure that DMA is enabled and supported on your drive. atacontrol cap is supposed to tell you if DMA is supported, and sysctl hw.ata will tell you if DMA is enabled in the kernel driver. I don't know the details on how to enable DMA if it is disabled, unfortunately. I would expect it would need to be something in the loader, rather than a run-time tunable. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Complete hangs while extracting source
Lowell Gilbert wrote on Sat, 01 Oct 2005 08:33: Jakob Breivik Grimstveit [EMAIL PROTECTED] writes: My entire system stops completely while unpacking large tar files, like when building OpenOffice2.0-devel. Sound goes into a loop with a timeframe of 0.001 seconds, mouse pointer stops responding, everything goes to a halt. This lasts about 0.5 seconds, then it starts working again, and then halts again and so on, until the tarfile is completely uncompressed. Is this expected behaviour? Is it a broken scheduler? Am I incuding something in the kernel which I shouldn't have? Sounds more like an interrupt issue. Yes, it does, but doing a top while extracting the Mozilla Thunderbird v1.0.6 source bz2-file (which provokes the sound and mouse jitter problem to occur) the interrupt level consistently stays below 2% of CPU. However, at times during the extraction of the file, bsdtar and bzip2 produces system CPU time of between 30% and 50%. It's during these system peaks that the mouse and sound starts being jerky. Getting any wiser with this explanation? :-) [EMAIL PROTECTED] conf]$ grep -i sched /usr/src/sys/i386/conf/NUSSE options SCHED_4BSD # 4BSD scheduler options _KPOSIX_PRIORITY_SCHEDULING # POSIX P1003_1B real-time extensions [EMAIL PROTECTED] conf]$ uname -a FreeBSD nusse.starshipping.com 5.4-RELEASE-p7 FreeBSD 5.4-RELEASE-p7 #17: Tue Sep 13 19:31:11 CEST 2005 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/NUSSE i386 Are you seeing interrupt storms? Nope. What kind of controller are you using on that hard disk [EMAIL PROTECTED] ~]$ dmesg | grep ATA atapci1: GENERIC ATA controller port 0xc800-0xc80f,0xb60-0xb63,0x960-0x967,0xbe0-0xbe3,0x9e0-0x9e7 irq 20 at device 9.0 on pci0 atapci2: GENERIC ATA controller port 0xb000-0xb00f,0xb70-0xb73,0x970-0x977,0xbf0-0xbf3,0x9f0-0x9f7 irq 23 at device 10.0 on pci0 [EMAIL PROTECTED] ~]$ dmesg | grep ad4 ad4: 190782MB ST3200822AS/3.01 [387621/16/63] at ata2-master UDMA33 Mounting root from ufs:/dev/ad4s2a Motherboard: MSI K8N NEO2 Platinum-54G S939. and is anything else sharing the same interrupt? How can I tell? I doubt this to be the problem, since I experience same problem on two other machines as well. Only thing they share (except same lousy administrator :-) is that they all run on AMD CPUs (1700+, 1800+ and 3500+). -- Jakob Breivik Grimstveit, http://www.grimstveit.no/jakob/, 48298152 Besøk Newsergalleriet: http://www.newsergalleriet.no/ It's kind of fun to do the impossible. -- Walt Disney ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Complete hangs while extracting source
On 10/2/05, Jakob Breivik Grimstveit [EMAIL PROTECTED] wrote: How can I tell? I doubt this to be the problem, since I experience same problem on two other machines as well. Only thing they share (except same lousy administrator :-) is that they all run on AMD CPUs (1700+, 1800+ and 3500+). dmesg will show you irqs, but you can also find it using systat's vmstat mode. It'll also allow you to monitor interrupt activity. It's best used from the console of the machine. I would guess you'd want to watch for two drivers, sharing an interrupt, and both incrementing at the same rate. (I've seen that before, and it was associated with severe performance problems, but unfortunately I do not recall how it was fixed.) ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Complete hangs while extracting source
Jakob Breivik Grimstveit [EMAIL PROTECTED] writes: My entire system stops completely while unpacking large tar files, like when building OpenOffice2.0-devel. Sound goes into a loop with a timeframe of 0.001 seconds, mouse pointer stops responding, everything goes to a halt. This lasts about 0.5 seconds, then it starts working again, and then halts again and so on, until the tarfile is completely uncompressed. Is this expected behaviour? Is it a broken scheduler? Am I incuding something in the kernel which I shouldn't have? Sounds more like an interrupt issue. [EMAIL PROTECTED] conf]$ grep -i sched /usr/src/sys/i386/conf/NUSSE options SCHED_4BSD # 4BSD scheduler options _KPOSIX_PRIORITY_SCHEDULING # POSIX P1003_1B real-time extensions [EMAIL PROTECTED] conf]$ uname -a FreeBSD nusse.starshipping.com 5.4-RELEASE-p7 FreeBSD 5.4-RELEASE-p7 #17: Tue Sep 13 19:31:11 CEST 2005 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/NUSSE i386 Are you seeing interrupt storms? What kind of controller are you using on that hard disk, and is anything else sharing the same interrupt? ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Complete hangs while extracting source
Lowell Gilbert wrote: Jakob Breivik Grimstveit [EMAIL PROTECTED] writes: My entire system stops completely while unpacking large tar files, like when building OpenOffice2.0-devel. Sound goes into a loop with a timeframe of 0.001 seconds, mouse pointer stops responding, everything goes to a halt. This lasts about 0.5 seconds, then it starts working again, and then halts again and so on, until the tarfile is completely uncompressed. Is this expected behaviour? Is it a broken scheduler? Am I incuding something in the kernel which I shouldn't have? Sounds more like an interrupt issue. [EMAIL PROTECTED] conf]$ grep -i sched /usr/src/sys/i386/conf/NUSSE options SCHED_4BSD # 4BSD scheduler options _KPOSIX_PRIORITY_SCHEDULING # POSIX P1003_1B real-time extensions [EMAIL PROTECTED] conf]$ uname -a FreeBSD nusse.starshipping.com 5.4-RELEASE-p7 FreeBSD 5.4-RELEASE-p7 #17: Tue Sep 13 19:31:11 CEST 2005 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/NUSSE i386 Are you seeing interrupt storms? What kind of controller are you using on that hard disk, and is anything else sharing the same interrupt? ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED] I've always had this problem (it sounds like the same thing at least) with FreeBSD, on every system I have tried it on, be it 4.x or 5.x. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
Complete hangs while extracting source
My entire system stops completely while unpacking large tar files, like when building OpenOffice2.0-devel. Sound goes into a loop with a timeframe of 0.001 seconds, mouse pointer stops responding, everything goes to a halt. This lasts about 0.5 seconds, then it starts working again, and then halts again and so on, until the tarfile is completely uncompressed. Is this expected behaviour? Is it a broken scheduler? Am I incuding something in the kernel which I shouldn't have? [EMAIL PROTECTED] conf]$ grep -i sched /usr/src/sys/i386/conf/NUSSE options SCHED_4BSD # 4BSD scheduler options _KPOSIX_PRIORITY_SCHEDULING # POSIX P1003_1B real-time extensions [EMAIL PROTECTED] conf]$ uname -a FreeBSD nusse.starshipping.com 5.4-RELEASE-p7 FreeBSD 5.4-RELEASE-p7 #17: Tue Sep 13 19:31:11 CEST 2005 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/NUSSE i386 Thanks in advance. -- Jakob Breivik Grimstveit, http://www.grimstveit.no/jakob/, 48298152 Besøk Newsergalleriet: http://www.newsergalleriet.no/ Just think -- blessed SCSI cables! Do a big enough sacrifice and create a +5 blessed SCSI cable of connectivity. -- Lionel Lauer ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to [EMAIL PROTECTED]
