Re: Identifying disk activity
At 07:30 28/08/2011, Polytropon wrote: Since I have installed my new system (FreeBSD/i386 8.2-STABLE), I have found some kind of disk activity I've never had before on my home system. As this PC is a very cheap product, it doesn't have a HDD LED. Instead I have to listen to the disk. This is the strange sound: four groups of short bt sounds within a second, with a short pause between them. #-#-#-#- = 1 s This can be heared over several seconds, then silence. From time to time, a brrrt sound appears for 3 seconds in one long rush. Your disk is thinking or trying to do something. In morse code -- is 'm' so -- -- -- -- is mmm. When you hear the 'h' (four dots) or 'ph' (.--.) it's finishing thinking. A typical sesion will be: mmhhh or p. Similar sounds are made by lot of people in the bathroom while poping, you may want to investigate perhaps beastie is poping there. HTH Couldn't resist ;D ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Identifying disk activity
On Sat, 27 Aug 2011 23:54:32 -0600 (MDT), Warren Block wrote: On Sun, 28 Aug 2011, Polytropon wrote: Since I have installed my new system (FreeBSD/i386 8.2-STABLE), I have found some kind of disk activity I've never had before on my home system. As this PC is a very cheap product, it doesn't have a HDD LED. Instead I have to listen to the disk. This is the strange sound: four groups of short bt sounds within a second, with a short pause between them. #-#-#-#- = 1 s This can be heared over several seconds, then silence. From time to time, a brrrt sound appears for 3 seconds in one long rush. That could be a t-cal, thermal calibration. Depends on the age and model of the drive, some drives don't do it. Could be other internal drive activity. WD drives like to park heads often, loudly, and for no good reason. I didn't mention it's a WD, how did you guess? :-) % dmesg | grep ^ad ad4: 305245MB WDC WD3200AAJS-56M0A0 01.03E01 at ata2-master UDMA100 SATA 1.5Gb/s Strangely, I didn't hear that sound when the system was running AMD64, it seems to have appeared on i386 and _sometimes_ in conjunction with Opera. I have removed the side panel of the PC box, but the sound still sometimes can be heared. The disk doesn't become spectacularly hot, but thermal problems are quite common to modern home consumer hardware - and this special PC seems to be _very_ modern. I have used _many_ WD drives in the past, but none of them has even made such sounds. Seems to be a modern feature. Is there a way to force synchronous disk activity? Turning off soft updates will help, but not make disk writes totally synchronous. Can this be done easily (e. g. tunefs -n disable devices in SUM to the unmounted partitions), without data loss? I have to admit that I've never tried that before. -- Polytropon Magdeburg, Germany Happy FreeBSD user since 4.0 Andra moi ennepe, Mousa, ... ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Identifying disk activity
On Sun, Aug 28, 2011 at 12:30 AM, Polytropon free...@edvax.de wrote: I am not sure this is related to a program, but I'd like to find it out. As FreeBSD's I/O subsystem does not work in real-time, I cannot conclude from actual program file I/O to physical disk I/O. Is there a way to force synchronous disk activity? well there is the mount -o sync option. My idea is to watch open files and running programs as precise as possible (as root: top -St -s 0). Which tools (e. g. top, htop, lsof) would you suggest to narrow down _which_ program is accessing _which_ file, causing the sound? top -Hm io has revealed pesky apps in the past for me without using mount options. IIRC, you have to use procstat to find if something is using swap. -- Adam Vande More ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Identifying disk activity
Date: Sat, 27 Aug 2011 23:54:32 -0600 (MDT) From: Warren Block wbl...@wonkity.com Subject: Re: Identifying disk activity Is there a way to force synchronous disk activity? Turning off soft updates will help, but not make disk writes totally synchronous. 'mount -o sync' does, however. grin ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Identifying disk activity
On Sun, 28 Aug 2011, Polytropon wrote: Is there a way to force synchronous disk activity? Turning off soft updates will help, but not make disk writes totally synchronous. Can this be done easily (e. g. tunefs -n disable devices in SUM to the unmounted partitions), without data loss? I have to admit that I've never tried that before. SUM isn't needed, just have the filesystem unmounted. Never had data loss from it. Disabling write caching will also help to make filesystem writes more synchronous. There are several sysctls and some information on soft updates at http://www.freebsd.org/doc/handbook/configtuning-disk.html. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Identifying disk activity
Since I have installed my new system (FreeBSD/i386 8.2-STABLE), I have found some kind of disk activity I've never had before on my home system. As this PC is a very cheap product, it doesn't have a HDD LED. Instead I have to listen to the disk. This is the strange sound: four groups of short bt sounds within a second, with a short pause between them. #-#-#-#- = 1 s This can be heared over several seconds, then silence. From time to time, a brrrt sound appears for 3 seconds in one long rush. I am not sure this is related to a program, but I'd like to find it out. As FreeBSD's I/O subsystem does not work in real-time, I cannot conclude from actual program file I/O to physical disk I/O. Is there a way to force synchronous disk activity? I don't mind if this makes the system run slower, because it's just for diagnostics, but I'd like file I/O done by a program to cause immediate disk activity. My idea is to watch open files and running programs as precise as possible (as root: top -St -s 0). Which tools (e. g. top, htop, lsof) would you suggest to narrow down _which_ program is accessing _which_ file, causing the sound? I already do suspect Opera (due to opera-linuxplugins-11.50, linux-f10-flashplugin-10.3r183.5, nspluginwrapper-1.4.4 maybe), but I'd like to know _where_ exactly this strange sound came from, as there is no 1:1 relation (running Opera does not imply the sound to appear). I have already checked smartctl -a /dev/ad4 which doesn't show any malicious behaviour of the disk itself (sometimes also the reason for strange sounds). -- Polytropon Magdeburg, Germany Happy FreeBSD user since 4.0 Andra moi ennepe, Mousa, ... ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
Re: Identifying disk activity
On Sun, 28 Aug 2011, Polytropon wrote: Since I have installed my new system (FreeBSD/i386 8.2-STABLE), I have found some kind of disk activity I've never had before on my home system. As this PC is a very cheap product, it doesn't have a HDD LED. Instead I have to listen to the disk. This is the strange sound: four groups of short bt sounds within a second, with a short pause between them. #-#-#-#- = 1 s This can be heared over several seconds, then silence. From time to time, a brrrt sound appears for 3 seconds in one long rush. That could be a t-cal, thermal calibration. Depends on the age and model of the drive, some drives don't do it. Could be other internal drive activity. WD drives like to park heads often, loudly, and for no good reason. Is there a way to force synchronous disk activity? Turning off soft updates will help, but not make disk writes totally synchronous. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org
