Re: Bug madness: High frequency of load/unload cycles
hi, Am Montag, den 12.05.2008, 23:14 +0200 schrieb Milan Bouchet-Valat: > I'd like to raise the developers' awareness about bug 59695 [1]: High > frequency of load/unload cycles on some hard disks may shorten lifetime > please see this article: http://lwn.net/Articles/257426/ it is not aything ubuntu specific but a hardware vendor setting, note as well that the relatime mount option is the default with hardys 2.6.24 kernel. ciao oli signature.asc Description: Dies ist ein digital signierter Nachrichtenteil -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Bug madness: High frequency of load/unload cycles
Le mardi 13 mai 2008 à 11:22 +0200, Oliver Grawert a écrit : > hi, > Am Montag, den 12.05.2008, 23:14 +0200 schrieb Milan Bouchet-Valat: > > I'd like to raise the developers' awareness about bug 59695 [1]: High > > frequency of load/unload cycles on some hard disks may shorten lifetime > > > please see this article: > http://lwn.net/Articles/257426/ > it is not aything ubuntu specific but a hardware vendor setting, note as > well that the relatime mount option is the default with hardys 2.6.24 > kernel. If you think it's been solved with the relatime option in Hardy, then please go there and explain it to the subscribers, and set the bug to Invalid. Don't let people talk again and again about these issues. But AFAIK the problem is still here, and some users are returning to Windows because they say then their Hard Cycle Count is not rising so quickly - whether this can be Ubuntu's fault or just a result a kind of bug in Windows, I cannot tell. Anyway, if it is real that on Windows heads don't park, and on Linux they do, so no matter of the cause, we should at least provide an easy workaround. Can you claim there is nothing in this bug but nonsensical fear from users? I'm not taking side here, I can be sure of anything. What I know is that we should tell them if something has to be done, will be done, and why. Thanks for you work -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Bug madness: High frequency of load/unload cycles
How about a script that does some sort of binary search for the ideal amount of time? So it would start off by setting `hdparm -B 128`, then wait 10 minutes, check whether there's been more than $CYCLE_MAX cycles, do `hdparm -B 64` or -B 192, then start over again. If the script finished with a value of 255, it could then warn the user that there was something seriously wrong with their HDD. With a range of 0..254, you could find a good number this way in 80 minutes. This is obviously just a more rigorous version of the workaround, but if the workaround is what people trust, it does no harm to do it right. - Andrew -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Bug madness: High frequency of load/unload cycles
hi, Am Dienstag, den 13.05.2008, 22:42 +0200 schrieb Milan Bouchet-Valat: > Le mardi 13 mai 2008 à 11:22 +0200, Oliver Grawert a écrit : > > hi, > > Am Montag, den 12.05.2008, 23:14 +0200 schrieb Milan Bouchet-Valat: > > > I'd like to raise the developers' awareness about bug 59695 [1]: High > > > frequency of load/unload cycles on some hard disks may shorten lifetime > > > > > please see this article: > > http://lwn.net/Articles/257426/ > > it is not aything ubuntu specific but a hardware vendor setting, note as > > well that the relatime mount option is the default with hardys 2.6.24 > > kernel. > If you think it's been solved with the relatime option in Hardy, then > please go there and explain it to the subscribers, and set the bug to > Invalid. Don't let people talk again and again about these issues. please read the article again... relatime might help as part of a workaround but the article makes pretty clear that there is no sane way you can solve it on the software side in a nonintrusive way, its a hardware vendor setting and hardware vendors need to fix it. ciao oli signature.asc Description: Dies ist ein digital signierter Nachrichtenteil -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Bug madness: High frequency of load/unload cycles
On Wed, 2008-05-14 at 07:58 +0200, Oliver Grawert wrote: > please read the article again... relatime might help as part of a > workaround but the article makes pretty clear that there is no sane way > you can solve it on the software side in a nonintrusive way, its a > hardware vendor setting and hardware vendors need to fix it. Could a list of vendors who make crappy hard drives (ie ones with this issue) be made so we all can avoid them? I can say my Western Digitals don't have the issue, though they do have a tendency to die anyway (bad sectors and dead circuitry, no clicking or oddly high disk i/o). -- Mackenzie Morgan http://ubuntulinuxtipstricks.blogspot.com apt-get moo signature.asc Description: This is a digitally signed message part -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Bug madness: High frequency of load/unload cycles
ke, 2008-05-14 kello 02:11 -0400, Mackenzie Morgan kirjoitti: > Could a list of vendors who make crappy hard drives (ie ones with this > issue) be made so we all can avoid them? I can say my Western Digitals > don't have the issue, though they do have a tendency to die anyway (bad > sectors and dead circuitry, no clicking or oddly high disk i/o). Don't all vendors occasionally make bad hardware? I know that I've had bad disks from IBM/Hitachi, Seagate, WD, and Maxtor, at least. (Not all of them laptop disks, though.) (My philosophical stance is that all hardware is not just bad, but evil, and actively designed to make your life a misery, but let's ignore that, for now. My question above is serious.) -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Bug madness: High frequency of load/unload cycles
On Wed, 2008-05-14 at 10:11 +0300, Lars Wirzenius wrote: > ke, 2008-05-14 kello 02:11 -0400, Mackenzie Morgan kirjoitti: > > Could a list of vendors who make crappy hard drives (ie ones with this > > issue) be made so we all can avoid them? I can say my Western Digitals > > don't have the issue, though they do have a tendency to die anyway (bad > > sectors and dead circuitry, no clicking or oddly high disk i/o). > > Don't all vendors occasionally make bad hardware? I know that I've had > bad disks from IBM/Hitachi, Seagate, WD, and Maxtor, at least. (Not all > of them laptop disks, though.) I've had 2 Western Digital bricks since Nov and 2 Western Digital bad-enough-to-corrupt-things-and-crash-the-system drives since January. One of the bricks was a year old, the other 6 months old. One misbehaving drive is 1.5 years old. The other was misbehaving from day 1. -- Mackenzie Morgan http://ubuntulinuxtipstricks.blogspot.com apt-get moo signature.asc Description: This is a digitally signed message part -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Bug madness: High frequency of load/unload cycles
Il giorno Mon, 12 May 2008 23:14:33 +0200 Milan Bouchet-Valat <[EMAIL PROTECTED]> ha scritto: > But it has not fixed all issues. And it appears to cause some HD to > heat, which can possibly be dangerous. But this may just be an > illusion. This is not an illusion: my laptop hangs after a couple of hours of work if I set the disk that way. I tend to consider this more bad design of my laptop than a software bug, but it seems that there is no static hdparm -B value that fits me (load cycles increases 100 times per hour if I set it to lower values). I still see a solution: the hdparm setting should be dynamic and influenced by both hddtemp results and the load_cycles count. There should be a daemon which monitors the derivatives of load_cycles and the drive temperature and sets the -B flag of hdparm to 255 whenever load_cycles is increasing too fast, and to lower values whenever the hard disk temperature is increasing too. Vincenzo -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Bug madness: High frequency of load/unload cycles
Hi Vincenzo Ciancia wrote: > I still see a solution: the hdparm At best that seems like a workaround. I'm pretty sure I saw Matthew Garrett say that the underlying cause of the high parking rate is that Linux's disk IO pattern is quite different to that of Windows, and that we ought to be able to change that so the disk is touched less frequently. Cheers, -- Chris Jones -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss
Re: Bug madness: High frequency of load/unload cycles
Olá Vincenzo e a todos. On Thursday 16 October 2008 12:04:35 Vincenzo Ciancia wrote: > Il giorno Mon, 12 May 2008 23:14:33 +0200 > Milan Bouchet-Valat <[EMAIL PROTECTED]> ha scritto: > > > But it has not fixed all issues. And it appears to cause some HD to > > heat, which can possibly be dangerous. But this may just be an > > illusion. > > > This is not an illusion: my laptop hangs after a couple of hours of work if I > set the disk that way. Eversince hdparm changed it default value to 254/255 my disk is much hotter. I tried different values for it, but anything bellow 127 makes my disk start to many times. I have records on my syslog of HD temp every 10 minutes, if anyone is interested in checking. -- BUGabundo :o) (``-_-´´) http://LinuxNoDEI.BUGabundo.net Linux user #443786GPG key 1024D/A1784EBB My new micro-blog @ http://BUGabundo.net ps. My emails tend to sound authority and aggressive. I'm sorry in advance. I'll try to be more assertive as time goes by... signature.asc Description: This is a digitally signed message part. -- Ubuntu-devel-discuss mailing list Ubuntu-devel-discuss@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss