Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
On Tuesday, June 12, 2001 02:00:36 AM +0200 Jens Benecke <[EMAIL PROTECTED]> wrote: > Hello, > > when working on files (i.e. having open files) on my laptop reiserfs > accesses the disk every 5 seconds. this effectively prevents the disk from > spinning down, i.e. APM modes to take over, even when I'm not writing > anything. > > Why does 'sync' write to disk even if there are no dirty buffers? I can > provoke almost continuous disk LED activity by calling Because thats how reiserfs does sync. If anyone is interested in making it more efficient, I can give pointers. You should be able to prevent bdflush/kupdate from trigger disk writes by suspending them. Andrea has patches that allow you to send both signals to make things easier on laptops. -chris
Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
Chris Mason wrote: > On Tuesday, June 12, 2001 02:00:36 AM +0200 Jens Benecke <[EMAIL PROTECTED]> wrote: > > > Hello, > > > > when working on files (i.e. having open files) on my laptop reiserfs > > accesses the disk every 5 seconds. this effectively prevents the disk from > > spinning down, i.e. APM modes to take over, even when I'm not writing > > anything. > > > > Why does 'sync' write to disk even if there are no dirty buffers? I can > > provoke almost continuous disk LED activity by calling > > Because thats how reiserfs does sync. If anyone is interested in making it > more efficient, I can give pointers. > > You should be able to prevent bdflush/kupdate from trigger disk writes by > suspending them. Andrea has patches that allow you to send both signals > to make things easier on laptops. > > -chris I do not understand. After a sync(), there is no dirty page to by synced to disk. So when the next sync happens, there should not be anything to be sent to disk? So what makes the disk-access happen? The journal? Xuân.
Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
Xuan Baldauf wrote: > > Chris Mason wrote: > > > On Tuesday, June 12, 2001 02:00:36 AM +0200 Jens Benecke <[EMAIL PROTECTED]> >wrote: > > > > > Hello, > > > > > > when working on files (i.e. having open files) on my laptop reiserfs > > > accesses the disk every 5 seconds. this effectively prevents the disk from > > > spinning down, i.e. APM modes to take over, even when I'm not writing > > > anything. > > > > > > Why does 'sync' write to disk even if there are no dirty buffers? I can > > > provoke almost continuous disk LED activity by calling > > > > Because thats how reiserfs does sync. If anyone is interested in making it > > more efficient, I can give pointers. > > > > You should be able to prevent bdflush/kupdate from trigger disk writes by > > suspending them. Andrea has patches that allow you to send both signals > > to make things easier on laptops. > > > > -chris > > I do not understand. After a sync(), there is no dirty page to by synced to disk. So >when > the next sync happens, there should not be anything to be sent to disk? So what >makes the > disk-access happen? The journal? > > Xuân. I agree with Xuan's skepticism that this is a necessary behavior rather than a bug. Please elaborate. Hans
RE: [reiserfs-list] Frequent disk accesses (sync?) on laptop
> Xuan Baldauf wrote: > > > > Chris Mason wrote: > > > > > On Tuesday, June 12, 2001 02:00:36 AM +0200 Jens Benecke > <[EMAIL PROTECTED]> wrote: > > > > > > > Hello, > > > > > > > > when working on files (i.e. having open files) on my > laptop reiserfs > > > > accesses the disk every 5 seconds. this effectively > prevents the disk from > > > > spinning down, i.e. APM modes to take over, even when > I'm not writing > > > > anything. > > > > > > > > Why does 'sync' write to disk even if there are no > dirty buffers? I can > > > > provoke almost continuous disk LED activity by calling > > > > > > Because thats how reiserfs does sync. If anyone is > interested in making it > > > more efficient, I can give pointers. > > > > > > You should be able to prevent bdflush/kupdate from > trigger disk writes by > > > suspending them. Andrea has patches that allow you to > send both signals > > > to make things easier on laptops. > > > > > > -chris > > > > I do not understand. After a sync(), there is no dirty page > to by synced to disk. So when > > the next sync happens, there should not be anything to be > sent to disk? So what makes the > > disk-access happen? The journal? > > > > Xuân. > I agree with Xuan's skepticism that this is a necessary > behavior rather than a > bug. Please elaborate. > > Hans > It's the atime update. every time you run 'sync', the sync program's atime is updated. the next sync writes this atime update, then sync gets updated again... - Brent Graveland
Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
On Fre, 15 Jun 2001, Jens Benecke wrote: > ugh... ugly. Can you mount reiserfs 'noatime', like ext2? IIUC, this would > solve my problem. sure you can! Dirk
Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
On Fri, 15 Jun 2001 at 00:57, Dirk Mueller wrote: > > Can you mount reiserfs 'noatime' > sure you can! I've been mounting my ReiserFS partitions using "noatime" for quite awhile, and everything seems to be in order. I am wondering, though, there is a mount option for ext2fs, "nodiratime", which turns off access time updates for directories. Is this a valid mount option for ReiserFS, too? While we're at this, what are the valid mount options for ReiserFS? (Maybe we could have a FAQ entry explaining these? Or is this in some document that I should read maybe in the kernel sources?) TIA! --> Jijo -- Linux, MS-DOS, and Windows NT ... ... also known as the Good, the Bad, and the Ugly
Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
Federico Sevilla III wrote: > I've been mounting my ReiserFS partitions using "noatime" for quite > awhile, and everything seems to be in order. I am wondering, though, there > is a mount option for ext2fs, "nodiratime", which turns off access time > updates for directories. Is this a valid mount option for ReiserFS, too? Both are global VFS options and are available for all filesystems. The following global options are known in 2.4.6-pre3: noexec, nosuid, nodev, sync, noatime, nodiratime > While we're at this, what are the valid mount options for ReiserFS? (Maybe > we could have a FAQ entry explaining these? Or is this in some document > that I should read maybe in the kernel sources?) For a complete list see fs/reiserfs/super.c, but I guess the official list are the one on the web site... (see configuration -> mount options) -- Henrik Nordstrom MARA Systems
Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
Federico Sevilla III writes: > On Fri, 15 Jun 2001 at 00:57, Dirk Mueller wrote: > > > Can you mount reiserfs 'noatime' > > sure you can! > > I've been mounting my ReiserFS partitions using "noatime" for quite > awhile, and everything seems to be in order. I am wondering, though, there > is a mount option for ext2fs, "nodiratime", which turns off access time > updates for directories. Is this a valid mount option for ReiserFS, too? > > While we're at this, what are the valid mount options for ReiserFS? (Maybe > we could have a FAQ entry explaining these? Or is this in some document We do: http://www.namesys.com/mount-options.html it's accessible from "Man Pages" menu option: http://www.namesys.com/manp.html > that I should read maybe in the kernel sources?) parse_options() function in fs/reiserfs/super.c :) > > TIA! > > --> Jijo Nikita. > > -- > Linux, MS-DOS, and Windows NT ... > ... also known as the Good, the Bad, and the Ugly
Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
On Saturday 16 June 2001 21:37, Hans Reiser wrote: > > It's the atime update. every time you run 'sync', the sync program's > > atime is updated. the next sync writes this atime update, then sync > > gets updated again... > > thanks brent, nikita make a faq entry out of this. I have my ReiserFS partitions mounted with noatime running 2.4.5. Every time I run sync I get disk access so atime does not explain the situation. -- http://www.coker.com.au/bonnie++/ Bonnie++ hard drive benchmark http://www.coker.com.au/postal/ Postal SMTP/POP benchmark http://www.coker.com.au/projects.html Projects I am working on http://www.coker.com.au/~russell/ My home page
Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
if you're using bash, it's writing to ~/.bash_history. On Fri, Jun 22, 2001 at 11:33:29AM +0200, Russell Coker wrote: > I have my ReiserFS partitions mounted with noatime running 2.4.5. Every > time I run sync I get disk access so atime does not explain the situation. -- Tom Vier <[EMAIL PROTECTED]> DSA Key id 0x27371A2C
Re: [reiserfs-list] Frequent disk accesses (sync?) on laptop
On Sunday 24 June 2001 19:26, Tom Vier wrote: > if you're using bash, it's writing to ~/.bash_history. Good guess but the .bash_history file is only updated on logout. See the following: rjc@lyta:~$ tail -f .bash_history & [1] 25664 fg man bon_csv2html cd cd debian/bonnie++-1.92b/ less bon_csv2html.1 bug ppp cd fg su - exit rjc@lyta:~$ sync rjc@lyta:~$ sync rjc@lyta:~$ sync rjc@lyta:~$ sync rjc@lyta:~$ sync rjc@lyta:~$ sync rjc@lyta:~$ sync > On Fri, Jun 22, 2001 at 11:33:29AM +0200, Russell Coker wrote: > > I have my ReiserFS partitions mounted with noatime running 2.4.5. > > Every time I run sync I get disk access so atime does not explain the > > situation. -- http://www.coker.com.au/bonnie++/ Bonnie++ hard drive benchmark http://www.coker.com.au/postal/ Postal SMTP/POP benchmark http://www.coker.com.au/projects.html Projects I am working on http://www.coker.com.au/~russell/ My home page