i have my laptop hard drive set to spin down with hdparm -S 120. it works great. the problem is it spins back up after 10 minutes, and then starts the sequence all over again. how can i find out what is causing it to wake up every ten minutes without... waking it up?!

Wow.. there are many programmes that could be starting.. screensavers etc..
But I suspect you'll find it in the /etc/cron.d directory.

Most distros have a cron job that runs every 10 minutes..

There is a daemon that periodically syncs the disks. For more inspiration take a look at this

Groetjes,

Kees-Jan

excellent, thanks for the link and replies. the rundown is i tried turning off swap, but that didn't work. cron.hourly has kmod in it, which i guess is a process to remove modules that aren't being used. that seems okay, as the problem is not once per hour, but every ten minutes after the disk spins down. crontab lists a /usr/bin/run-parts command that just seems to be telling the /etc/cron.d/ stuff to run.

the update/bdflush daemon looks promising, but there is no entry that i can see in either /etc/inittab or /etc/rc.d/rc.S for update. could it be somewhere else (slack 10.1)? syslog.conf already has the - sign preceding all the entries.

regarding the init command section, i have no idea what the author is talking about when he mentions removing the sync () call and recompiling the source code. what source code? source code for init? also, that procedure sounds a little extreme, and possibly even risky. wouldn't there be a whole bunch of stuff building up in some buffer somewhere if things aren't synced? anyway, it seems like an extreme measure to take for what might not even be the problem. maybe i would consider it as a last resort.

perhaps it's something to do with X, but i don't use a screensaver and dpms shuts off the screen after 10 minutes, so it's not screensaver related at least. maybe X is writing something to the disk every 10 minutes? seems unlikely, but i guess maybe.

finally i don't use emacs, and i don't have apache running.

i wish i could report that one of those was the answer (it could still be init, i guess) but so far no success. any other ideas before i start messing with init?


>>edit: i also have noatime as an option for all my disk mounts.

I had the same problem of not being able to find the call to update. So I did some more googling and came up with this.

Init is the process that starts all others. It does so following the instructions from inittab. Apparently, it calls sync before starting a process. However, it shouldn't be starting processes all that often (depending on inittab contents)

Having said that, its actually fairly safe to remove the sync, provided your system doesn't crash or shut down. This is because data gets written to disk periodically (for example by init), or when memory becomes full. So removing the sync will cause data to be written later, but it will be written. On the other hand: before doing a shutdown, you should do a sync (which is also described in the document) because init won't be doing it for you, and you still want to have the data written to disk before the system goes down.

Groetjes,

Kees-Jan

problem solved (partially solved -- problem identified, i should say ). i believe it was in fact syslogd, even though the dash is in front of the entries in /etc/syslog.conf. i turned off that and something called pdflush, and the hard drive did not spin up again. when i looked at ps ax, pdflush was back, so it must be syslogd. i'll just have to find some other way to disable that, or change the settings or something. i really don't like the idea of not having things logged, but i also don't see any need to write to the log every ten minutes even when the computer is in standby. maybe i can find a script somewhere that turns it on and off along with dpms or some other workaround. thx for your help.