The "removing watch on '/dev/sda'" is my fault. I added this rule: KERNEL=="sda*", OPTIONS-="watch", to /etc/udev/rules.d/99-user.rules , which I'm seeing it doesn't have the effect I wanted. I'm running udevadm monitor right now, so if you need that information I will mail it to you later. Thanks. Tom escribió: > Kay Sievers escribió: >> On Mon, Mar 30, 2009 at 16:43, Tom <tomgparchaur@xxxxxxxxx> wrote: >>> I have a spare hard disk I don't touch the most of the time. Due to a >>> strange reason, a change uevent is triggered some random time of the day >>> (like 4 or more times) and then, of course, udev apply the rules, being >>> one of them a rule which runs vol_id which then wakes up my spinned-down >>> hard disk (debugged with echoing a 1 to /proc/sys/vm/block_dump and >>> udevadm monitor). >> Does "monitor" show a kernel event, or only a udev event? >> >> Any other events at the same time? >> > I don't remember exactly (if I remember correctly, it was one uevent > everytime). Here I attach the logs. Later I could tell you better. I > will be more aware of that. > >> Are you running smartd? > Yes, but I have smartd configured to ignore the hard disk if it's in > standby. In syslog I have: > smartd[7004]: Device: /dev/sda, is in STANDBY mode, suspending checks > > And thanks for answering so fast (and wasting your time too). > >> Kay >
Attachment:
signature.asc
Description: OpenPGP digital signature
