How to reduce unnecessary disk activity?

[pags]

Yes, I did reboot completely.  As of now, I have also had the chance to test same on a couple of other machines - still no difference; the disk light still flashes on a freshly-booted-but-otherwise-untouched machine, and gkrellm indicates persistent activity also.  In comparing between several different installation layouts, it seems to me as though the activities are likely accessing something in the /var partition.  Could it be related to system monitoring, or logging functions?

Additional observation:  my original EeePC does not suffer from such behaviour.  It is an old install of Minime 2008.1 with KDE 3.5.9 that has not been updated in quite a while.  It is partitioned as follows:

sda1 --- /
sdb1 --- /tmp
sdb2 --- /var
sdb3 --- /home

With all partitions being primary, formatted with ReiserFS, and also with notail,noatime set in fstab.  The internal 4GB SSD is /dev/sda, whereas /dev/sdb is an 8GB SDHC card.

What does

Code: [Select]

ps aux | grep haldreturn?

I would think you'd see something similar to:

root      4754  0.0  0.0   5424  1836 ?        S    14:05   0:01 hald-addon-storage: polling /dev/sr0 (every 2 sec)

amongst the returned items.  If /dev/sr0 is replaced with the device that is your SD slot, then it would appear you need to disable the polling.

The only likely candidate I see in Services is the haldaemon, but it might handle more than just removable media (someone else with more authority may know better).  I suppose you could try just stopping the service, without changing the startup status, so if it did break, a reboot should get you back up and running

[Was_Just19]

Code: [Select]

root      3217  0.0  0.0   5424  1832 ?        S    20:05   0:00 hald-addon-storage: no polling on /dev/hda because it is explicitly disabled
root      3240  0.0  0.0   5424  1836 ?        S    20:05   0:00 hald-addon-storage: polling /dev/sdc (every 2 sec)
root      3241  0.0  0.0   5424  1836 ?        S    20:05   0:00 hald-addon-storage: polling /dev/sdd (every 2 sec)
root      3242  0.0  0.0   5424  1840 ?        S    20:05   0:00 hald-addon-storage: polling /dev/sdf (every 2 sec)
root      3243  0.0  0.0   5424  1836 ?        S    20:05   0:00 hald-addon-storage: polling /dev/sr1 (every 2 sec)
root      3244  0.0  0.0   5424  1840 ?        S    20:05   0:00 hald-addon-storage: polling /dev/sde (every 2 sec)
root      3245  0.0  0.0   5424  1840 ?        S    20:05   0:00 hald-addon-storage: polling /dev/sr0 (every 2 sec)

hda is a Zip drive
sdc and up are connected via USB ..... sdc to sdf a 4 port hub
sr0 & sr1  two optical drives ..... one an internal SATA and the second a USB connected drive
Neither of the two internal SATA HDDs are in the list .... sda & sdb

Just posted for comparison   

[pags]

JohnBoy,
Did you change a configuration to enable:

no polling on /dev/hda because it is explicitly disabled

or was that handled entirely by the system?

Just curious.

[Was_Just19]

JohnBoy,
Did you change a configuration to enable:

no polling on /dev/hda because it is explicitly disabled

or was that handled entirely by the system?

Just curious.

No I did not change anything ..... I suspect it may be the drive itself or a setting for Zip drives in the system.

[vc]

What does

Code: [Select]

ps aux | grep haldreturn?

It returns (specifically):

root      2250  0.0  0.0   5428  1836 ?        S    18:48   0:00 hald-addon-storage: polling /dev/sdb (every 2 sec)

/dev/sdb1 is for /home - it is primary, formatted ReiserFS, with notail,noatime in /etc/fstab.

I would think you'd see something similar to:

root      4754  0.0  0.0   5424  1836 ?        S    14:05   0:01 hald-addon-storage: polling /dev/sr0 (every 2 sec)

amongst the returned items.  If /dev/sr0 is replaced with the device that is your SD slot, then it would appear you need to disable the polling.

The only likely candidate I see in Services is the haldaemon, but it might handle more than just removable media (someone else with more authority may know better).  I suppose you could try just stopping the service, without changing the startup status, so if it did break, a reboot should get you back up and running

Perhaps... I should examine it more closely, I suppose.  What .rc files are specifically associated with hald polling, I wonder?