MinimServer & HDD Sleep

[MikeSpragg]

(19-04-2014 16:20)simoncn Wrote:  

(19-04-2014 12:33)MikeSpragg Wrote:  Hi Simon, think that's a red-herring - it appears to be a log of disk usage (the QNAP records disk usage) in :

[/] # find / -name vg*.log
/mnt/HDA_ROOT/.config/storage_usage_history/vg1.log
/mnt/HDA_ROOT/.config/storage_usage_history/vg2.log
[/] # cd /mnt/HDA_ROOT/.config/storage_usage_history
[/mnt/HDA_ROOT/.config/storage_usage_history] # ls -l
-rw-r--r-- 1 admin administ 16686 Apr 19 01:40 CACHEDEV2_DATA.log
-rw-r--r-- 1 admin administ 16046 Apr 19 01:40 vg1.log
-rw-r--r-- 1 admin administ 16634 Apr 19 01:40 vg2.log

Which, if the disk is being written to - would explain why this is being logged (as it's usage).

Surely this will prevent the disk from spinning down, as writing the disk update history to the disk will spin up the disk, and this disk I/O to the logs will trigger writing a further history update after 10 minutes, etc., etc.

I just rechecked the log1.log and log2.log files that you produced yesterday to see if the same thing was happening. The log2.log file (wthout MinimServer running) was truncated by Ctrl-C at the 9/100 stage, but the log1.log file was allowed to terminate normally after completing the final 99/100 test. This could explain why log2.log is much smaller than log1.log.

Not quite. It only writes the usage log after something has made it write to the disk (no write, no log).

It's just something (whilst Minim/JRE) is running that is doing this (that's why I say it's a red herring) write to disk (and then the usage is also updated).

Also, I stopped it at 9/100 as, because, it had already gotten to the half hour mark (with log2) - I waited 30 mins and there was nothing else happening. The ctrl+c was at the 30 minute marker (20:53), not 20:30 as it would appear from the log. The monitor program only writes to the log if there is activity.

With log1 it exceeded the 100 test within 30 mins. In other words - both logs are for the same period (30 mins) and log1 exceeds the test capacity of 100 tests within that period - and log2 only gets to 9/100.

In reality, these aren't "tests" at all - they are reports of disk interaction reported from the OS and then picked up by the monitor. So, they only pickup the activations as they occur.

As per readme.txt>>

Log2.log
--------

Now, with JRE/MinimServer both switched off. As you can see - a much quieter log, only 9 disk activations (at start) for the same period
and, also, disks spun down twice during the test. On the last spin down at 20:35 they stayed spun down.


20:28 - disks shutdown - Status LED Off - at 20:30 they restart (reasons unknown, spurious as.... next)
20:35 - disks shutdown - Status LED Off - and stay shutdown until end of test period (20:52) ....

-----------