Standby not working 9.1.0.1

[dnar]

This one is odd, as I have 2 x NAS running 9.1.0.1. Standby works reliably on one box but not the other.

Both boxes are using mobo SATAII ports in AHCI mode.

The troublesome box sometimes spins the drives down but not all the time. I have tired issuing both ataidle and camcontrol commands manually, which also works sometimes.

I am using the following script on both boxes to monitor drive status. Any tips?

Code: Select all

#!/bin/sh

date
echo
echo "-------------------------------------------------------------------------------------------------------"
echo "Disk    Disk State   Temp  MB/s  Hours  Load Count  Pending  Offline   Health   Last Test Result"
echo "-------------------------------------------------------------------------------------------------------"
camcontrol devlist | awk -F\( '{print $2'} | awk -F\, '{print $1}' | grep ada|while read LINE
do
	CM=$(camcontrol cmd $LINE -a "E5 00 00 00 00 00 00 00 00 00 00 00" -r - | awk '{print $10}')
	if [ "$CM" = "FF" ] ; then
		STATE="++ SPINNING ++"
	elif [ "$CM" = "00" ] ; then
		STATE="   STANDBY    "
	else
		STATE="** UNKNOWN ** "
	fi
	smartctl -a /dev/$LINE > /tmp/$LINE.txt
	BUS=`camcontrol inquiry ada0 -R | cut -f2 -d " " | cut -b1-3`
	TEMP=`grep Temperature /tmp/$LINE.txt | cut -b88-90`
        HOURS=`grep Power_On_Hours /tmp/$LINE.txt | cut -b88-99`
	PEND=`grep Pending /tmp/$LINE.txt | cut -b88-99`
	LOAD=`grep Load_Cycle_Count /tmp/$LINE.txt | cut -b88-99`
	OFFLINE=`grep Offline_Uncorrectable /tmp/$LINE.txt | cut -b88-99`
        HEALTH=`grep overall-health /tmp/$LINE.txt | cut -b51-99`
	LAST=`grep "# 1" /tmp/$LINE.txt | grep -v newer | cut -b26-51`
	printf "%-5s %-13s %3sc %4s %7s %11s %8s %8s %8s %s %s\n" $LINE "$STATE" $TEMP $BUS $HOURS $LOAD $PEND $OFFLINE $HEALTH "  $LAST"
done
echo "-------------------------------------------------------------------------------------------------------"
echo

Script output:

Code: Select all

$ diskstat
Sun Apr 28 15:37:44 WST 2013

-------------------------------------------------------------------------------------------------------
Disk    Disk State   Temp  MB/s  Hours  Load Count  Pending  Offline   Health   Last Test Result
-------------------------------------------------------------------------------------------------------
ada0  ++ SPINNING ++  34c  300   29027      149072        0        0   PASSED   Completed without error    
ada1  ++ SPINNING ++  38c  300   28698      135455        0        0   PASSED   Completed without error    
ada2  ++ SPINNING ++  35c  300   28764      134961        0        0   PASSED   Completed without error    
ada3  ++ SPINNING ++  40c  300    8623       31721        0        0   PASSED   Completed without error    
ada4  ++ SPINNING ++  39c  300   28739      138142        0        0   PASSED   Completed without error    
ada5  ++ SPINNING ++  38c  300   20303      100642        0        0   PASSED   Completed without error    
-------------------------------------------------------------------------------------------------------

[dnar]

Well after a lot of trial and error, I think I understand the root cause of my issue. I suspect this is due to my drives being modified using wdidle3.exe to change the idle time from 8s to 300s before the heads are parked.

This does not work (command used in /etc/rc.d/ataidle at startup):

Code: Select all

ataidle -I 120 /dev/ada0

This does work, the only downside is the drives are spun down immediately at startup, but they do spin up and down as required, after 120 minutes.

Code: Select all

ataidle -S 120 /dev/ada0

Next I am going to pull the drives out and rerun wdidle to delete the idle timer (rather than set it to 300 seconds as I have). I'll post the results.

[raulfg3]

Thanks to share your experience.

[dnar]

Ok, so just to wrap this up, I have tonight rerun wdidle3.exe on all drives disabling idle. No difference.

Setting standby with "ataidle -I 120 /dev/ada0" does not work.

Setting standby with "ataidle -S120 /dev/ada0" does work.

I have now written a script to run as a PostInit script and all is well.

I wonder if this is related to my mobo SATAII ports? I'll find out soon as I have ordered a HighPoint 2720 SAS controller so I can wire up my 6 x drive caddy disk activity LED's.

[RedAntz]

Ok, so just to wrap this up, I have tonight rerun wdidle3.exe on all drives disabling idle. No difference.

Setting standby with "ataidle -I 120 /dev/ada0" does not work.

Setting standby with "ataidle -S120 /dev/ada0" does work.

I have now written a script to run as a PostInit script and all is well.

I wonder if this is related to my mobo SATAII ports? I'll find out soon as I have ordered a HighPoint 2720 SAS controller so I can wire up my 6 x drive caddy disk activity LED's.

Have you tried camcontrol ?

[jones_chris]

Nice script dnar, but if I'm reading it right, your bus speed is reading the ada0 one, for all drives.

Code: Select all

BUS=`camcontrol inquiry ada0 -R | cut -f2 -d " " | cut -b1-3`

ada0 needs changing to $LINE

[dnar]

Ok, so just to wrap this up, I have tonight rerun wdidle3.exe on all drives disabling idle. No difference.

Setting standby with "ataidle -I 120 /dev/ada0" does not work.

Setting standby with "ataidle -S120 /dev/ada0" does work.

I have now written a script to run as a PostInit script and all is well.

I wonder if this is related to my mobo SATAII ports? I'll find out soon as I have ordered a HighPoint 2720 SAS controller so I can wire up my 6 x drive caddy disk activity LED's.

Have you tried camcontrol ?

Yes, exactly the same result.

[dnar]

Nice script dnar, but if I'm reading it right, your bus speed is reading the ada0 one, for all drives.

Code: Select all

BUS=`camcontrol inquiry ada0 -R | cut -f2 -d " " | cut -b1-3`

ada0 needs changing to $LINE

Whoops! Thanks for this.

Updated script, added clear screen (printf "\f"):

Code: Select all

#!/bin/sh
printf "\f"
date
echo
echo "-------------------------------------------------------------------------------------------------------"
echo "Disk    Disk State   Temp  MB/s  Hours  Load Count  Pending  Offline   Health   Last Test Result"
echo "-------------------------------------------------------------------------------------------------------"
camcontrol devlist | awk -F\( '{print $2'} | awk -F\, '{print $1}' | grep ada|while read LINE
do
	CM=$(camcontrol cmd $LINE -a "E5 00 00 00 00 00 00 00 00 00 00 00" -r - | awk '{print $10}')
	if [ "$CM" = "FF" ] ; then
		STATE="++ SPINNING ++"
	elif [ "$CM" = "00" ] ; then
		STATE="   STANDBY    "
	else
		STATE="** UNKNOWN ** "
	fi
	smartctl -a /dev/$LINE > /tmp/$LINE.txt
	BUS=`camcontrol inquiry $LINE -R | cut -f2 -d " " | cut -b1-3`
	TEMP=`grep Temperature /tmp/$LINE.txt | cut -b88-90`
        HOURS=`grep Power_On_Hours /tmp/$LINE.txt | cut -b88-99`
	PEND=`grep Pending /tmp/$LINE.txt | cut -b88-99`
	LOAD=`grep Load_Cycle_Count /tmp/$LINE.txt | cut -b88-99`
	OFFLINE=`grep Offline_Uncorrectable /tmp/$LINE.txt | cut -b88-99`
        HEALTH=`grep overall-health /tmp/$LINE.txt | cut -b51-99`
	LAST=`grep "# 1" /tmp/$LINE.txt | grep -v newer | cut -b26-51`
	printf "%-5s %-13s %3sc %4s %7s %11s %8s %8s %8s %s %s\n" $LINE "$STATE" $TEMP $BUS $HOURS $LOAD $PEND $OFFLINE $HEALTH "  $LAST"
done
echo "-------------------------------------------------------------------------------------------------------"
echo