SMART Monitoring of 3Ware physical drives

[fallen2109]

I am having the following issue - I use a 3Ware 96904I4E with total of 6 physical arranged in arrays as follows:

- Array 0 (RAID 0) - 2 x WD Velociraptors as boot and swap partition on ports 0,1 respectively
- Array 1 (RAID 5) - 3 Seagate 3.0 TB Barracudas on ports 4, 5 and 6
- Single drive - as a backup drive for the most critical data on port 7

I see these arrays as 3 drives called:
- /dev/da0
- /dev/da1
- /dev/da2

When I try to use SMART however the smartd.conf gets wrong configuration for the first 2 arrays:
/dev/twa0 -a -d 3ware,0 -T permissive -s (S/../.././..) -W 25,50,60 -n standby,q -m mg.vassilev@quicknet.nl -M exec /etc/mail/smartdreport.sh -M test
/dev/twa0 -a -d 3ware,0 -T permissive -s (S/../.././..) -W 25,50,60 -n standby,q -m mg.vassilev@quicknet.nl -M exec /etc/mail/smartdreport.sh -M test
/dev/twa0 -a -d 3ware,7 -T permissive -s (S/../.././..) -W 25,50,60 -n standby,q -m mg.vassilev@quicknet.nl -M exec /etc/mail/smartdreport.sh -M test

where it should be:
/dev/twa0 -a -d 3ware,0 -T permissive -s (S/../.././..) -W 25,50,60 -n standby,q -m mg.vassilev@quicknet.nl -M exec /etc/mail/smartdreport.sh -M test
/dev/twa0 -a -d 3ware,1 -T permissive -s (S/../.././..) -W 25,50,60 -n standby,q -m mg.vassilev@quicknet.nl -M exec /etc/mail/smartdreport.sh -M test
/dev/twa0 -a -d 3ware,4 -T permissive -s (S/../.././..) -W 25,50,60 -n standby,q -m mg.vassilev@quicknet.nl -M exec /etc/mail/smartdreport.sh -M test
/dev/twa0 -a -d 3ware,5 -T permissive -s (S/../.././..) -W 25,50,60 -n standby,q -m mg.vassilev@quicknet.nl -M exec /etc/mail/smartdreport.sh -M test
/dev/twa0 -a -d 3ware,6 -T permissive -s (S/../.././..) -W 25,50,60 -n standby,q -m mg.vassilev@quicknet.nl -M exec /etc/mail/smartdreport.sh -M test
/dev/twa0 -a -d 3ware,7 -T permissive -s (S/../.././..) -W 25,50,60 -n standby,q -m mg.vassilev@quicknet.nl -M exec /etc/mail/smartdreport.sh -M test

Can someone help me figure out how to get this issue corrected? What am I doing wrong? Shall I change some part of the system configuration?

I tried to edit manually smartd.conf file, but it gets recreated on every system boot. And I do not how even to restart S.M.A.R.T. service manually in order to
test my edited smartd.conf.

Thank you in advance for your time and efforts.

[fallen2109]

Well as I did not find any other means - I modified the /etc/rc.d/smartd script to stop it from deleting and overwriting /var/etc/smartd.conf every time the smartd daemon is started .

# Clear existing file
# REMARKED by M.Vassilev
#/bin/cat /dev/null > ${smartd_conf}

and

# Write a single line to smartd_conf
# REMARKED by M.Vassilev
#echo "${_devicefilepath} ${_param}" >> ${smartd_conf}

Then I modified my /var/etc/smartd.conf as in my previous post and voila - I can see all my physical drives being monitored.

UNFORTUNATELY - this idea does not hold water as on reboot the /etc/rc.d/smartd script is OVERWRITTEN. Apparently this is mechanism to keep the integrity of the build.
Can someone please help and let me know - how am I supposed to resolve this? I do not have anymore ideas as I do not know how to tamper with the NAS4Free build .

[fallen2109]

Finally managed to come up with a workaround:

1) Created a boot directory on my data volume

2) Put the following files there
smartd (modified as above)
smartd.conf (modified as above)
smrtinit (with eXecutable rights)
Contents of the smrtinit file:

#########################################################
# smrtinit - created by M. Vassilev for customizing the SMART
# monitoring service configuration file
#########################################################
cp /mnt/NASVAS1_DATA_VOL1/boot/smartd /etc/rc.d/
cp /mnt/NASVAS1_DATA_VOL1/boot/smartd.conf /var/etc/
/etc/rc.d/smartd restart

3) Added the following Startup command via the System > Advanced > Command Scripts menu (run PostInit)
/mnt/NASVAS1_DATA_VOL1/boot/smrtinit

May be it is not perfect, but it works and most importantly - it will work with any future release of Nas4Free .

smartd daemon is enabled by allowing at least one of the logical disks to be SMART monitored in the Disks > Management menu

[sunshine]

I need the same thing - Please see viewtopic.php?f=56&t=8670

[efes]

I ended with config.xml modification with my LSI controlles. Only devicepath entry nedded to be modified for smartd to recognize the masked passX devices.

See: viewtopic.php?f=59&t=8615

[sunshine]

I ended with config.xml modification with my LSI controlles. Only devicepath entry nedded to be modified for smartd to recognize the masked passX devices.

See: viewtopic.php?f=59&t=8615

Am I correct in saying you modified the section highlighted below? If so, did that enable smart monitoring of the devices? Were you able to schedule SMART tests too?

Code: Select all

       <disks>
                <disk>
                        <uuid>fdf85bf9-c7ca-46c2-b6c1-3fc1d9cc87a3</uuid>
                        <name>mfid20</name>
                        <id>20</id>
                        <devicespecialfile>/dev/mfid20</devicespecialfile>
                        <model>n/a</model>
                        <desc>n/a</desc>
                        <type>RAID</type>
                        <serial/>
                        <size>953344MB</size>
                        <harddiskstandby>0</harddiskstandby>
                        <acoustic>0</acoustic>
                        <apm>0</apm>
                        <transfermode>auto</transfermode>
                        <fstype>zfs</fstype>
                        <controller/>
                        <controller_id/>
                        <controller_desc/>
                        <smart>
                                <devicefilepath>/dev/mfid20</devicefilepath>  ###### HERE - Change to corresponding /dev/passX device? #######
                                <devicetype/>
                                <devicetypearg/>
                                <extraoptions/>
                        </smart>
                </disk>
                .
                .
                .