http://www.debian-administration.org/users/sneex/weblog

Standard:
/dev/sda:
 Timing cached reads:   5706 MB in  2.00 seconds = 2854.73 MB/sec
 Timing buffered disk reads:  1114 MB in  3.00 seconds = 370.78 MB/sec

deb:/boot# echo cfq > /sys/block/sda/queue/scheduler
deb:/boot# hdparm -tT /dev/sda | tee -a /tmp/output

/dev/sda:
 Timing cached reads:   5892 MB in  2.00 seconds = 2947.46 MB/sec
 Timing buffered disk reads:  1120 MB in  3.00 seconds = 373.00 MB/sec

echo 128 > /sys/block/sda/queue/max_sectors_kb
echo 512 > /sys/block/sda/queue/nr_requests
blockdev --setra 16384 /dev/sda

deb:/boot# hdparm -tT /dev/sda | tee -a /tmp/output

/dev/sda:
 Timing cached reads:   6184 MB in  2.00 seconds = 3094.18 MB/sec
 Timing buffered disk reads:  1732 MB in  3.00 seconds = 576.94 MB/sec


--------------------
Install web enabled monitoring software:

*important*
Newer binaries @ http://jonas.genannt.name/

wget http://ftp-master.debian-unofficial.org/other/openpgp/archive-key-2007.asc -O - | apt-key add -


# For getting 3dm2-package for 3ware
echo "deb http://ftp.debian-unofficial.org/debian etch main contrib non-free restricted" >> /etc/apt/sources.list
apt-get update
aptitude install 3ware-3dm2-binary

#In /etc/3dm2/3dm2.conf, the following defaults are set:
#  * 3DM 2 is listening on <https://localhost:888/>.
#  * Default password for both the user and the administrator is '3ware'.
#  * Remote access is disabled for security reasons (change 'RemoteAccess 0' to
#    'RemoteAcces 1' to enable it and restart the 3dm2 daemon).

---------------------------

Insert new hotspare into array


//beast> /c0 add disk=1 RaidType = spare
Creating new unit on controller /c0 ...  Done. The new unit is /c0/u1.
WARNING: This Spare unit may replace failed drive of same interface type only

//beast> show

Ctl   Model        (V)Ports  Drives   Units   NotOpt  RRate   VRate  BBU
------------------------------------------------------------------------
c0    9650SE-8LPML 8         8        2       0       3       3      -   
--------------------------------

example for automagic rebuild with hot spare:

Date: Sat, 16 Feb 2008 12:51:20 +0100
Subject: 3ware 3DM2 alert -- host: beast
20080216125114 - Controller 0
ERROR - Degraded unit: unit=0, port=1
20080216125114 - Controller 0
WARNING - Drive removed: port=1
Date: Sat, 16 Feb 2008 12:52:51 +0100                                                                                        
20080216125244 - Controller 0
INFORMATION - Rebuild started: unit=0
Date: Sat, 16 Feb 2008 16:29:08 +0100
20080216162907 - Controller 0
INFORMATION - Rebuild completed: unit=0

SO rebuild took 16:29-12:52= 3hrs37 minutes

-------------------
update firmware

//beast> update fw=/archive/software/3ware/fw3.08.03.016/prom0006.img

Warning: Updating the firmware can render the device driver and/or
management tools incompatible. Before you update the firmware,
it is recommended that you:

1) Back up your data.

2) Make sure you have a copy of the current firmware image so that
you can roll back, if necessary.

3) Close all applications.

Examining compatibility data from firmware image and /c2 ... Done.

New-Firmware        Current-Firmware    Current-Driver    Current-API
----------------------------------------------------------------------
FE9X 4.08.00.006    FE9X 4.06.00.004    2.26.02.011       2.04.00.003

Do you want to continue ? Y|N [N]: y
Downloading the firmware from file /archive/software/3ware/fw3.08.03.016/prom0006.img ... Done. 
The new image will take effect after reboot.
=======================================================================

 tw_cli /c0 show all|less


=======================================================================

#------------------------------------3ware
#miquels 3ware settings

#for deadline
# Limit queue depth somewhat
#echo 128 > /sys/block/sda/device/queue_depth
# Increase nr_requests
#echo 256 > /sys/block/sda/queue/nr_requests
# Don't use as for database-like loads
#echo deadline > /sys/block/sda/queue/scheduler
#CFQ seems to like larger nr_requests, so if you use CFQ, try 254
#(maximum hardware size) for queue_depth and 512 or 1024 for nr_requests.

#for cfq
# Limit queue depth somewhat
echo 254 > /sys/block/sda/device/queue_depth
# Increase nr_requests
echo 1024 > /sys/block/sda/queue/nr_requests
# Don't use as for database-like loads
echo cfq > /sys/block/sda/queue/scheduler
#CFQ seems to like larger nr_requests, so if you use CFQ, try 254
#(maximum hardware size) for queue_depth and 512 or 1024 for nr_requests.


echo 128 > /sys/block/sda/queue/max_sectors_kb
blockdev --setra 16384 /dev/sda
#------------------------------------/3ware