[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

ACNC SATA RAID tests



Hello All,

Here is a summary of some of the ACNC SATA RAID tests. In all of these tests the firmware was never lost. Please let me know if there other things you would like tested or things that need to be clarified.

thanks...

=======

 Swapping disk between different controllers the disk configuration and
 data follow the disk. There is no data loss. All files have the same MD5
 checksums.

 If the power cables are pulled while disk i/o is happening to all the
 mounted volume the firmware is not lost. Disk i/o includes
 mkfs, fsck and making a number of files a shell script. The CPU and SCSI
 need to be power cycled for the LINUX SCSI device (e.g. /dev/sdc1
 /dev/sdd1 /dev/sde1 /dev/sdf1 /dev/sdg1 /dev/sdh1) to be manipulable. If
 there was data the checksums on the target data stayed the same o that
 is once the  machine and RAIDs are up and have been fsck'ed . An example
 of how the target data was created is below.

 A volume can not be created 2 disk set on the same controller with the
 same SCSI ID. The user interface (i.e. LCD and button) does not allow this.

 There are 2 identical ACNC SATA RAIDs on the same SCSI chain. Both are
 configured with 3 RAID0 volumes. When moving disk around in the same
 controller there was no data loss.

 When moving disk around 1 for 1 in different controller many error like
 the one below occurred on all volumes


I/O error: dev 08:41, sector 0 SCSI Error: (1:10:0) Status=02h (CHECK CONDITION) Key=5h (ILLEGAL REQUEST); FRU=00h ASC/ASCQ=00h/00h "" CDB: 08 00 00 38 08 00 - "READ(6)"

 Then the volume went into "Failed" state and the machine would not
 reboot. The volumes needed to be deleted and re-created.


=====


The target data was create with the following shell script:

#!/bin/bash
num=1
tot=1000
while test $num -le $tot
do
touch /usr/local/raid/1/raid1.$num
echo "raid1.$num got to get out of it before i get into and you got into before you get out of it" > /usr/local/raid/1/raid1.$num
touch /usr/local/raid/2/raid2.$num
echo "raid2.$num got to get out of it before i get into and you got into before you get out of it" > /usr/local/raid/2/raid2.$num
touch /usr/local/raid/3/raid3.$num
echo "raid3.$num got to get out of it before i get into and you got into before you get out of it" > /usr/local/raid/3/raid3.$num
touch /usr/local/raid/4/raid4.$num
echo "raid4.$num got to get out of it before i get into and you got into before you get out of it" > /usr/local/raid/4/raid4.$num
touch /usr/local/raid/5/raid5.$num
echo "raid5.$num got to get out of it before i get into and you got into before you get out of it" > /usr/local/raid/5/raid5.$num
touch /usr/local/raid/6/raid6.$num
echo "raid6.$num got to get out of it before i get into and you got into before you get out of it" > /usr/local/raid/6/raid6.$num
num=`expr $num + 1`
done




!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!"!
Allen Bailey
The University of Michigan            |  http://www.umich.edu/~akbailey
ITCS Contract Services/GPCC           |  fax: 734-764-8533
akbailey@xxxxxxxxx                    |  cellphone: 734-355-9332