RAID0 problem array broken

[John Barbieri]

I am also having this problem under FreeBSD 5.3. IDE RAID is HPT372 and 
SATA is Silicon Image 3114

dmesg shows:

ad1: 19073MB <Maxtor 5T020H2/TAH71DP0> [38752/16/63] at ata0-slave UDMA100
acd0: CDRW <HL-DT-ST GCE-8160B/1.02> at ata1-master PIO4
ad4: 190782MB <WDC WD2000JD-00HBB0/08.02D08> [387621/16/63] at 
ata2-master SATA150
ad10: 190782MB <WDC WD2000JD-00HBB0/08.02D08> [387621/16/63] at 
ata5-master SATA150
ad12: 190782MB <WDC WD2000JB-00EVA0/15.05R15> [387621/16/63] at 
ata6-master UDMA100
ar0: ERROR - array broken
ar0: 381564MB <ATA RAID0 array> [48642/255/63] status: BROKEN subdisks:
disk0 READY on ad12 at ata6-master
disk1 DOWN no device found for this disk


First, the two SATAs are supposed to be a RAID0 array, but FreeBSD does 
not see them that way, so i have not played with any partitioning.

ad12: 190782MB <WDC WD2000JB-00EVA0/15.05R15> [387621/16/63] at 
ata6-master UDMA100<<   Thats one of the drives for the RAID, but the 
second drive is not showing up, what gives? The BIOS for the card can 
see it, but not FreeBSD.


In FreeBSD 4.x, the IDE RAID works fine, but I cant get the SATA Raid to 
work. Its always that way, I can get one work, but not the other. I need 
an OS that can do both  :(



what was the command you useed to disable the DMA? I want to give that a 
shot to.

I see that its sysctl, but what were the flags.


Thanks,
John



Matthias F. Brandstetter wrote:

>---------- quoting Karsten Fuhrmann ----------
>  
>
>>But now my ar0 stripe will not be recognised anymore.
>>/var/log/messages showed this :
>>Dec 17 03:24:44 darkstar kernel: ad6: FAILURE - ATA_IDENTIFY timed out
>>Dec 17 03:24:49 darkstar kernel: ad6: FAILURE - ATA_IDENTIFY timed out
>>Dec 17 03:24:49 darkstar kernel: ad6: deleted from ar0 disk1
>>Dec 17 03:24:49 darkstar kernel: ar0: ERROR - array broken
>>Dec 17 03:24:49 darkstar kernel: ad6: WARNING - removed from
>>configuration Dec 17 03:24:49 darkstar kernel: ar0: ERROR - array broken
>>    
>>
>
>I had a very similar problem with a RAID1 on a raid hardware controller. I 
>managed to re-build the broken array from within the controller's BIOS 
>during bootup.
>
>It seems this error comes from the broken DMA support on 5.3, so I disabled 
>DMA for my disks on this server via sysctl.
>
>Hey guys, what's the status of this DMA issue?
>
>Greetings, Matthias
>
>  
>