The only configuration change is that in Bootlog4 I switched the 4726
back to JBOD mode instead of RAID-1 mode. No physical cabling has been
changed.

I've attached a screenshot showing the configuration in JBOD mode from
the 3124's view.

When I rebooted this morning, there was no indication of ATA1.00 being
shown as down (attached Bootlog5.txt).

Tejun Heo wrote:
> Hello,
>
> Andrew Ryder wrote:
>   
>> Yes the issue is still there. I have also tried it with 2.6.24-rc6-git4.
>>
>> BootLog1.txt
>> 2.6.24-rc6-git4: machine booting up, drive array on before machine is
>> powered on, drive array is configured as one large disk (RAID-1).
>>
>> Bootlog2.txt
>> Same kernel as #1, system has been up for a day or two, drive array was
>> turned off after the boot failure in #1 and I turned it on again to see
>> if it made any difference.
>>
>> Bootlog3.txt
>> Booted system, drive array already on, kernel 2.6.24-rc6 and the drive
>> array was configured as one large disk (RAID-1).
>>
>> Bootlog4.txt
>> Same as #3, but drive array was configured in JBOD mode.
>>
>> If you have any not too dangerous fixes you want to test, let me know. I
>> can give them a try on my hardware here.
>>     
>
> Hmmmm... This is weird.  I have two reports on addonics reporting the
> same problem now.
>
> In boot log 1 and 3, the first fan-out port (ata1.00) is down while on 2
> and 4, the link is up.  Was there any hardware configuration change?
> 4726 is known to malfunction if the first port isn't occupied.
>
> Thanks.
>
>