[PC-BSD Testing] Support for J-Micron?

doverosx at gmail.com doverosx at gmail.com
Wed Dec 2 15:54:22 PST 2009


Jeff wrote:
> The RAID array is managed by an onboard ROM based controller, Intel 
> matrix.
>
> I previously had my system setup as RAID - 0 but it failed after a few 
> months. Assuming a drive had failed I replaced both and the array 
> broke again.  Gave up on RAID - 0.
>
> Guess I'll have to give up on eSATA for now and just use the USB 
> connection.
>
> ...Jeff
>
> --- On *Wed, 12/2/09, Josh Paetzel /<josh at tcbug.org>/* wrote:
>
>
>     From: Josh Paetzel <josh at tcbug.org>
>     Subject: Re: [PC-BSD Testing] Support for J-Micron?
>     To: "PC-BSD Testing list" <testing at lists.pcbsd.org>
>     Cc: "PC-BSD Testing list" <testing at lists.pcbsd.org>
>     Date: Wednesday, December 2, 2009, 7:21 AM
>
>
>     On Dec 2, 2009, at 1:09 AM, Jeff <dejamuse at yahoo.com
>     </mc/compose?to=dejamuse at yahoo.com>> wrote:
>
>>     I tried that at first but ad0 and ad1 are the two main drives and
>>     I'd assume they'd be the same since it's a RAID-1, mirrored.
>>
>>     Mounting either of those drives shows the contents of the system.
>>
>
>     Couple of things:
>
>     The last eSATA controller I had didn't support hot plugging in
>     FreeBSD. A drive attached when the system booted would be seen,
>     but the system would not recognize drives attached after the
>     system was booted. 
>
>     PATA supports two devices per controller, generally refered to as
>     a master and a slave. The controller cannot access both devices
>     simultaniously. In a perfect world the devices in your mirror
>     would be on seperate controllers. I don't know how smart the ar
>     driver is about reads, but theoretically having the drives on
>     seperate controllers would allow for two concurrent reads
>     simultaniously.  It could also allow for better resiliancy in the
>     case of an unclean shutdown. In your current setup, writes are
>     always serialized. If the system were to crash during a write the
>     mirror would almost certainly need to be resynced.  Anyways, you
>     might contemplate moving the drive that is the primary slave (ad1)
>     to being on the other IDE controller. It may or may not have some
>     tangible benefits, in theory it should. 
>
>     Thanks,
>
>     Josh Paetzel
>
>     -----Inline Attachment Follows-----
>
>     _______________________________________________
>     Testing mailing list
>     Testing at lists.pcbsd.org </mc/compose?to=Testing at lists.pcbsd.org>
>     http://lists.pcbsd.org/mailman/listinfo/testing
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Testing mailing list
> Testing at lists.pcbsd.org
> http://lists.pcbsd.org/mailman/listinfo/testing
>   
Unfortunately, that seems to be the case...only thing I can find is 
atacontrol..

# atacontrol detach ataX && atacontrol attach ataX


Brodey




More information about the Testing mailing list