Peter Mansvelder wrote:
>martin b wrote:
>
>
>>#0.) My experience was SPARC only, not x86 at all
>>#1.) I didn't know what I know today in terms of,
>>whether or not the recently added (over a year ago)
>>Solaris_sparc support includes LBA48 (and therefore
>>"big disks"), or not.
>>--->> Obviously not!
>>
>>I used the opportunity to test that first, a few
>>months ago when I purchased an additional hdd
>>(Seagate 250GB, ata133), before I eventually put that
>>disk into a USB/ieee1394a combo enclosure for use via
>>ieee1394 on a sb2k (do the that [or USB2.0], if you
>>can [that is, if you don't need that drive to be
>>bootable]).
>>
>>So LBA48 does not appear to be supported on
>>[EMAIL PROTECTED]
>>Only ca. 128GB or so were recognized, no matter
>>whatever I tried inside the format utility.
>>Just in order to correct myself after circa a year ...
>>
>>
>
>I had exactly the same experience: I bought two 500GB PATA drives to attach to
>my SunBlade 150 (Sparc [EMAIL PROTECTED]); at first I tried to connect them
>through FireWire with a USB2/FW box, but whenever the drives are accessed
>simultaneously, one of them locks up, and the system responds with:
>
>Sep 2 17:09:57 sunshine scsi: [ID 193665 kern.info] sd2 at scsa13941: target
>0 lun 0
>Sep 2 17:09:57 sunshine genunix: [ID 936769 kern.info] sd2 is /[EMAIL
>PROTECTED],0/[EMAIL PROTECTED],2/[EMAIL PROTECTED]/[EMAIL PROTECTED],0
>Sep 2 17:09:57 sunshine scsi: [ID 193665 kern.info] sd3 at scsa13942: target
>0 lun 0
>Sep 2 17:09:57 sunshine genunix: [ID 936769 kern.info] sd3 is /[EMAIL
>PROTECTED],0/[EMAIL PROTECTED],2/[EMAIL PROTECTED]/[EMAIL PROTECTED],0
>Sep 2 17:12:59 sunshine scsi: [ID 107833 kern.warning] WARNING: /[EMAIL
>PROTECTED],0/[EMAIL PROTECTED],2/[EMAIL PROTECTED]/[EMAIL PROTECTED],0
> (sd3):
>Sep 2 17:12:59 sunshine SCSI transport failed: reason 'reset':
>retrying command
>
>When I connect just one, everything works fine, I can see the drive through
>'rmformat', and I can format and use them as I do right now.
>Anyway, to be able to connect both drives, I tried a PCI IDE controller, and
>looking through the forums I read good experiences with the SIL 0680 chipset,
>so I bought one of those (they are quite cheap), and took care to choose the
>non-raid version.
>Well, my experience was like yours: it recognized the drive, but it gave weird
>messages when I started 'format', about the label being corrupt; when I tried
>to label it correctly, it 'saw' no more than 127GB, the infamous LBA48 limit.
>So it seems the driver in Solaris is to blame, because the card definitely
>supports drives > 127GB, and is even ATA6 compliant.
>So no show, I was hoping to use the drives with ZFS or the 'old' Solaris SVM,
>but it seems I need either newer hardware or some other trick.
>
>
_Try_ the following:
alternative #0)
ieee1394a
--->> Don't use "rmformat", but "format -e" instead. Repartition and
re-label the disks, "newfs" ufs to them (didn't the rmformat crap force
you to use fat, not sure?). Then disable volmgt (completely or specially
for those disks). It is actually unlikely to solve your problem, given
above error msg. However, perhaps it _could_ work?
Start by trying it out, just with volmgt disabled, before repartitioning
anything.
alternative #1.)
usb2
--->>
Go and buy a cheap supported USB2 pci controller (rather than an ide
sil0680 IDE one).
Solaris' (and Linux') FireWire driver is known to have certain
limitations, i.e. dvd+rw tools/growisofs do not work at all, certain
media cannot be mounted properly ("invalid ioctl") etc.
The USB2 driver - on the other hand - does not seem to have those
problems. It perfectly behaves like ide / scsi.
alternative #2.)
Figure out how to improve the driver src.
--->>
Look into the driver and what is different to other implementations of a
similar driver for other OS's running on sparc.
~m
www.martux.org
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
