Paul,
A quick glance through the sources indicates that the topology labels
have only been shipped for 4170 M3, not M2. I'm not enough of an expert
on this to know whether there's a workaround that can be applied; I can
guess at a couple though they'd require customizing the AI boot image.
I'd suggest filing a bug against croinfo (solaris/fma/io) to get this
evaluated.
I know we have issues with identifying boot_disk in some cases. That
needs to be looked into further, though I think in some cases we're
victims of firmware there.
You're right that derived manifests would be one way to work around the
problem.
Dave
On 10/21/11 14:01, Paul de Nijs wrote:
All,
And then you would THINK that this croinfo stuff is usable on x86 (in
this case a x4170M2)
root@x4170-220:/# smbios -t SMB_TYPE_SYSTEM
ID SIZE TYPE
1 146 SMB_TYPE_SYSTEM (system information)
Manufacturer: SUN MICROSYSTEMS
Product: SUN FIRE X4170 M2 SERVER
Version:
Serial Number: 1136FMM10J
UUID: ff200008-ffff-ffff-ffff-aa98d7282100
Wake-Up Event: 0x6 (power switch)
SKU Number: 30056022+1+1
Family:
root@x4170-220:/# /usr/sbin/croinfo -O cAR
root@x4170-220:/#
Well, that doesn't help much, does it ?
On the other hand, smbios sees something ....
root@x4170-220:/# smbios | grep HDD
External Reference Designator: /SYS/DBP/HDD0
External Reference Designator: /SYS/DBP/HDD1
External Reference Designator: /SYS/DBP/HDD2
External Reference Designator: /SYS/DBP/HDD3
External Reference Designator: /SYS/DBP/HDD4
External Reference Designator: /SYS/DBP/HDD5
root@x4170-220:/#
You would think that prtpicl/prtconf would do it ....
root@x4170-220:/# prtconf -v | grep HDD
root@x4170-220:/#
Bummer!
and prtdiag -v ???
No luck either.
And if you use the keyword "boot_disk" in your AI manifest and you have
more than one disk, it can't find it.
Great, now we have to "manually" give the boot disk a volname, and use
that in your ai_manifest.
Are there other options ? since the disks are really the WWN:
root@x4170-220:/# echo "" | format
Searching for disks...done
c0t5000C5003C157A4Bd0: configured with capacity of 279.38GB
c0t5000C5003C157DB3d0: configured with capacity of 279.38GB
AVAILABLE DISK SELECTIONS:
0. c0t5000C5003C157A4Bd0 <SEAGATE-ST930003SSUN300G-0B70 cyl 46872 alt 2
hd 20 sec 625>
/scsi_vhci/disk@g5000c5003c157a4b
1. c0t5000C5003C157DB3d0 <SEAGATE-ST930003SSUN300G-0B70 cyl 46873 alt 2
hd 20 sec 625>
/scsi_vhci/disk@g5000c5003c157db3
Specify disk (enter its number): Specify disk (enter its number):
root@x4170-220:/#
Which means every x4170M2 will need it's own ai_manifest.
I guess we have to use the 'derived manifest' for these systems ?
Paul
On 10/20/11 05:06 PM, Brian Phipps wrote:
Hi All,
Currently, in one of my manifests I'm using the following target
definition:
<disk whole_disk="true">
<disk_name name="/SYS/MB/HDD0" name_type="receptacle"/>
<slice name="0" in_zpool="rpool" action="create" force="true">
<size val="279gb"/>
</slice>
<slice name="1" action="create" force="true"/>
</disk>
My question is, can I use something more generic than "/SYS/MB/HDD0"
for the receptacle? Maybe something like "HDD0" since I'm finding that
the preceding path changes on differing machine types. Some have
/SYS/HDD0 and some have /SYS/MB/HDD0.
- Brian
Oracle <http://www.oracle.com/>
Brian Phipps | Software Engineer | +1.503.495.7716
OracleWorld Wide Operations
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
--
Paul de Nijs | Principal Software Engineer | Performance Technologies |
+1.503.495.7882
Oracle Strategic Applications Engineering (SAE)
3295 NW 211th Terrace | Hillsboro, OR 97124-7110
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
