On 10/11/11 14:48, Eric Timberlake wrote:
Hi,
Hey, I'm looking for documentation on the AI install manifest?
Have you looked at ai_manifest(4) yet?
The following should be the equivalent of your snippet:
<!DOCTYPE auto_install SYSTEM "file:///usr/share/install/ai.dtd.1">
<auto_install>
<ai_instance name="default">
<target>
<disk>
<disk_name name="c7t0d0"/>
<partition name="2" action="delete"/>
<partition name="3" action="delete"/>
<partition name="1" action="create">
<slice name="0" action="create" force="true" in_zpool="rpool">
<size val="260gb"/>
</slice>
<slice name="1" action="delete"/>
<slice name="3" action="delete"/>
<slice name="4" action="delete"/>
<slice name="5" action="delete"/>
<slice name="6" action="create">
<size val="100mb"/>
</slice>
<slice name="7" action="create">
<size val="100mb"/>
</slice>
</partition>
</disk>
<logical>
<zpool name="rpool" is_root="true">
<filesystem name="export" mountpoint="/export"/>
<filesystem name="export/home"/>
<be name="solaris"/>
</zpool>
</logical>
</target>
<software type="IPS">
But in general, there was a flag day announced for the AI manifest
changes in build 167 and a new-to-newer.py script provided. See:
http://hub.opensolaris.org/bin/view/Project+caiman/AI_Manifest_Changes
-ethan
Here's my deal - I need for various benchmarking purposes to have a
slice (or two) left free on the boot disk for SVM to hold SVM metadb's.
Up to build 168 we would use something like: (300gb disk w/Metadb
space example given..)
<!DOCTYPE auto_install SYSTEM "file:///usr/share/auto_install/ai.dtd">
<auto_install>
<ai_instance name="default" auto_reboot="true">
<target>
<target_device>
<disk>
<disk_name name="c7t0d0" name_type="ctd"/>
<partition name="1" action="delete"/>
<partition name="2" action="delete"/>
<partition name="3" action="delete"/>
<partition name="1" action="create"/>
<slice name="0" action="create">
<size val="260gb"/>
</slice>
<slice name="1" action="delete"/>
<slice name="3" action="delete"/>
<slice name="4" action="delete"/>
<slice name="5" action="delete"/>
<slice name="6" action="create">
<size val="100mb"/>
</slice>
<slice name="7" action="create">
<size val="100mb"/>
</slice>
</disk>
</target_device>
</target>
<software>
This worked fabulously to create just what we need for our testing -
zfs on slice 0, blank partitions to house metadb's that STAY with the
boot disk. This is important because we test between versions of
solaris, and sharing metadb's between s10 and s11 is problematic.
Skip forward to the latest manifest dtd for s11 and we find the above
manifest no longer works, which seems reasonable.
What we're missing is the process to specify leaving a portion of the
boot target free as separate slices. We've tried several variations,
but there appears to be more to the 'new' AI than meets the eye. My
personal thought is that somehow the keyword "whole_disk=true" is set
during manifest creation and that's what's causing it to ignore what
we think are valid partition data that appears later on in the manifest.
Can someone point me to a documentation repository that might help us
figure out how to keep using free slices of the boot disk under the
latest release of AI? (aka build 173 or greater...) The docs I can
see are rather thin on this sort of specialized knowledge and most
likely and rightly assume that most installations will simply use the
entire boot disk for rpool.
Sorry if this is the wrong forum - I figure you folks are closer than
just about anyone to how AI works and the knowledge of what we can and
can't do with it..
Thanks so very much for your time!!
Sincerely,
-ET-
_______________________________________________
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
