The reason why one would need to have the same dataset on two machine is when
you need to run a clone while the original is down, to make some tests, and 
then destroy
the clone and revert back to the original.
It's simply more work if you have to remove the dataset from the source machine 
and make
it available to the clone, then bring it back to the source when finished.
One option I see, is the clone code may clone both the zone and the delegated 
filesystems,
delegating the clones to the cloned zone.
Then, destroying the zone, should also destroy the associated cloned 
filesystems.
Gabriele.
----------------------------------------------------------------------------------
Da: Jim Klimov
A: [email protected]
Cc: Gabriele Bulfon
Data: 11 gennaio 2013 20.26.32 CET
Oggetto: Re: [discuss] Cloning a zone with same assigned dataset fail. Bug?
On 2013-01-11 16:59, Gabriele Bulfon wrote:
Now.
What was the cause of the failure?
The srczone had datasets of the global zone assigned with "add dataset".
I needed this dataset in the cloned zone, while the srczone was halted,
to make tests.
Because this dataset was present in two zones, the clone command did the
above problem!
I had to remove the dataset from the srczone before cloning, and the
clone worked as always correctly :)
I guess this is what I meant about "delegated datasets" in my additional
questions when I replied to you - where to store them, under zone root
or separately from it?
In fact, how was yours stored - from the listings in previous mails, I
do believe you kept the delegated dataset separately from the zone root
dataset hierarchy?
I think (not certain, because I mostly did these tasks of cloning data
and rewriting configs manually) that the zone-cloning would clone
datasets under the zone root - including the delegated one if it must.
I do also believe that several zones can't be natively delegated the
same dataset and/or block devices at once (lofs can be used to that
effect), so while it is arguable whether it's a bug or not that zone
cloning doesn't affect (some?) delegated datasets, it is certainly
correct not to allow both zones to access the same one (though it
may yet be incorrect or inconvenient in its ways of reporting what
it stumped on).
My2c, //Jim



-------------------------------------------
illumos-discuss
Archives: https://www.listbox.com/member/archive/182180/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182180/21175430-2e6923be
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21175430&id_secret=21175430-6a77cda4
Powered by Listbox: http://www.listbox.com

Reply via email to