Hi, Alessandro
I would recommend you to read this article:
http://support.citrix.com/article/CTX122978 Note parts about "VDI remapping".
I would probably do this way: remove all snapshots to be able to coalesce
VDI-s (look for "Advanced Coalescing Scenarios" chapter). Than CS should see VM
and their disks with original UUID-s. Then you can create snapshots using CS
interface.
Vadim.
-----Original Message-----
From: Alessandro Caviglione [mailto:c.alessan...@gmail.com]
Sent: Saturday, May 23, 2015 3:36 PM
To: users@cloudstack.apache.org
Subject: A little lesson about Snapshots
...so guys, let me tell you my little adventure with snapshots.
Our environment is ACS 4.4.2 with XS6.2SP1.
Yesterday we done some activities on a customer's server, foolishly we have
XenCenter interface already opened and we took a couple of snapshot during our
activities.
At the end, we've to change the service offering of the server, so we opened
ACS web interface, stopped the instance and change service offering.
When we start again the server the instance would not start... in ACS logs we
see:
Catch Exception
com.xensource.xenapi.Types$UuidInvalid :VDI getByUuid for uuid:
xxxxxxxxx failed due to The uuid you supplied was invalid.
After google a little, we understood that we stopped the instance with a couple
of active snapshots took from XenCenter instead of ACS.
So... maybe, in ACS db there was a wrong uuid of the Root Disk of the
instance...
In fact, in "Volume" table of ACS db we see that "path" record for that
instance was different from the one we see in "xe vdi-list" in XenServer.
So, this was our situation in XenServer:
uuid ( RO) : 0345bbf9-be9c-45ae-80e7-fa5fff31625b
name-label ( RW): ROOT-136
snapshot-of ( RO): a0de7625-f98a-4443-a13f-6c9577c2d958
snapshot-time ( RO): 20150522T21:20:39Z
sm-config (MRO): vhd-parent: 9ffe5381-5a0e-4e7c-8614-e19318d9dec9
uuid ( RO) : cf50f4ba-05b9-46f9-85e4-f1ac82cdb6b4
name-label ( RW): ROOT-136
snapshot-of ( RO): a0de7625-f98a-4443-a13f-6c9577c2d958
snapshot-time ( RO): 20150522T12:49:44Z
sm-config (MRO): vhd-parent: 083364dc-b43e-401b-ab73-1959a23a9d1
uuid ( RO) : a0de7625-f98a-4443-a13f-6c9577c2d958
name-label ( RW): ROOT-136
snapshot-of ( RO): <not in database>
snapshots ( RO): 0345bbf9-be9c-45ae-80e7-fa5fff31625b;
cf50f4ba-05b9-46f9-85e4-f1ac82cdb6b4
Now, I need to understand a couple of things...
First of all... how Snapshots works??
Instinctively we change the "path" record in ACS db with the first one
(0345bbf9-be9c-45ae-80e7-fa5fff31625b), you can see that snapshot-time is later
than others.
When instance started we see that was not tha "last situation" of the
instance... so we stopped again and changed "path" record with the
"a0de7625-f98a-4443-a13f-6c9577c2d958".
So far so good, now the instance was in the last situation and all works fine!!
Now... why the last situation was in the ROOT DISK that is has not parent
snapshot?
I thought that last Snapshot ROOT DISK has the last instance situation...
Second... since now the instance is running fine but nor in ACS, nor in
XenCenter we can see any Snapshot (that in fact there are two), how can we
delete those Snapshots??
Thank you!!
