Hi Trevor,
I got everything working for the most part. (thanks again) Even the
golden imaging with the automatic customizations. Since the basic
functionality is working, I do not know if this is a big issue but I
can't seem to get rid of these errors from the VDA service. (even after
reloading from scratch, once I was a little more comfortable with the
setup)
I don't think these errors should create any major problem with your
operation, though I'd agree having them fill up your log might be a bit
annoying.
This first one that happens a lot when there's activity in the pooling..
==================================================================
Feb 26, 2007 12:16:24 PM
FINER: THROW
java.lang.NullPointerException
at
com.sun.via.ClusterLocator.findNewMachine(ClusterLocator.java:24)
at com.sun.via.Locator.findAvailable(Locator.java:49)
at com.sun.via.Locator.locate(Locator.java:18)
at com.sun.via.RequestHandler.getVmIP(RequestHandler.java:82)
at com.sun.via.RequestHandler.run(RequestHandler.java:35)
at java.lang.Thread.run(Unknown Source)
==================================================================
The test environment I am using is a single host. The only way I could
get the service to load was by making a cluster of one, and using this
line in the vda.properties which points to my cluster name:
Folders.Cluster = "/vm.demo/host/x4600"
This first error occurs, as far as I can see, when there's no machine
available in the pool, i.e when all machines are currently being used,
prepared or built (from the VI client you shouldn't see any machine
directly in the pool folder, only machines in its subfolders). This
triggers this exception (we should get rid of it) but it doesn't really
change much the outcome of the operation, since even if we caught it
there still wouldn't be a machine for that user.
Doesn't seem to be a critical error but sometimes this error will appear
every couple seconds for while. Also in the VI client the error
"Recommended Power on Hosts - x4600 - The operation is not supported on
the object." comes up.
The second one seems to appear once everytime a host is assigned a VM
that is sitting in the pool that's unassigned and ready for use.
<snip stack>
That second exception is related to that VI message about the operation
not being supported. Whenever we try and power on a VM, we use a cluster
feature which is supposed to determine, presumably based on load and
other considerations, which host in the cluster is more suitable for
running the VM on. If we fail to get a recommendation, we just power on
the machine without a host suggestion, which if I'm not mistaken means
the machine is started on the host on which it's stored.
If we encounter an error while trying to get the recommendation, we log
it, and that generates that stack. It shouldn't, however, impact the
actual startup of the machine, especially considering that you only have
one host in your cluster to begin with. I'm a bit surprised, though,
that your cluster object doesn't support the command we're trying to
use, I thought this to be a basic feature of the VC. May I ask which
versions of Virtual Center and ESX you're using for your host?
Regardless of this, these errors shouldn't prevent you from working, you
may need to have a couple more machines in your pool to avoid the first
one (depends on your usage of the system), and if you're satisfied that
the system works and just want to avoid having your log fill up, you can
lower the log level to INFO for instance (in vda.properties. property
General.LogLevel), which would stop the logging of such errors. Note
that you'll need to restart the VDA service for this change to take
effect, and that you won't see other errors either after it, though.
Thanks for the feedback, we'll definitely try to get rid of that first
exception in future versions.
Cyrille
_______________________________________________
SunRay-Users mailing list
[email protected]
http://node1.filibeto.org/mailman/listinfo/sunray-users
