Archive for December, 2012


vCOPS for Monitoring

We started “testing” vCOPS a few months back because of the projects we have going on. I am always happy with doing an install that requires me to just import an OVF and give some IPs or set to DHCP and click finish. Of course there’s some settings in the app that you have to do but that’s not out of the norm.

Now here’s where some people have a difficult time understanding. Why does VMware recommend letting it run for 21+ days before going with the stats it gathers day one? vCOPS doesn’t just pull metrics from vCenter and trust it. As an admin I would want it to bake for at the very least a month and still monitor the changes a monitoring tool tells me.

The changes in version 5.6 are improving the GUI for users and giving us more power to create groups. VMware is also giving away a foundation edition to all vSphere customers. Now yes, it is limited but it gives you the opportunity to try it out against all your hosts and VMs instead of just one host and limited number of VMs like some other vendors are doing.

If you are also a View shop, check our vCOPS for View. It gives you info from the endpoint to the host to the connection manager to the VDI. I was able to pull stats as to why a user was experiencing slowness and point it to the fat client getting packet loss.

I will go more into detail about our experience but want to give a shout out to Sean O’Dell from VMware for coming out and showing us the new features as well as giving us ideas on what else Hyperic can monitor with vCOPS.

Advertisements

Why do we at times of tragedy attack others for their opinions? Why do we not be quiet and understand someone else’s pain and grieve with them? Whether you agree or not, do not attack others with your words but show love and support.

Or stay out of it. Simple isn’t it?

Today was suppose to be an easy day of racking and setting up a Isilon in our Norman lab.  Things were going off without a problem except for the last rail, screw wouldn’t go in.  But that was an easy work around.  But back to the long subject in the post.  I needed to get our blades in our Dell M1000E up and going for some testing.  Storage isn’t ready just yet as we have been waiting on connectivity to finish up since we can’t run iSCSI on wifi.  🙂  But I did have a local 146GB drive in each blade that’s not being used since I installed ESXi onto the SD card.  Easy right?  Just create a datastore like you’ve done hundreds of times, no problems.  Or so I thought.

Going to the first blade after connecting directly to the host with the vSphere Client and selecting the LUN in the Storage tab and create Datastore.  All is going and thinking that this was going to be a homerun.  About 5 minutes after it’s spinning at “loading..” on creating the VMFS5 partition the vSphere Client just bombs with an error that I pressed OK too fast to read.  Thinking it was network related, again our connectivity team is still finishing up their part as well as getting the firewalls and DNS/DHCP stuff knocked out, we start down a path of trouble shooting the network.  We noticed some NAT translations going on that didn’t need to be after adding the hosts to our lab vCenter.  Got that resolved but the issue was still there.  Now thinking it’s the firewalls and get Juniper support on the line.  They notice some issues but nothing really related to the ESXi dropping.

Now here comes the part I SHOULD have noticed on the first crash.  The error of Call fails for “HostDatastoreSystem.QueryVmfsDatastore- CreateOptions” for object “ha-datastoresystem”  Yea.. um.. what? That wasn’t what I thought it said the first couple times it crashed.  But my stupidity of thinking it was a network problem (cause it’s always a network problem.. right storage guys? or was that “it’s always a storage problem” from the network guys.. I forget).  Simple Google search and guess what, there was some Dell partitions on that local drive.  D’oh!! Deleted those and boom, hosts with local datastores done.

Thank you Eversity Blog for typing something up 2 years ago to make me face palm myself.

OKC VMUG

Well, started thinking I need to get into the blogging world. With all the projects going on it’s really the best way to keep up with them all.

Back in November, I started up the OKC VMUG again. The group had been dead for over 2 years. Within the last 3 years I have been taking over more of a virtualization subject matter expert here at OUHSC. So I have been craving the opportunity to meet others in the same field and bounce ideas around. This is why I decided to become the leader and get the VMUG going to not just learn more myself but also help others that their job may not get to afford to send to VMworld or training. So this is where I begin in the blogging community.

20121210-191103.jpg