Some questions after an hour of use...

[HenrikElm]

Am I the first one? Ok! I have installed Chargeback and started some testing.

So.. All this looks very good and handy I must say! But of course there are questions:

1) Does it not use ODBC? It looks like it does not (to talk to the chargeback DB)?

2) Can it use Windows integrated SQL authentication? I could not get past the install without turning on SQL auth on the SQL Server and using a SQL account! The same goes for configuring the Chargeback connection to the VC DB. How can this be? I hope I am wrong here?

3) Is there a way to automatically add a VM (or host) into a chargeback hiearchy folder? Or is it always a manual thing to drag/drop newly created VMs in there? It would be great if there was an automated way, say for folders or resource pools to have a "twin" in the chargeback hierarchy where it gets added.

4) Can you specify start and end times of the day, or does it always calculate whole days? (Maybe want to charge parts of days).

5) Is it ok to use SQL 2005 Express in production for the CB DB? The DB seem to get quite small?

6) If I choose to not replicate the VC Data to the CB DB, will CB work online and create load on the VC DB Server? (But also save me from some diskspace usage)?

I guess I will find more as I work through the product, but it seems easy enough to use and it looks very very promising for us to use at work! Good job!

/Henrik

[admin]

1) It uses JDBC

2) version 1.0 does not support windows authentication

3) Yes, You can have a twin for any vCenter Server entity in chargeback. Once you drag drop a folder or resourcepool or any other vCenter entity, any of its descendents automatically gets added / removed / synchronized.

Let's say you added a host H1 from vCenter Server to chargeback hierarchy . Any new VM that gets created under this host will automatically appear in chargeback hierarchy under this host.

Also, you can choose to have entire vCenter hierarchy as chargeback hierarchy. You get this option while creating a chargeback hierarchy

4) Our user interface does not allow you to specify time of the day. It always takes entire day.

5) SQL 2005 Express is not a supported database management system. We do not recommend it for production use. You can get a list of supported systems in the release notes.

6) If you choose not to replicate , chargeback will go to VC DB for information. This is more work for VC database. But yes, it does save you disk space

Finally, Thanks for appreciation.

-naveen

[HenrikElm]

1) OK

2) OK

3) But I added two hosts and no VMs got billed? I created a folder of its own in the CB hierarchy and added the VMs and could set other rules for them. I liked it that way. It is very possible that you want different billing for hosts and its VMs so I hope that my findings are right? Are you 100% sure a VM gets added to the hierarchy when the host is added? A folder or a Respool I can accept, but a host= Does not make all that sense?

Also, when I set billing policy on a CB folder, it is also applied to the children in it. This is good. If I set different policy on a, say, single VM, will it be billed twice? Once for the settings from the folder, once for the unique settings for the VM? Earliy testing gives that this is true, but I haven't had the time to test it too much.

4) OK

5) OK. I will use Express during the eval, it works quite fine. Is there an automatic mechanism to clean/purge older data from the CB DB? Like in the VC DB where you set it under "Statistics" settings? Would be great to keep it from growing out of control?

6) OK

/Henrik

[admin]

Hi Henrik,

3) There are couple of things you have discussed here. Lemme take one at a time

a) When you add a host to Chargeback hierarchy , all the VMs under that host automatically gets added. You can see the VM under the host in chargeback hierarchy. Now, if you generate a report it should have details of the VMs added. However note that the report is generated considering the hierarchy creation time . If the duration chosen for report is before that of hierarchy creation or VM addition, the report would not have VM details. One easy way to check whats happening is to select "Show Info Messages " in the "Cost info" tab while generating report.

b) Does VM belonging to host make sense ?

If the vCenter Server is added with "Host and Clusters" view, the VM can be below any of ResourcePool, Host or Cluster. When any of these entities are added, the children VM automatically gets added. Note that, if you drag and drop a host under a cluster into Chargeback hierarchy, the VMs which belong to cluster ( in this case they belong to cluster not host) , do not come in Chargeback hierarchy automatically.

c) How Billing policies work

You do not associate a Billing policy with any entity in chargeback hierarchy. Billing policy is a part of cost model and you choose the right cost model (If you have created one beyond the default cost model) during report generation. .If you are referring to entity cost configuration (configure cost tab) the the configuration at VM will override the one at folder.At no time will a single VM be billed twice.

5) There is no setting in the Chargeback through whichyou can purge old data. However, We do provide some database management scripts. These scripts let you purge/archive old data.

Hope this helps. Do post if you have more questions.

- naveen

[HenrikElm]

Ah yes, you spotted it. I have my hosts in a HA/DRS cluster..

I will try it all out more of course and learn by hands-on lab, but I have one more question..

All the metrics (CPU cost, bytes xfered and so on) are they only able to bill VMs? Let's say I would like to bill a host that way instead, like billing the aggregated network traffic from a host no matter what number of VMs on it or what they are used for? The scenario might be that I host a host (no pun ) for a client and he creates as many VMs as he likes, but I have responsibility for all technical/practical costs associated with the host itself?

I may ask noobish or more advanced questions in the future as I move along with the evaluation. I really want to get to know the whole vSphere family including CB.

/Henrik

[HenrikElm]

Also, the CB plug-in in vSphere client does not show any of the same info (like created hierarchies) or even all the same tabs as when I log in to the CB web GUI. Shouldn't the whole CB application be visible within the vSphere client?

/Henrik

[admin]

Hi Henrik,

yes, you can get the cost (and the actual usage) at a host level. But, the cost and the usage you see at a host level is a summation of the corresponding numbers for all VMs below it.

If you are hosting a host, all you need to do is to generate a report at the host and you have all information :smileyblush: .

1) What I said above is true for any entity, ie folder, cluster, resourcepool etc which cocntains VMs

2) Suppose you are runing a cluster. Now generating a report at a host does not give you any cost associated with host usage. This is because the VM does not belong to host. The usage cost is not associated with cluster

Feel free to ask questions

- naveen

[admin]

Hi Henrik,

wrt : CB plug-in

The user would login to vSpehere client using vSphere Server credentials. we do not want to reveal more information that what he is authorized for.

So we show the user the hierarchy equivalent to what he would see in vSphere Client (considering his privileges on these entities). This requires us to limit few features.

Hope this answers you question.,

-naveen

[Deval]

Hi

Good to see some communication there on Chargeback. I have set up my chargeback server and generated few sample cost reports. But i need validation of my understanding on following:

1. vCenter Server is managing VM's and hosts resource utilization data. I mean data is available with vCenter Server all the time

2. chargeback server picks up host and VM's utilization data from vCenter server to generate reports and Chargeback server also maintains all its data in its own DB instance.

3. If i have installed and configured my CB server today I can generate a report for Past month for any VM instance or hosts.

Please correct me if i am wrong i want to generate reports for past months.

Thanks and regards

Deval Kulshrestha

[admin]

Hi Deval,

1. vCenter Server is managing VM's and hosts resource utilization data.

I mean data is available with vCenter Server all the time

Yes Data is available in vCenter Server all the time unless the VM or host is deleted. Once deleted, the data will be lost.

2. chargeback server picks up host and VM's utilization data from

vCenter server to generate reports and Chargeback server also maintains

all its data in its own DB instance.

Chargeback server operates in two modes.

a. Synchonization enabled while adding VC. - In this case the chargeback server will maintain all utilization data in its own DB and do not go to vCenter Server while generating reports. It would have replicated the data in back ground b. Synchonization enabled while adding VC (on-demand) - In this mode, chargeback will not replicate the utilization data. It will go to vCenter Server for utilization data every time the report is generated.

In both modes, chargeback will not write anything to VC database. It maintains all its data in its own DB.

3. If i have installed and configured my CB server today I can generate a report for Past month for any VM instance or hosts.

Please correct me if i am wrong i want to generate reports for past months.

Unfortunately no.The report will only consider the duration in which the VM / Host belonged to chargeback hierarchy. If you install CB server today, the VM can belong to CB hierarchy only from today and hence report will have costing applicable from today.

Thanks and regards

Naveen

[Deval]

Dear Naveen

Thanks to make me understand the architectural part of it.

While testing Chargeback i also found that whenever i generate any report based on variable cost model , as per defined in Chargeback Evaluator guide. It seems to be not calculating memory and CPU usage for given entity and its showing 0.0 as an utilization value for CPU and memory. I am failed to understand how come a running machine is not occupying any memory , i mean it can not run on 0 memory.

Please suggest if there are any further tuning required.

Let me know if you need any other info in order to debug this issues.

Thanks and regards

Deval Kulshrestha

[abhinavkatiyar]

Deval,

Do you still see this issue?

Please note that you can generate reports starting from the time when an entity (i.e. VM) was added to chargeback hierarchy.

-Abhinav

[Deval]

Hi abhinav

Thanks !

Infact this question was answered previously. My recent question was abt. the values of utilization and amount being charged in reports.

How come a machine is live and up but not consuming memory and cpu cycles.

It was shown as 0, 0 in reports

Thats what i wanted to know

Thanks and regards

Deval Kulshrestha

+919818682509