Manually upgrading the vpxa agent of an ESX server

During my ESX 2.5 Upgrade project I ran into some VM migration problems.
As soon as the vmdk was too big to be able to complete the migration in 25 minutes, the operation would fail.

A cause could have been that the version of the vpxa agent on the ESX 2.5 agent was of a different build than the vpxa agent on the ESX 3.5 destination host. So I needed to manually upgrade the vpxa agent of the ESX 2.5 host.

As you can figure, documentation on how to do this isn’t very widely spread, so I decided to write a short blog post on it:
1.Log into the VC server locally and browse to the “Upgrade” folder. Default: C:\Program Files\VMware\Infrastructure\VirtualCenter Server\Upgrade
2.Browse to the correct vpx upgrade file for your ESX server version you need to upgrade the vpxa agent on.
My ESX 2.5 was version 2.5.2 so I needed vpx-upgrade-esx-4-linux-*. I found this info in the bundleversion.xml file

3.Copy this file to the ESX host you need to upgrade the vpxa agent on via a winSCP or PenguiNet or something like it.
4.Log into the ESX as root and browse to the folder where you have copied to upgrade file
5.Run the command: service vmware-vpxa restart
This will stop and start the agent and automatically upgrade it’s version. This shouldn’t take more than 5 – 10 seconds.
6.Now log back into the VirtualCenter server locally and restart the services:
a. VMware License Server
b. VMware VirtualCenter Server

OK, that’s it. Your vpxa agent is now upgrade to the version and build you have selected. It could be that in VC itself you will have to disconnect the ESX host and then connect it again.

ESX command line commands

I am currently busy with designing a migration strategy for a large ESX 2.5.2 migration to ESX 3.5 for one of our customers. (yeah I know, a little late ... but at least they will migrate immediately on the vShpere4 :))

While doing my tests and type the commands mostly only once and from then on use the arrow keys to go up to previous commands. You know how it goes: as lazy as you can get it ... :)

Anyways, I decided I would post most of these commands here. Not just for you guys, but admittingly also for myself as I regularly find myself looking through my memory for correct syntaxes ... :)

List files:
vmware-cmd –l
List path and names of .registered VM vmx files on the present host

Get state vm
vmware-cmd /vmfs/volume///.vmx getstate
Retrieve power state of the VM: off, on, suspended, stuck

Reboot vm
vmware-cmd vmfs/volume///.vmx reset trysoft/hard
Reboot the VM. First try a nice shutdown (trysoft), then if necessary force a shutdown before reboot (hard).

Power on vm
vmware-cmd vmfs/volume///.vmx start
Power on the VM

Shutdown vm
vmware-cmd vmfs/volume///.vmx stop trysoft/hard
Shutdown/halt the VM. First try a nice shutdown (trysoft), then if necessary force a shutdown (hard).

Suspend vm
vmware-cmd vmfs/volume///.vmx suspend
Suspend the VM

Verify snapshot
vmware-cmd vmfs/volume///.vmx hassnapshot
Query if VM has a snapshot

Create snapshot
vmware-cmd createsnapshot name description quiesce memory
Quiesce will quiesce file system writes, while Memory will grab the memory state

Revert to snapshot
vmware-cmd vmfs/volume///.vmx revertsnapshot
Revert to previous created snapshot (you loose the current VM state!)

Remove snapshot
vmware-cmd vmfs/volume///.vmx removesnapshots
Remove previous created snapshots (you keep the current VM state!)

Register vm
vmware-cmd -s register vmfs/volume///.vmx
Register VM (add to inventory)

Unregister vm
vmware-cmd -s unregister vmfs/volume///.vmx
Unregister VM (remove to inventory)

Answer vm
vmware-cmd vmfs/volume///.vmx answer
answer pending request for userinput

Extend virtual Disk
vmkfstools -X 12G ./testing.vmdk
To extend an existing Virtual Disk to 12GB.
Be aware if the shrinked size is smaller as the partition size in the guest there might be a data losse or a corrupted system resulting!

Copy virtual disk
vmkfstools -i /vmfs/volumes///.vmdk /vmfs/volumes//.vmdk
Copy vmdk from one vmfs to another datastore

Export virtual disk
vmkfstools -i /vmfs//.vmdk -d 2gbsparse //.vmdk
Export vmdk to ext3 partition

Rename files
vmkfstools -E
Rename files associated with a specified virtual disk

Delete Virtualdisk
vmkfstools -U
Delete files associated with the specified virtual disk

Delete folders
rm –R –f /vmfs/volumes//VM folder>
Delete non-empty folders

Find functionality
| grep –i “”
Example: vi *440*.vmx | grep –I “version”
Find a word with a file. In this example “version”

There are of course a whole lot more vmkfstools and vmware-cmd commands, but I think this gives a good start ...

Enjoy!

SMS Site System Status Summarizer still cannot access storage object. The operating system reported error 2147942405: Access is denied.

Now that I've got my OCS 2007 R2 successfully integrated with Cisco Call Manager 7, I thought I'll expand our test environment and start implementing SCCM 2007 R2 for monitoring the servers installed and facilitate the deployment of future clients.
I chose to set up my SCCM environment on 4 servers:
SCCM1 = MP, FSP, SLP, Site server
SCCM2 = DP, PXE, SUP
IIS1 = Reporting Point, SQL Reporting
SQL1 = DB server

I installed the first SCCM1 server and from within the management console I deployed the other SCCM services (SCCM2 and IIS1). For this I followed the Technet deployment and configuration guides for SCCM 2007 R2. No problems there.
Also with the installation and configuration of Reporting on IIS1 all went ok. Running the reports works fine. So all ok you would think?!

Still, I kept receiving these annoying messages under Site Status - Site System Status - IIS1\C$\SMS
(in the print screen below, the status is now OK, because of course by now I resolved the problem ...)
(The Site System Status is CRITICAL because I have installed the PXE service, but haven't configured it yet, I'll do that next ...)

So anyway, here I would get these messages:

SMS Site System Status Summarizer still cannot access storage object "\\BTLABIIS1\C$\SMS" on site system "\\BTLABIIS1". The operating system reported error 2147942405: Access is denied.

Everywhere you look online it will tell you to make sure the SCCM site server computer account (SCCM1 in my case) is member of the local Administrators group on the remote SCCM server.

Unfortunately for me, that didn't do the trick. Reading the documentation on Technet, I knew I had this before anything else, but still with the result displayed above.

To resolve this, there is a second account that needs to be member of the Administrators group on the remote server: the service account used to deploy the SCCM service with. You can look it up Site management - Site Settings - Site Systems - - Role properties.

When you add this account to the Administrators group, this error message will dispear.

Hope it helps.

OCS 2007 R2 Documentation

You also noticed that when you search for OCS 2007 R2 documentation, you get a lot of OCS 2007 documents? Drove me crazy, whenever I followed a link on a OCS 2007 R2 site, I got to OCS 2007 knowledge base.

So I decided to make a collection of true OCS 2007 R2 document libraries, to facilitate my OCS 2007 R2 - Cisco Call Manager integration project:
1. Microsoft Office Communications Server 2007 R2
The mother load, from the Microsoft download center. I just don't understand why this doesn't show up when you Google it.
2. Microsoft Office Communications Server 2007 R2 Documentation
A CHM file with technical documentation to help you understand, plan, deploy, and operate Microsoft Office Communications Server 2007 R2 servers.
Remember that you might have to "Unblock" the content.
If you need help on that, here is the Microsoft support page: You cannot open HTML Help (.chm) files from Internet Explorer
3. Microsoft Office Communications Server 2007 R2 online documentation
If you don't want to mess about with the .chm file, you can find the same documentation online, on the Technet pages.

Hope this can help you on your OCS search.

Exchange 2010 beta available: a list of the new features

So yesterday (14/04/09) we were able to download the first beta of the new email system that is used by 65 % of the companies. In Q3 of 2009 the distribution of the RTM version should be a fact.

First thing I noticed is that the admin interface has not been majorly changed, compared to the new interface we got from E2K7. That's good, cause I like this new interface much better than the old one.

Based on Microsoft documentation, I've set off testing and playing around in my test lab. The improvements they have made (27 to be exact) are divided by Microsoft in 3 big pilars:
1. Flexibility and reliability
2. Anywhere access
3. Protection and compliance

Of course, one of the most eye catching improvement is the support for OWA on Internet Explorer 7 & 8, Safari 3 and Firefox 3. But another cool feature that I think will be much used is the "MailTips". It will protect end users from sending personal mails to large mailing groups. The last thing that caught my eye yesterday already is the "Consolidated view" and "Conversation Mute". Finally we are able to view all e-mails concerning a single topic in a single node (in Outlook ànd OWA).

Below is a list of the other 27 improvements. The ones I think are really interesting and that I'll be looking into I've indicated in bold:
1. Added internet browser support for OWA
2. Answer/forward status is being kept by the server and can be shown on all clients
3. MailTips to prevent personal mails from being distributed to large communities
4. Conversation view for the threating of messages
5. Calendar sharing available in OWA and for federated users
6. Sharing of contacts outside the company and outside the desktop
7. Voice Mail Preview: automatic written preview of a received voice mail
8. Call Answering Rules to administer phone calls just like e-mails
9. Rights Management in OWA as well as Outlook
10. Federation Services to connect your Exchange servers to those of other organizations
11. Page patching: automatic fixing of corrupted DB pages
12. I/O Optimalization: less I/O bursts, support for SATA disks
13. JBOD (Just a Bunch Of Disks) support instead of only RAID
14. Database Availability groups: redundant DB's for mailboxes which provide automatic recovery
15. Failover on DB level, clustering not longer necessary to provide high-availability, so more uptime
16. Online Move-Mailbox: moving MBX'en while the user is logged in
17. Rules for transport protection: an administrator can change the IRM protection AFTER they have been sent
18. Moderation: a transport rule that allows you to send a message to a reviewer before the message is actually sent
19. Rules for protection Outlook which allows you to automatically assign RMS templates to e-mails
20. Role-based access control for Outlook
21. Exchange Control Panel to assign end users specific levels of control
22. Message tracking for end users (without having to contact the helpdesk)
23. Distribution Groups can be created, modified and deleted by end users
24. Block/Allow list for mobile devices
25. Protected voicemail: you are able to block the sending of voicemails outside the organization
26. Personal archive: moving of PST files to a secundary mailbox for better performance and compliance
27. Multi mailbox search: finally you are able to search multiple mailboxes with a single click

As you can see, there are quite a few cool improvements.
Can't wait to get started ... :)

OCS 2007 - Failed to send SIP request: outgoing TLS negotiation failed; HRESULT=-2146893022

So now that I have my OCS server installed, a few troubleshooting tasks have to be done (in my case anyway).

First error I saw when running the Validate Front End Server Configuration wizard, is this:

Looking around online I see everywhere that the certificate is probably wrong. I configured the certificate with the FQDN of the OCS server, which looks to be correct, especially when I see the entries above the check user logon section, where it shows it succeeds in connecting to the OCS pool.

So anyway, I tried recreating a new certificate, this time with the FQDN of the OCS pool instead of the OCS server. After signing the new certificate with my CA and assigning it to the OCS server, it still Completes with failures. Only this is very interesting: now my login is successful but I can't connect to the OCS pool anymore. Complete opposite of what I had before!

What the f***??!! Before you start trying: assigning the certificate with the server FDQN to the OCS server and the OCS FQDN to the IIS (or visa versa) does not work either.

Instead, create a new certificate with these settings:
1. Subject name: FQDN of the OCS pool
2. Alternate name: not that important
3. Remember to check the Automatically add local machine name to Subject Alt Name, that way you create a multi-homed certificate

After assigning this certificate to your OCS server and in the IIS manager, you should be good:

Hope it can help...