Anyone upgraded to VC 2.0.2?

[peetz]

Hi,

since yesterday VirtualCenter 2.0.2 is available (s. Release Notes[/url]).

Anyone upgraded already? Any problems? Will the performance data and the custom attributes be preserved?

\- Andreas

[wallbreaker]

no, the dns return name1.domain.com

[peetz]

Why have you put static entries in /etc/hosts if you have a functioning DNS setup?

Anyway, if you leave the entries in /etc/hosts then make sure that they use exactly the same spelling as the DNS server (both lowercase or both uppercase).

[wallbreaker]

It's OK now, everything in lower case, change the hostname of my ESX hosts with the command hostname

Message was edited by:

wallbreaker

[TCronin]

Did you use a fully qualified domain name when you entered your hosts into VC?

Best practice is to use hostname.domainname.com

[wallbreaker]

yes, but it was in lower case...

[Chris_S_UK]

The notes only say to create a new database if your VC tables are in the SQL "master" database, i.e. if you don't have a dedicated VC database. This is very bad practice anyway!

Chris

[esiebert7625]

I see one good thing, you can increase the task timeout from the previously hard-coded 15 minutes. Handy when committing large snaps.

Task Timeout Value Can Be Manually Set in the vpxd.cfg File

The default task timeout for VirtualCenter Server is set to 900 seconds. This value can be manually set in the vpxd.cfg file by adding:

The timeout value is specified in seconds.

Note: You must restart VirtualCenter Server after making changes to the vpxd.cfg file.

[ldiederich]

Hi,

today i've updated 5 esx servers with new VC agents and VC Server, too.

i've no problems atm

cu

Lars

[Fuzweezel]

Does anyone have an easy way to migrate everything from 'master' to a dedicated VC database ? running in sql 2000 sp4.

[esiebert7625]

See this link, you might have security permissions you'll need to fix afterward.

http://technet.microsoft.com/en-us/library/ms187858.aspx

[JohnGibson]

Upgrade on nearly all my hosts went OK, a mix of 3.0.1 2.5.3 and 2.5.4.

Made sure I had a /tmp/vmware-root folder on all my ESX 3 Hosts 1st.

Upgraded the license server 1st, and confirmed the old VC client worked against the new VC server before upgrading (it does).

Had to change my hostnames to lowercase in my /etc/hosts files to get HA working again - only realised this after all my HA servers wouldnt re-enable and got the multipleshorthostnames error.

Used this script command:

#!/bin/sh

mv /etc/hosts /etc/hosts.old

sed -e 'y/ABCDEFGHIJKLMNOPQRSTUVWXYZ/abcdefghijklmnopqrstuvwxyz/' /etc/hosts.old > /etc/hosts

and then ran it against all my ESX 3 hosts using

plink -ssh -m lowerhosts.sh root@esxhostname1 -pw esxholstpassword1

plink -ssh -m lowerhosts.sh root@esxhostname2 -pw esxholstpassword2

plink -ssh -m lowerhosts.sh root@esxhostname3 -pw esxholstpassword3

etc...

HA now working for me ok, a lot less painful than last time.

Custom fields retained ok.

More VMs are now showing heartbeat statuses also.

And it actually feels a little quicker, but that could be bacause im the only one using it instead of the usual 8+ users.

Still cant connect to my NAT'd ESX Hosts from VC though - terribly dissapointed on this bit.

John

[CDRippe]

I can confirm this issue as well.

I did not disable HA before doing the VC upgrade. When complete one of eight servers would not join the HA cluster with "multipleshorthostname" error.

The affected server had both its hostname and /etc/hosts file capitalized where it wasn't before. I am absolutely certain it was not before as I keep an off-line copy of the /etc/hosts to re-distribute when a new host is added to the cluster.

Strange

I renamed the host and then edited the /etc/hosts file to reflect the correct, lowercase, names and reconfigured host for HA.

All is well again.

C. D. Rippe

[ffalcone]

I am planning an upgrade but via a clean install on a newer box with SQL 2005 SP2. I have performed the ESX pre-work and am aware of the HA bug (thank you all for the good info) but have a question regarding the agent:

"If there are issues with the new VCenter and I have to roll back to the old server, can the new VCenter agent on the ESX boxes connect to the old VCenter box?"

Thanks

[wallbreaker]

3 of my VM have there status stucked.

\- two Warning for Virtual machine memory Usage

\- one Alert for Virtual machine CPU usage.

Didn't see this before.

I stop one VM to see what happen, the status is gone.

Then I restart the Virtualcenter server service, no more status for my 36 VM.

one migration between two of my esx hosts and there is a status for all the Vm one theses two hosts, but the status is frozen.

Message was edited by:

wallbreaker

[peetz]

I can confirm this problem. For several VMs the status display (the traffic light column in the VI client) does not work anymore after the upgrade.

If you restart the VC service it is normal that you won't see the status for some minutes, until the service finished collecting information from all VMs. However, I have not yet restarted the service since the upgrade.

[wallbreaker]

The host's status doesn't work more.

[MattyP]

Just upgraded to 2.0.2 from 2.0.1

Followed all instructions in the release notes, and made sure the /tmp/vmware-root directory existed.

I have an environment will all 3.0.1 hosts patched to the latest patches as of July, and DRS / HA enabled.

The upgrade went smoothy, and HA etc reconfigured itself without issue.

The cluster seems to be running as per normal.

All of my custom attributes and performance data were preserved, and my chargeback reports are still working too!

[northwest]

I disabled HA prior to the upgrade. I have 6 hosts.

I had an issue with one VM which had an alert (red CPU) that would not clear until I did a shutdown.

I also had one host that did not automatically update and connect after VC came up, but I just right clicked and chose connect and all is fine. I checked the version of the VC agent on that host and it was updated.

[VirtualNoitall]

Hello,

You basically want to:

Create new database

Stop VCMS services

copy all VPX* objects from master to new database

update the ODBC DNS to point to the new database

Start VCMS services

[sipsipi]

WOOHOO!!

Thanks guys, I had to change in /etc/hosts

from

ip.address.number.here HOSTNAME.domain.com

to

ip.address.number.here hostname.domain.com

The caps of the local host name in /etc/hosts was mad.