Upgrade from ESX 4.0 to ESX 4.1 fails. Server ESX won't start after reboot.

[bachelor37e]

I tried to upgrade to upgrade a ESX 4.0 server to ESX 4.1 using upgrade manager and upgrade-from-ESX4.0-to-4.1.0-0.0.260247-release.zip (tested md5-hash, was correct). In update manager the progressbas stopped at 25% and timed out after a while. On the console I could see that it was hanging while trying to start VMNIX. In failsafe mode it tried to boot ESX 4.0 but during the boot it changed to 4.1 and gave a bunch of version error messages.

I did a clean 4.1 install on the server, it worked. Then I tried to update the next ESX 4.0-server. Same problem. I got 10 more servers to upgrade and I don't want to get the same problems. Any suggestions how I can avoid it?

The servers are HP ProLiant BL460c G6 blades.

[virtualdive]

Hi,

In this case you might wanna try with Host Update Utility and see if the similar thing comes up?

thanks

[stanleykwan]

For me, I won't do this on update manager,

why don't use CLI command, it save and easy by using vihostupdate.pl script,

I upgrade 4 ESX4 to ESX4.1 and 2 ESXi4 to ESXi4.1 without any issue and problem.

[nkrishnan]

Hi,

Could you please provide the detailed steps and logs from Update manger as well as esxupdate.log

1) What is the update manger and vCenter server are you using it

2) In the Home->Solutions and Applications-> Update Manger go to host Upgrade Releases tab.

3) Import the Upgrade ZIp bundle

4) Create the Base line and attach with Host

5) Scan the host against base line, it should show the base line as Non-compliant.

6) start the remediation,

If you still face the problem, Upload the VUM and esxupdate.logs

Thanks

[bachelor37e]

Hi,

Could you please provide the detailed steps and logs from Update manger as well as esxupdate.log

1) What is the update manger and vCenter server are you using it

2) In the Home->Solutions and Applications-> Update Manger go to host Upgrade Releases tab.

3) Import the Upgrade ZIp bundle

4) Create the Base line and attach with Host

5) Scan the host against base line, it should show the base line as Non-compliant.

6) start the remediation,

If you still face the problem, Upload the VUM and esxupdate.logs

Thanks

Thanks for all the replies!

vCenter 4.1.0 and a updated update manager.

I did the same steps in Update Manager as you describe.

So I tried updating from the console with esxupdate instead:

First I applied the pre-upgrade, it went fine:

/usr/sbin/esxupdate --bundle pre-upgrade-from-ESX4.0-to-4.1.0-0.0-260247-release.zip update #gave no errors

Then I tried to appy the upgrade:

/usr/sbin/esxupdate --bundle upgrade-from-ESX4.0-to-4.1.0-0.0-260247-release.zip update

Went fine, until it gave this error:

Running ...

Encountered error PostScriptError:

The error data is:

Script - esxcfg-boot -b

Exitcode - 127

Errno - 19

Description - A post-transaction script failed with a nonzero exit code.

Examine esxupdate logs for more details.

esxupdate log

[bachelor37e]

I tried updating with vihostupdate.pl

C:\Program Files (x86)\VMware\VMware vSphere CLI\bin>vihostupdate.pl --server XXXXXX -i -b "c:\temp\ESX 4.1\pre-upgrade-from-ESX4.0-to-4.1.0-0.0-260247-r

elease.zip"

Enter username: root

Enter password:

Please wait patch installation is in progress ...

Host updated successfully.

C:\Program Files (x86)\VMware\VMware vSphere CLI\bin>vihostupdate.pl --server XXXXXX -i -b "c:\temp\ESX 4.1\upgrade-from-ESX4.0-to-4.1.0-0.0.260247release.zip"

Enter username: root

Enter password:

Please wait patch installation is in progress ...

Host was not updated, no changes required.Skipping bulletin ESX410-GA-esxupdate; it is installed or obsoleted.

The update completed successfully, but the system needs to be rebooted for the changes to be effective.

I rebooted and it was hanging at the same place as before.

[sc_2111]

Hi ,

I just did the same with the same HW ( bl460c ) .

I used update manager to roll out 4.0 updates before updating to 4.1 .

Don't know if that was required but worked for me

HTH

stefano

[bachelor37e]

In esxupdate.log it sad that esxcfg-boot could not be fond. It is located in /usr/sbin/. /usr/sbin/ and /sbin/ is not in the $PATH-variable on my esx servers, so I added it before i tried to update one more of my servers to 4.1 with esxupdate. No error on esxcfg-boot in the log file, but the server still hangs on reboot.

[bachelor37e]

Esxupdate can't find vsish. I checked and it is nowhere to be found at the server. Shall a ESX 4.0 server have vsish? is there some way I can install it?

INFO: cos: vsish command /usr/sbin/vsish cannot be found

INFO: cos: No hardware vendor/model information obtained

[johu]

Got exactly same problem when upgrading from ESX 4.0.0 261974 to ESX 4.1.0 260247 with upgrade manager. I have 10 servers - three FSC RX300S3, five FSC BX620S3 and two HP BL460c G1. All RX300S3's and four BX620S3's are old installs originally done back in 2007. One FSC BX620S3 and both HP BL460c G1's are newer installs from 2009. I have successfully upgraded those three newer ESX installs but seems like rest get stuck in starting vmnix. So far killed two of them (one RX300S3 and one BX620S3). Server is not actually completely dead as keyboard still echos on screen. I know these old FSC servers are on HCL only up to ESX 4.0 U2 so actually this isn't even supposed to work. Guess I'll need to do clean install next rather than upgrade to see if it works.

[bachelor37e]

I would be interested in seeing the esxupdate.log for one of the server you got the same error on. So if you are going to try to update one more, please copy out the /var/log/vmware/esxupdate.log before you reboot if you can.

[admin]

Can you post a screenshot of where it fails loading VMnix?

[johu]

Attached image how screen looks. Second image is without "quiet" kernel parameter. CtrlAltDelete can be used to restart server when it's stuck on this "vmnix" stage. Troubleshooting mode results broken mix of 4.0 and 4.1. I'm running out of servers to break so no esxupdate.log to post.

[admin]

From the screenshot it looks like esxcfg-boot was the problem. The initial ramdisk wouldn't have been created correctly, and that would cause it to hang when loading VMnix. I'm not sure why esxcfg-boot wasn't in your PATH though -- that seems suspicious.

--Patrick.

[nkaelin]

Hi

We had the same problem.

We are using Fujitsu Primergy RX300S5's.

After a BIOS update from 1.06 to 1.12 it worked without any issues.

Best regards

Norbert

[scottmlew]

I encountered the same problem, with a report of "esxcfg-boot -b" having failed, and I manually ran the command before rebooting and everything seems to be OK after rebooting (I say with fingers crossed).