gi-minni
Contributor
Contributor

patch update failed on esxi 3.5 Update 1

After looking inside your patch website I discovered a newer patchfile named ESXe350-200804401-O-BG-PATCH.

The difference is only at firmware level going from build#:82664 to build#:85332 and I tried to update the running system.

But after reaching 11% of completeness I got the following error:

Failed to Install Firmware update: Unable to copy install files to host.

What have I done so far:

  • switched esxi server in maintenance mode

  • disabled my own firewall otherwise I could not download nothing

  • started the VIC update manager

  • checked designated host for update

  • pressed check now

  • clicked on Install Updates

  • entered user and password

After looking around I found the following error inside the message logfile on the server:

NFc_AuthdServerLoop saw error: File error -- Failed to write to the target file: there is no space left on the device

IMHO obviously the size for storing this files is not large enough and the image must be readjusted!

Could someone explain how to enlarge the space and which of the partition (Hypervisor1,2,3 ?) is affected?

I booted from an 4GB USB-stick!

0 Kudos
16 Replies
Dave_Mishchenko
Immortal
Immortal

I don't beleive you'll have any option to enlarge the bootbank / altbookbank partitions or that you particularly will need to . It may be the /tmp folder that is getting ful although I haven't looked into the details of which folders ESXi is using when you copy up the patch files. Which build are you trying to upgrade from? Have you applied any patches already? I would suggest trying to install the 85332 patch from the command line. Basically extract the zip files into a folder and you should see the program remoteinstall.exe. See the bottom of this page for instructions on running that - http://www.vm-help.com/esx/esx3i/Patch_3i_without_VC.html.

0 Kudos
ocremel
Hot Shot
Hot Shot

Are you running without scratch ? (Configuration, Advanced Settings, ScratchConfig, is the location /tmp/scratch ?)

(It should not be needed, just trying to narrow down your issue)

0 Kudos
jparnell
Hot Shot
Hot Shot

I'm also unable to update our embedded ESXi host. The task just fails with "Remediation failed for vms-3.swwater.co.uk: SingleHostRemediate: Platform Configuration error"

The host is completely new, only config thats been done is network settings.

Scratch is set to /tmp/scratch

0 Kudos
gi-minni
Contributor
Contributor

Hi Dave,

> Which build are you trying to upgrade from

From a vanilla ESXi 3.5 Update 1 (Build 82664) -> ESXe350-200804401-O-BG (Firmware patchlevel:85332)

> Have you applied any patches already?

No, I haven't

> I would suggest trying to install the 85332 patch from the command line.

> Basically extract the zip files into a folder and you should see the program remoteinstall.exe.

> See the bottom of this page for instructions on running that - http://www.vm-help.com/esx/esx3i/Patch_3i_without_VC.html.

Ok, thx for this suggestion I will try and report back my experience.

0 Kudos
gi-minni
Contributor
Contributor

Are you running without scratch ? (Configuration, Advanced Settings, ScratchConfig,...

No, I never changed this entry. BTW it points to /vmfs/volumes/48270ea4-de88c1da-9974-001641e2fe01/.locker

looking inside the ScratchConfig.ConfiguredScratchLocation and this is my local datastore with 90GB empty space on it.

For your info ScratchConfig.CurrentScratchLocation points to the same directory and ScratchConfig.ConfiguredSwapState

and ScratchConfig.CurrentSwapState are not checked.

What role do this scratch directory play? Is it only for swap space or could it be used also for updates? Strange is

that looking at the kernel messages I read always /tmp, but scratch points to another directory?

... is the location /tmp/scratch) ?

Yes, I have a /tmp/scratch and a symbolic link /scratch pointing to that location. But this is not configured inside

the ScratchConfig variable

Executing cat /etc/mtab I got:

none /tmp visorfs 2,128,tmp,rw,suid,,dev,exec,async,atime,diratime,loud 0 0

After doing a df -h I got:

Filesystem Size Used Avail. Use% Mounted on

unknown 155.1M 87.9M 67.1M 57% /

Could be that the downloaded and extracted zipfile is bigger than 67.1M ?

0 Kudos
jparnell
Hot Shot
Hot Shot

Ive managed to find the esxupdate.log on the host, and here is what I'm getting:

Installing packages...

/altbootbank/cim.tgz failed on MD5 digest check;

Expected :fdd8724876f4c52e8cd454a3396e84c4; got 46709d9f9774427dc274fb3d0fcfae77

Install script returned with status 1

Cleaning up temporary install files...

Installation of package ESXe350-200804401-I-BG failed.

Failed to parse version from build value in /altbootbank/boot.cfg

Expecting :

Got :

Failed to parse rel from build value in /altbootbank/boot.cfg

Expecting :

Got :

PackageDB :unable to convert bootstate in altbootbank to integer

bootstate is :

Ive tried downloading the update again, and it fails again with the same error messages.

0 Kudos
ocremel
Hot Shot
Hot Shot

Is your machine an HP machine and did you get ESXi embedded directly from HP ?

0 Kudos
jparnell
Hot Shot
Hot Shot

Yes, the server is a DL585G2 and its HP's embedded (USB) ESXi

0 Kudos
ocremel
Hot Shot
Hot Shot

There was a glitch and the second bootbank (where the update goes before the bootbanks are swapped) is corrupted on those machines. A special patch is being prepared to correct the problem.

It is simple enough to fix from the support shell but I cannot officially tell you how. Should be easy to find out but the special patch should be available soon anyway.

0 Kudos
jparnell
Hot Shot
Hot Shot

Thanks, good to know what the problem is. I'll hang on for the patch.

0 Kudos
cenintagevm
Contributor
Contributor

Any luck on a patch being issued yet? Who should we contact, HP or VMware?

Thanks,

Josh

0 Kudos
ocremel
Hot Shot
Hot Shot

The P04 patch will take care of this issue (amongst other things). As luck would have it, it should be out tomorrow (Jun 03).

0 Kudos
cenintagevm
Contributor
Contributor

Thanks very much for the information. I'll look for it tomorrow.

0 Kudos
lexonb
Contributor
Contributor

I have same problem with update HP ESXi embeded usb key on HP Proliant BL460c (CPU E5440). The P04 patch don't solve this trouble. The Shift-R combination don't normally work in HP Hypervisor and destroy our usb key after upgrade. HP Support engineer send us recovery image for HP USB VMWARE Hypervisor. I think that HP Vmware USB Hypervisor have incompatible version with generally Vmware ESXi embeded. We are don't have support for HP ESXi USB now and don't know officially information about this case. I know that HP ESXi USB include HP Agent for HP SIM, but i wanna know full different between HP ESXi and VMWARE ESXi. HP ESXi USB don"t have manual how VMWARE (ESX Server 3i Embedded Setup Guide) Smiley Sad

So, waiting for a solution this problem Smiley Sad

0 Kudos
ocremel
Hot Shot
Hot Shot

What exactly happened to you ? You applied P04 and got into trouble ? What kind of trouble ? You then rebooted and tried shift -r and got a non-bootable system ? What were the symptoms ?

The image shipped by HP contains HP CIM providers that are not in the generic VMware image. That is the only difference and it should be transparent.

0 Kudos
AdamSnow
Enthusiast
Enthusiast

I am having this problem to on a BL460c. I tried to run update manager against a vanilla (installed by HP) 3i image. It corrupted the install. so I downloaded the recovery image, reinstalled using it, but now Update Manager won't even run. I get the "platform configuration error" too.

0 Kudos