VMware Cloud Community
darrelleddy
Enthusiast
Enthusiast
Jump to solution

Cannot re add host

I had two host in a cluster in VC with an iSCSI SAN. I got an error that one of the host was unavailable. I then removed the host and tried to reconnect with no success. I then wen to the console and tried userdel -r vpxuser. It comes back with the following error: userdel: unable to lock group file. How do I fix this?

Darrell Eddy

0 Kudos
1 Solution

Accepted Solutions
admin
Immortal
Immortal
Jump to solution

Run df -h and see which partition is full. Free some space by moving some files to another location and restart hostd. This should resolve the issue.

-Sandeep

View solution in original post

0 Kudos
26 Replies
AndreTheGiant
Immortal
Immortal
Jump to solution

You must remove the VC agent, not only the user.

Witch version of ESX do you have?

Andre

**if you found this or any other answer useful please consider allocating points for helpful or correct answers

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
darrelleddy
Enthusiast
Enthusiast
Jump to solution

3.5.4

Darrell

0 Kudos
AndreTheGiant
Immortal
Immortal
Jump to solution

You can follow this instruction (could be fine also for 3.5):

Andre

**if you found this or any other answer useful please consider allocating points for helpful or correct answers

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
darrelleddy
Enthusiast
Enthusiast
Jump to solution

I ran the command then rebooted the host. I watched the ocnsole and I saw something flash by about a disk full error. How can I review the console log?

Darrell

0 Kudos
AndreTheGiant
Immortal
Immortal
Jump to solution

You can use:

tail -f /var/log/messages

or

dmesg (if are kernel errors)

or

tail -f /var/log/vmware/*log (for vmware related logs)

Andre

**if you found this or any other answer useful please consider allocating points for helpful or correct answers

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
darrelleddy
Enthusiast
Enthusiast
Jump to solution

In the var/log directory, there are several versions of ksyms., spooler., vmkernel.*. I cannot connect to the host via winscp because the network services will not log. A message flashed on the console about being out of space. I could not read it because it was too fast. What files can I safely delete in the /var/log directory to regain space? WHat other directory should I check for space issues?

Darrell

0 Kudos
admin
Immortal
Immortal
Jump to solution

Run df -h and see which partition is full. Free some space by moving some files to another location and restart hostd. This should resolve the issue.

-Sandeep

0 Kudos
darrelleddy
Enthusiast
Enthusiast
Jump to solution

OK,

I deleted some ISO's to free space. I rebooted the host and I still cannot connect va the VC client. Ping from a windows box does not respond. When I run esxcfg-vswif -l there some how DHCP is enabled. A quick on the other host shows the defined settings. Is there a configuration file where this is stored on the host that I can use to change it back to what it was or do I just manually put it back?

Darrell

0 Kudos
darrelleddy
Enthusiast
Enthusiast
Jump to solution

tail var/log/vmkwarning reveals the following "Initialization for vmfs2 failed with -1"

I did not find anything in the forums on this but does it mean?

Darrell

0 Kudos
AndreTheGiant
Immortal
Immortal
Jump to solution

VMFS2 is the old filesystem of ESX2.

I do not think that this could be the problem.

Andre

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
darrelleddy
Enthusiast
Enthusiast
Jump to solution

esxcfg-vswif -l reveals that it is set to DHCP. This is not how I had it setup originally.

0 Kudos
AndreTheGiant
Immortal
Immortal
Jump to solution

This could be a problem.

Fix the IP, fix the FQDN DNS resolution and add again.

Andre

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
darrelleddy
Enthusiast
Enthusiast
Jump to solution

Ok,

Fixed IP, checked DNS and all seems good. Not when I try to reconnect, the alarms show "Login failed due to a bad username and password". I have an identical second host that connects just fine. There are two VM's that are assigned to this disconnected host. Is htere way to manually move them tothe second HOST if I can't connect to the disconnected host via the VI client. I can only access the console.

Darrell

0 Kudos
AndreTheGiant
Immortal
Immortal
Jump to solution

To manually remove the old agent follow this infos:

(see the rpm part where you can remove the package)

Andre

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
darrelleddy
Enthusiast
Enthusiast
Jump to solution

When I do this: "rpm -qa | grep vpxa" it returns nothing like it is not installed. I restarted the host and tried to add the host to VC and get the same error.

Darrell

0 Kudos
AndreTheGiant
Immortal
Immortal
Jump to solution

Have you tried to use VIC from VC to connect to the esx, using the same name that you use when you try to add into VC?

Just to be sure that VC can reach (at least with VIC) the host.

Andre

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
darrelleddy
Enthusiast
Enthusiast
Jump to solution

Sure did. I tried it both from withing VC and VIC. I can only access the console from my KVM. No remote SSH or anyting. I can ping out from the box, but I cannot ping into the box.

0 Kudos
AndreTheGiant
Immortal
Immortal
Jump to solution

First you have to solve the network issue.

Ping and VIC must work.

Check netmask settings using esxcfg-vswif -l

Check also the cable.

Andre

Andrew | http://about.me/amauro | http://vinfrastructure.it/ | @Andrea_Mauro
0 Kudos
darrelleddy
Enthusiast
Enthusiast
Jump to solution

Forgot to mention. When I do an add host, it adds the host to VC, but is gets addedd as "disconnected". That is when I see the error "login failed due to bad username or password" listed under recent task.

Darrell

0 Kudos