Blog

Posted on

vRealize Log Insight Admin Alert: SSL Certificate Error (Host =vrli.vrmware.nl)

This time a short post about a vRealize Log Insight (vrLI) configuration issue that took too long to solve. In the end, the solution is simple after I found the documentation. Finding the right documentation was the hardest part.

Just briefly the reason of this setup. I want ESXi hosts to use Syslog over SSL to send logging encrypted to vRLI.

While adding the vCenter I configured the hosts to use SSL.

After configuring, everything seemed to work fine, until I got a vRLI Admin mail with the following alert:

This alert is about your Log Insight installation on https://vrli.vrmware.nl/

SSL Certificate Error (Host = vrli.vrmware.nl) triggered at 2023-04-16T09:23:53.412Z

This notification was generated from Log Insight node (Host = vrli.vrmware.nl, Node Identifier = de568ad3-d4e3-7f8a-b543-cef17632af11).

Syslog client esx01.vrmware.nl disconnected due to a SSL handshake problem. This may be a problem with the SSL Certificate or with the Network Time Service. In order for Log Insight to accept syslog messages over SSL, a certificate that is validated by the client is required and the clocks of the systems must be in sync.

Log messages from esx01.vrmware.nl are not being accepted, reconfigure that system to not use SSL or see Online Help for instructions on how to install a new SSL certificate .

This message was generated by your Log Insight installation, visit the Documentation Center for more information.

Time couldn’t be the issue in my case. So it had to be a certificate issue. The issue was caused by the vRLI certificate that wasn’t in the ESXi host truststore.

Per ESXi host, the following steps should be taken to solve the issue. Step 3 is only a verification step.

  1. openssl s_client -connect [FDQN or IP vRLI]:1514 < /dev/null | openssl x509 -outform PEM >> /etc/vmware/ssl/castore.pem
    Example: openssl s_client -connect vrli.vmware.nl:1514 < /dev/null | openssl x509 -outform PEM >> /etc/vmware/ssl/castore.pem
  2. esxcli system syslog reload
  3. esxcli network ip connection list | grep 1514

If ESXi hosts have the vRLI certificate in their truststore, the vRLI Admin mail (1x per day per vRLI node) should no longer occur.

Here is the link to the VMware documentation. This documentation is actually for vRLI Cloud which is a different product than standard vRLI although they overlap in some areas. The documentation for vRLI will be updated according to VMware GSS.

So this is probably why the vRLI documentation on this topic was so hard to find. Hopefully this blog post will save you a lot of time.

Posted on

vRealize Log Insight persistent static route on second network interface on behalf of Log Forwarding

Recently I wanted to test whether it is possible to configure vRealize Log Insight (vRLI) log forwarding to a second network interface to reach a log target in another network segment that could not be reached from the default vRLI appliance ip address.

The first step is adding a second network interface to the appliance. In this example we use the following network configuration.

  1. VMnic1
    Vlan10, IP 10.1.1.10, Subnetmask 255.255.255.0, Gateway 10.1.1.1
  2. VMnic2
    Vlan20, IP 20.2.2.20, Subnetmask 255.255.255.0, Gateway 20.2.2.1
  3. In this example the log forwarding target ip address is 30.3.3.233

To configure the second network interface open a SSH session to vRLI appliance. Move to /opt/vmware/share/vami/ and run the network configuration script. vami_config_net. Eth1 is now also available for configuration. First select ‘0’ for a configuration overview. In the results is an error on eth1 displayed. This error keeps us from being able to configure eth1.

After some ‘Trial & Error’ research I noticed the following error during reconfiguring eth1 “can’t open /etc/systemd/network/10-eth1.network”

In the directory /etc/systemd/network is the file “10-eth1.network” not present. The name of the file could be different then here in this example. It depends on the number of network interfaces. I fixed this issue by creating this file manual.

  1. touch /etc/systemd/network/10-eth1.network
  2. chmod 644 /etc/systemd/network/10-eth1.network
  3. Config the second network interface. Go to the directory /opt/vmware/share/vami/ and run the network configution script. vami_config_net. Eth1 is now also available for configuation.
  4. Check the new configuration by selecting option 0. If Ok press 1 to exit
  5. Restart the network, systemctl restart systemd-networkd.service

Now this issue is fixed we can move on to configure the persistant static route for vRLI log forwarding.

Edit /etc/systemd/network/10-eth1.network

The file shlould look like this before editing:

[Match]
Name=eth1
[Network]

Gateway=10.1.1.1
Address=20.2.2.20/24
DHCP=no

[DHCP]
UseDNS=false

Now add route information at the end of the file: 

[Match]
Name=eth1
[Network]

Gateway=10.1.1.1
Address=20.2.2.20/24
DHCP=no

[DHCP]
UseDNS=false
[Route]
Gateway=20.2.2.1
Destination=30.3.3.233/24

Save the file and restart the network service.

systemctl restart systemd-networkd.service

Check if the new route is present.

route -n

Test if you reach the destionation from CLI. I used Syslog over UDP port 514.

nc -v 30.3.3.233 514

Answer if configuration is working:

[30.3.3.233 514] open

The last step is to configure the vRLI Log Forwarding Destination.

Send a test event and check if the event is received by the target.

Keep in mind that multic-nic configurations in vRLI are not officially supported.

Also credits to this blog post that pushed me in the right direction.

Posted on

vRealize Log Insight failed to start Journal Service

Recently, I started an ssh session on a vRealize Log Insight (vRLI) appliance. After entering name and password a message was prompted to change the password immediately. After I did this, the ssh session was dropped and I had to log in again. The password appeared to be unchanged. I repeated this step several times and then decided to restart the vRLI appliance. Hoping this would fix the problem.

After waiting for some time, the appliance does not appear to come online and appears to be in a loop during startup. The screenshot below shows that the Journal Service does not want to start.

I have seen similar issue more often and I suspect a partition has filled up. I couldn’t log in because the appliance won’t start. In the following action plan, I explain step by step how I solved this problem.

Action Plan:

1, Take a snapshot
2. Edit the VM options in the next step before restarting the vm
3. Change Boot Options, Change Boot Delay in 1000 milliseconds and Force Bios Setup

4. Restart the VM
5. Bios Exit Discarding Changes

6. Press ‘E’ direct during the start
7. Type “rw init=/bin/bash” at the end of the line and press “F10”

8. Type “df -h”. In our case we see that the root partition is full

9. Let’s clean up some old audit and authentication logs

10. Let’s check the root partition again. Type “df -h”.

11. Reboot the appliance and we’re done.

After cleaning up the root partition, the appliance started normally. I was also able to change the root password without any problems. Don’t forget change the Boot Options, Change Boot Delay back in 0 milliseconds and remove the snapshot if everything is working fine and you’re happy with this solution.

You can always open a VMware SR if you need help. 🙂

Posted on

Cannot install the vCenter Server agent(HA) service. Unknown installer error (UPDATE)

Almost two years ago I wrote a blogpost about a failure during installing the vCenter Server agent(HA) service. This post is one of most read articles on my blog. You can find the orginal post here. Recently, I ran into this problem again. This time I could not solve the problem using my earlier post.

I have found a workaround to this problem that is easy to implement and works well. I have been able to use it successfully several times recently.

  • Put the ESXi host in Maintenance Mode
  • SSH to the ESXi host
  • esxcli software vib remove -n vmware-fdm (no reboot needed)
  • Wait a few minutes and the result should like this:

Removal Result
   Message: Operation finished successfully.
   Reboot Required: false
   VIBs Installed:
   VIBs Removed: VMware_bootbank_vmware-fdm_7.0.3-19193900
   VIBs Skipped:

  • Check if the vCenter Server agent is removed
  • esxcli software vib list | grep fdm (result should be empty)
  • Exit Maintenance Mode, HA agent will be installed
  • Check if the vCenter Server agent is installed
  • esxcli software vib list | grep fdm (result should be like this:
    vmware-fdm 7.0.3-19193900 VMware VMwareCertified 2022-10-12
  • Done
Posted on

Slow startup and hung during reboot Ubuntu 22.04 LTS on Maxtang EHL30

Last week was VMware Explore Europe in Barcelona, Spain. Here vExperts were able to pick up a mini PC, the Maxtang EHL30. The mini PC was offered by the vExpert community and Cohesity. This as a gift for all the work vExperts do for the vCommunity.

The mini PC still had to be fitted with DDR4 memory and an M.2 SATA SSD. Since I wanted to keep it low budget I bought memory and an SSD on Amazon for about €50.

After inserting the memory and SSD, the Maxtang boots up. I have installed Ubuntu 22.04 LTS and used Rufus to create a bootable Ubuntu install USB media. After the setup was completed the system hung during the reboot. After a coldboot I observed that the startup was also take some minutes, much to slow in my opinion. After searching for a while I found the solution for the slow startup and hung during reboot/shutdown.

  1. enter command “sudo vi /etc/modprobe.d/blacklist.conf” in terminal
  2. add a new line “blacklist pinctrl_elkhartlake” , save and exit editor mode
  3. enter command “update-initramfs –u” in terminal
  4. reboot system, coldboot to apply change

Now the Maxtang EHL30 with Ubuntu 22.04 LTS Reboot and startup in a few seconds.

Posted on

vRealize Log Insight 8.10 bad performance

For a customer, I deployed a new vRealize Log Insight 8.10 cluster as a test. After initial installation and configuration of the first vRLI node I experienced a performance that was particularly bad. It wasn’t always this bad, but it frequently occurred.

The bad performance initially revealed itself when choosing one of the options in the menu on the left side of the screen. It seemed like the whole screen was frozen. After 10 till 30 seconds the requested information appeared.

I decided to first complete the cluster setup. After adding some vCenters and importing Content Packs, technically everything worked as designed. The bad performance was still there even after temporarily adding additional memory to the nodes.

During one of the moments that the performance was decreased I saw something in the lower left-hand corner. There were the following messages displayed:

  • Read <vRLI FQDN>
  • Transferring data from <vRLI FQDN>
  • Connecting to cdn.pendo.io

The last bullet was the key to resolve this issue. I searched some information about “cdn.pendo.io” and found this article about “Join or Leave the VMware Customer Experience Improvement Program“. At the bottom of the article is a section “What to do next”

After CEIP is enabled, when a user logs in to vRealize Log Insight, they see a banner at the top of their window that asks whether they want Pendo to collect data based on their interaction with the user interface.

  • If the user clicks Accept, Pendo collects their data and sends it to VMware
  • If the user clicks Decline, Pendo does not collect their data

In the General Configuration section of the vRLI configuration I deselected the Usage Reporting (Join the VMware Customer Experience Improvement Program). This resolved the issue for me.

Finally, in this case there was no need to join VMware CEIP. So, therefore it is an acceptable solution for me.

Posted on

VxRail V7.x log collection failed

Recently I ran into a problem when I needed to generate a VxRail V7.x support log bundle. The log collection failed after I went to the VxRail log section.

Cluster > Configure > VxRail > Troubleshoot > Create

After hitting the Create button I selected the required data. In my case the VxRail Manager data was enough.

When start to Generate the log bundle nothing happened. The Create button was greyed out for a few minutes and no log bundle was generated.

Solution:

  • Take a snapshot of the VxRail Manager vm
  • Open a ssh session to the VxRail manager
  • Login with mystic, switch to root
  • Cd /tmp/mystic/dc
  • Look for old log bundles for example VxRail_Support_Bundle_12d4235c-1458-d47f-e68b-a3221bca3bc3_2022-01-01_12_00_00.zip
  • Delete all old log bundles in /tmp/mystic/dc
  • Systemctl restart runjars
  • Systemctl restart vmware-marvin
  • Wait about 5-10 min.
  • Try again collecting VxRail manager logs

Now the VxRail Support log bundle was successfully generated.

If you are not familiar enough you can always open a Dell Support request. They will then help you further.

Posted on

Set ProductLocker Location with PowerCLI

Something I struggled with in the past is setting the productlocker, hassle with scripts and ssh and avoiding host reboots. Today we had to configure another 20 hosts with this and I was pointed to a solution doing it with the MOB. So my next thing was, how to automate this. And with 2 simple foreach loops you can set it and check it and be done!

Setting all hosts from a cluster with a specific path for the productlocker

$allhosts = get-cluster "clustername" | get-vmhost

foreach ($hostname in $allhosts){
Get-VMhost -Name $hostname | %{$_.ExtensionData.UpdateProductLockerLocation_Task("/vmfs/volumes/pathname")}
}

Query all hosts in a cluster to check what location is set

$allhosts = get-cluster "clustername" | get-vmhost

foreach ($hostname in $allhosts){
Get-VMhost -Name $hostname | %{$_.ExtensionData.QueryProductLockerLocation()}
Write-Host $hostname -ForegroundColor Green
}

Source: Hollebollevan

Posted on

Replacing the VxRail Manager SSL Certificate failed

Cause:
Replacing the VxRail Manager SSL Certificate through the Certificate Management Gui ended up with an error.

I was used to replacing the VxRail certificate using the cli. This was the preferred way somewhere up to version VxRail 4.7.x. All preparations for using the gui are the same as when replacing the certificate through the cli:

  1. Config OpenSSL.conf with the required information
  2. Create RSA key
  3. Creare CSR file
  4. Request a Certificate signed by your CA
  5. Download Certificate CA chain

The certificate and CA certificate chain must still be in PEM format.

So I start with a snapshot of the VxRail manager and copied the VxRail certificate, RSA key, certificate chain and entered a password in the gui. After pressing the Update button I received the following error.

Solution:
I will briefly tell you what went wrong. As told earlier, I used the same way to create the certifcates as I did when using the cli.

The error was caused by the certificate CA chain. After downloading the certificate CA chain file, the certificates are listed from top to bottom. The Root CA was at the bottom. This is order has always worked when replacing the VxRail manager certificate using the cli.

I was led to the solution after reading the certificate chain content information twice, see the screenshot above here. I assumed that the certificate format was causing the issue. There is also information about the certificate format next to the information about the order of the CA certificates in the chain file. I changed the order of the CA certificates in chain file with the Root CA on top.

At the next try I copied again the VxRail certificate, RSA key and certificate chain and entered a password in the gui. After pressing the Update button I received a notification that the VxRail Manager certificate is successful replaced. It takes about 5 minutes, before the VxRail manager is back online.

During this 5 minutes the VxRail SSL Thumbprint in the cluster in vCenter will be updated and services will be restarted. You can see the VxRail SSL Thumbrint on the cluster summary page in vCenter.

This auto replacement of the VxRail SSL Thumbprint is an enhancement compared to the old way when the certificate was replaced using the cli. The VxRail SSL Thumbprint then had to be manually read and copied from the VxRail manager to the cluster custom attribute in vCenter.

Return to the VxRail System info page after 5 minutes. You should now see the VxRail information.

Done.

Posted on

VMware Vrealize Log Insight upgrade from CLI

Did you know that you can also upgrade vReaze Log Insight from the CLI?

  • Upload the .PAK file to /tmp on the Log Insight appliance
  • Open 2 ssh sessions to the Log Insight appliance
  • Run the following Python script: /usr/lib/loginsight/application/sbin/loginsight-pak-upgrade.py /tmp/VMware-vRealize-Log-Insight-8.6.0-18703301.pak
  • Open the second ssh session to follow the progress in realtime in the upgrade log.
    tail -f /var/log/loginsight/upgrade.log

When the upgrade is finished the ssh sessions will be ended while the Log Insight aplliance is rebooting.

Done!