By:
On:
In: Cisco, NSX, VMware
With: 1 Comment

We have a typical design with Nexus 7K Border-Leaf and a pair of NSX Edge Gateways enabled for Equal-Cost Multi-Path, and finally the Logical Router which is also enabled for ECMP. The issue I needed to resolve recently at a customer was to remove 0.0.0.0/0 from the OSPF routing table of the Border Leaf Cisco Nexus 7706’s.  This was being redistributed improperly by the NSX Edge Services Gateway (ESG). You see even if we do not enable default originate we still get the 0.0.0.0/0 because we told the ESG’s to redistribute static and connected.  In the NSX Edge when we select redistribute static the defaultRead More →

By:
On:
In: General, Networking, VMware, vSphere
With: 0 Comments

Came across this in my current engagement.  Was happening on one of the servers in the Payload Cluster, which was a fresh install of 6.0.0, 2494585 After upgrading to or installing ESXi 6.0.x and ESXi 6.0 Update 1, you may experience these symptoms: ESXi 6.0 host network connectivity is lost randomly ESXi enters into a non-responding state and becomes unmanageable until reboot Network connectivity does not return until the ESXi 6.0 host is rebooted After rebooting, the issue is temporarily resolved for a period of time, but occurs again after a random interval unless the workaround (script) is implemented In the /var/log/vmkernel.log file, there isRead More →

By:
On:
In: General, HP, LACP, LACP, Networking, VMware, vSphere
With: 0 Comments

So you have decided to go with Enhanced LACP.  Great choice, if you don’t plan to use auto-deploy and host profiles to do it.  This my friend is a manual process, well at least for those that don’t do Python scripting.  The focus of this article is to document how I got vSphere 6 and HP to behave with vSphere 6 Enhanced LACP. The first step is to prepare the environment for LACP.  If this configuration is taking place where host management, vCenter, and workloads that you want to actually not get disconnected it is recommended to configure this on a separate set of uplinksRead More →

By:
On:
In: NSX, VMware
With: 0 Comments

During a recent NSX deployment when the hosts, NSX Manager, and vCenter resided on different networks with firewalls between them I provided the following KB to the security team to add the required ports to the firewalls. Well it turns out this KB is missing one very important port requirement.  The vCenter server needs access to NSX Manager via port 80/tcp for host preparation for NSX. vCenter – > NSX Manager -> tcp/80 for Host Prep There are several other reasons host prep may fail, most notability the availability of Update Manager to approve the installation. Installing VXLAN Agent fails with ESX Agent Manager displayingRead More →

By:
On:
In: Linux, rsyslog
With: 0 Comments

I have a customer who doesn’t have a syslog server configured in his vSphere environment.  I first recommended we install vRealize Log Insight, it’s a great product and has ton of insightful information when used with vCenter, vSphere and NSX.  But the customer didn’t want to worry about licensing, so we decided to stand up a CentOS rsyslog server to show the power of having a centralized syslog.  This is especially true for NSX installations, I mean using syslog in the DFW alone will save a ton of troubleshooting time. This article assumes you know how to use vi or something similar in the LinuxRead More →

By:
On:
In: Cisco, LACP, LACP, Networking, VMware, vSphere
With: 0 Comments

So you have decided to go with Enhanced LACP.  Great choice, if you don’t plan to use auto-deploy and host profiles to do it.  This my friend is a manual process, well at least for those that don’t do Python scripting.  The focus of this article is to document how I got vSphere 6 and Cisco IOS to behave with vSphere 6 Enhanced LACP. The first step is to prepare the environment for LACP.  If this configuration is taking place where host management, vCenter, and workloads that you want to actually not get disconnected it is recommended to configure this on a separate set ofRead More →

By:
On:
In: Logical Switch, NSX, VMware
With: 0 Comments

When using IP discovery on logical switches in VMware NSX for vSphere 6.2.0, you experience these symptoms: The ESXi 5.x and 6.x host fails with a purple diagnostic screen The purple diagnostic screen contains entries similar to: @BlueScreen: CPU xx /World xxxxx tried to re-acquire lock dvfg.sec.confg PTEs:0xnnnnnnnn;0xnnnnnnnn;0x0; 0xnnnnnnnn:[0xnnnnnnnn]PanicvPanicInt@vmkernel#nover+0x37e 0xnnnnnnnn:[0xnnnnnnnn]Panic_NoSave@vmkernel#nover+0x4d In order to protect from this issue you must disable IP discovery for all logical switches.  This prevents the issue from occurring. NOTE:  Disabling IP discovery disables broadcast suppression for ARP traffic.  Which may, or may not cause a performance hit depending on the environment.  Additionally if you used new features in NSX 6.2 to allowRead More →

By:
On:
In: VCA, VMware
With: 9 Comments

In my case I had a network outage in my lab which resulted in the VCA, NSX Manager, Platform Services Controller, nearly everything to crash or hang.  Upon restart of the VCA and PSC I received the some errors associated with /dev/mapper/log_vg-log partition.  Asking me to fix the error manually. The problem was if I follow the instructions to enable the shell so I can run fsck, I simply get an unknown command error. Here is what I had to do to get moving again. Hit Control-D or Reboot the vCSA using the vSphere Client. When the GRUB bootloader appears, press the spacebar to disable autoboot.Read More →

By:
On:
In: Cluster Design, NSX, VMware
With: 0 Comments

While waiting for another session to start at TechSummit some colleagues @markmeulemans Bill Mansfield and I had the discussion around when is a dedicated Edge Cluster actually necessary.  This is the outcome of that discussion. Like many other VMware products, NSX for vSphere can actually be deployed in a relativity small footprint.  This is especially true for Proof of Concept / training / nested lab environments.  The decision to do things like a dedicated Management, Edge, and Payload cluster are typically “for us” associated with customer imposed constraints.  These constraints are often physical data center layout, compute limitations, or legacy design. Dedicated Edge Cluster HavingRead More →

By:
On:
In: NSX, VMware
With: 1 Comment

There are considerations on where to place the “Active” DLR Control-VM.  This is due to an issue where “if” the active DLR Control-VM and any Active Edge are on the same host during a host failure the DLR Control-VM will not be able to provide the controller-cluster with routing updates. This will cause a black-hole effect because the routes associated with the failed Edge router have not been removed. By separating them, if the host with the active DLR Control-VM goes down there is no need to update routes as all Edge devices are still up and available.  The same goes if a host withRead More →