VSAN 6.1 boasts new features, promises greater…


VSAN 6.1 boasts new features, promises greater end-user security (via @anthony_poh)

VSAN 6.1 boasts new features, promises greater…

Looking to the past to plan for the future: how VSAN 6.0 and VSAN 6.1 updates improve upon VMware’s Virtual SAN model to increase shared storage capability.


VMware Advocacy

VSAN enabling Sky to be fast / responsive / agile…


VSAN enabling Sky to be fast / responsive / agile… @vmwarevsan

VSAN enabling Sky to be fast / responsive / agile…

Over the last couple of months I’ve been talking to a lot of VSAN customers. A while ago I had a very interesting use case with a deployment on an Oil Platform . This time it is a more traditional deployment: I had the pleasure of talking to James Cruickshank who works for Sky. Sky is Europe’s leading entertainment company, serving 21 million customers across five countries: Italy, Germany, Austria, the UK and Ireland.


VMware Advocacy

Introducing the New VMware Virtual SAN Partner…


Introducing the New VMware Virtual SAN Partner Campaign @vmwarevsan

Introducing the New VMware Virtual SAN Partner…

Visit www.vmwaredemandcenter.com for more information.


VMware Advocacy

Security for the New Battlefield


Security for the New Battlefield @VMwareNSX

Security for the New Battlefield

What will be our security challenge in the coming decade? Running trusted services even on untrusted infrastructure. That means protecting the confidentiality and integrity of data as it moves through the network. One possible solution – distributed network encryption – a new approach made possible by network virtualization and the software-defined data center that addresses some of the current challenges of widespread encryption usage inside the data center.


VMware Advocacy

A look inside VMware’s hyper-converged solutions


A look inside @VMware’s hyper-converged solutions

A look inside VMware’s hyper-converged solutions

VMware, Inc. is stepping it up when it comes to storage. The company announced at VMworld2015 not only the latest version of VSAN (virtual shared storage software) – 6.1 – but a beta version of the next evolution.


VMware Advocacy

Los dominios de fallos en vSAN 6


Los dominios de fallos en #vSAN 6

Los dominios de fallos en vSAN 6

Los dominios de fallos en vSAN 6¿Cómo crear un plan de contingencia con VMware vSphere? Parte 3/5VMware View Auto-Connection Utility para reutilizar nuestros viejos Desktops¿Cómo aportar capacidad en vSAN vía ESXCLI?¿Cómo crear un plan de contingencia con VMware vSphere? Parte 2/5VMware ViewDbChk al rescate del administrador de ViewLos recursos disponibles en VMware vSAN¿Cómo crear un plan de contingencia con VMware vSphere? Parte 1/5Mi experiencia con la certificación VCP6-NV (Network Virtualization NSX)El estado de las políticas en VMware vSAN Publicado por: Leandro Ariel Leonhardton: julio 10, 2015En: almacenamiento, Curso Online Gratuito, VMware vSphere, vSANSin Comentarios Unas de las características nuevas a destacar de VMware vSAN…Read More


VMware Advocacy

Restarting management services may result in VSAN network partition if LACP is configured


An issue was observed where a VSAN node gets partitioned after running at the host’s console:

#services.sh restart

The host rejoins the cluster after the services restat.

The root cause was attributed to restarting the “lacpd” service which results in tearing down the LAG group then re-forming it. During which time, the host is disconnected from the VSAN network. I will update this post when a fix is available.

ESXi 5.5 EP07 posted. A must have before upgrading to VSAN6 in certain configurations


A critical issue was reported with upgrading VSAN from ESXi 5.5 with Patch 4 or Express Patch 6 to vSphere 6 GA.

The problem was introduced in the above patches which attempted to fix an incorrect SSD capacity reporting when using RVC command vsan.disks_stats.

EP07 backs out the above fix.

Here is the possible sequence of events:

  1. Upgrade vCenter to 6.0
  2. Enter maintenance mode on one VSAN cluster node (Ensure availability option)
  3. Upgrade to ESXi 6.0 GA
  4. Exit maintenance mode
  5. vMotion some VMs to the upgraded node

The ESXi 6 node is unable to read the SSD UUID correctly on the remaining VSAN 5.5 nodes. This results in failure to access the VSAN objects on the 5.5 nodes and eventually, VMs become inaccessible and my disappear from vCenter inventory.

A look through the vmkernel.log file would show events like this:

2015-03-23T15:59:08.093Z cpu10:33023)WARNING: exprmsh: Error unmarshaling structure: Bad parameter
2015-03-23T15:59:08.093Z cpu10:33023)WARNING: Parse error at token 17. Message: Unexpected value type.
Expression up to error: …l0 i10 i15 l1600000 l0 l16777216 i0 52289c49-30c9-1ec7-16a9-57fc7ba21d7f “5422e4c9-32259101-000f-ac162d9dec28” l0

 

This problem is more evident with 3 nodes clusters. If there are more nodes, the chances of hitting this issue are less depending on FTT policy and on which nodes the replicas are stored.
Examples:

Cluster nodes count: 3
FTT=1

Components placements:
5.5 node 1: Replica
5.5 node 2: Replica
6.0 node: Witness

or

5.5 node 1: Replica
5.5 node 2: Witness
6.0 node: Replica

In either case, more than 50% of the components are on 5.5 nodes. As long as the associated VMs run on 5.5 nodes, they would be accessible. However, availability is not guaranteed if one of the remaining 5.5 nodes experiences a hardware component failure.

If you have 4+ nodes cluster, all 3 components (2 replica and a witness) may be on the 5.5 nodes. As long as the majority of the components are on the 5.5 nodes, the object should still be accessible.

The best approach to upgrading, with least disruption to the running VMs, is:

  1. Upgrade vCenter to 6.0
  2. Place one node in maintenance mode (Ensure accessibility option)
  3. Install EP07
  4. Reboot and exit maintenance mode
  5. Wait for resync to complete
  6. Repeat steps 2 through 5 for each node in the cluster.
  7. Place one node in maintenance mode (Ensure accessibility)
  8. Upgrade to ESXi 6.0 GA
  9. Exit maintenance mode.
  10. Wait for resync to complete
  11. Repeat steps 7 through 10 for each node in the cluster.

What if I already upgraded one host prior to installing EP07?

You have 2 choices:

  1. Shutdown all VMs on the cluster and upgrade the remaining nodes to 6.0. Make sure to choose “No data Migration” option when entering maintenance mode on the 5.5 nodes.
  2. Roll-back the 6.0 node to the previous version of 5.5

The first approach needs the least administrative effort but will disrupt all running VMs

The second approach is the least disruptive but will involve a rollback, installing EP07 then upgrading to 6.0

To rollback an upgrade, as long as you have not made any other configuration changes, you may follow VMware KB 1033604