How To – RedLogic Tech Blog series
At RedLogic we like to share our knowledge. You can find us at many events as speakers on various VMware related topics. We also organize inhouse events for our Dutch clients and Tech Days for exchanging technical knowledge between colleagues. We feel it’s time to also share our extensive knowledge with other tech enthusiasts all over the world. At the moment we’re working hard on developing a technical blog website, creating a platform for our experts to share their knowledge with you. Below the first of many blog posts, stay tuned for more.
Identify NVMe disks in an HPE vSAN environment
Replacing a failed disk or identifying an internal disk to be replaced is always a delicate task, because we don’t want to introduce additional failures. Especially in vSAN enabled clusters with different failure to tolerate settings. Of course, a vSAN enabled server can be placed into maintenance mode with full data migration and VMware has done a great job in creating visibility of the used components in a vSAN environment, including the identify disk led.
But sometimes things just don’t add up and you need additional confirmation to be absolutely sure that the correct disk is replaced. Therefore I was looking for a way to match the information displayed on the ILO page of an HPE DL380 Gen10 with the information displayed in vCenter for a vSAN NVMe disk.
On the ILO page the model and serial number are shown for a given NVMe disk.
In vCenter the model number for an NVMe disk can be easily retrieved. But unless you have completely different disks in the servers, the model number isn’t very useful.
The NVMe disk as presented in vCenter shows as t10.NVME__<model number>_______<GUID>. This GUID can’t be found in the ILO. So how can we retrieve the serial number for a specific NVMe disk?
From the ESXi CLI several commands can be used to retrieve information about a disk drive. Since this is a vSAN host I started with the standard vSAN command: esxcli vsan storage list
With this command the model number can be easily found, but the serial number isn’t listed.
After some Google work and a discussion with an HPE employee we found that we needed the command: esxcli nvme device list instead.
With this command the NVMe disks and their related vmHBA numbers are shown. When the vmHBA adapter is added to the previous command, we get the serial number of the disk: esxcli nvme devicve get -A vmhba1
The last step is to relate the vmHBAx number to the correct disk.
This can be done with the command: esxcfg-scsidevs -A
Now we can be completely confident that when we’re replacing the disk, we will get the correct one. This is also extremely useful when instructing a colleague on-site, or when using the remote hands in a datacenter.
Hopefully these steps provide you with some additional guidance and insurance when replacing or identifying NVMe disks.
Robin van Altena
Infrastructure Consultant at RedLogic
With over 15 years of experience in virtualization Robin is an experienced VMware Consultant. He has worked on designing, building and supporting various environments. Robin focuses primarily on Software Defined DataCenters, his expertise is mostly on vSphere and NSX(V+T), but he is also familiar with vSAN and other VMware products.
He is a modest blogger, public speaker and vExpert for the last two years. In his spare time, he likes practicing Aikido or watching movies.
