It is not good practice for a vSphere administrator when his production VM are having problem due to datastore storage availability issue. There is couple of things which leads to this storage problem.
- When you are using thin provision storage in your environment and it’s overcommitted.
- Using snapshot in your environment.
Know issues when datastore is out of space
When your storage has space issue then you will face following VM issues
- Cannot power on a VM.
- Cannot create a new VM.
- Cannot create snapshot of a VM.
- Cannot vMotion a VM.
- Current VMs in the datastore are suspended.
- VMs cannot read/write on the Datastore.
Working and analyze the datastore alarm
In above previous objective of blue print “Configure datastore alarms” .I have created an alarm which analyze and alert you on vSphere as well email the the vSphere admin when datastore is overcommitted at certain level. Datastore provision alarm is helpful mostly for those datastore which have support for thin provision. It is because NFS has the ability to over commit the storage from his defined size in the vSphere.
In “Configure datastore alarms”, I have created a datastore alarm with following condition. It will trigger and email your as well show you a warning sign on datastore in vSphere when datastore is provisioned above 100% and turn into red when reached above 150% of the actual datastore capacity.
Identify the Storage Issue?
As you can see below a yellow warning sign has appeared on my one of datastore name “SSD-Tag-Testing”
What cause the alarm?
Alarm triggered because this datastore is supported thin provisioning. If you see below in the general setting of the datastore and specifically look the “capacity” section. Datastore total capacity is “9.50 GB” and provisioned space is “11.94 GB” which is more than his 100% of his available capacity.
If you go into the event log of this datastore it will give the addition information about the alarm.
How to resolve Storage space Issue?
There are some solutions available when you are having this problem
- Create an alarm for datastore for storage monitoring.
- Extend the datastore space (increase the LUN space if arrays supported or add another extend).
- Use storage vMotion to migrate some VMs to other datastore.
- Delete some unused ISO and snapshots.
- If your VM are paused than resume it.