Hi BertrandFR,
We don’t have quite same situation but our teams still thinking that snapshot is a safe way and they have backup during some tasks such as upgrade. They are still asking us to create snapshot before upgrade and keep the snapshot for some days or even weeks. We had lot of virtual machines with very old snapshots before.
I’ve explained for them that snapshot is not safe and they need to have backup with some restore points and we can restore last backup quickly by Quick Rollback but they are still asking and we have to do it for them.
Actually, when you have critical data on a VM like a database server, having standby server is a good solution to keeping data safe, also backup the data on different storage device is a complementary solution. Taking backup from VM or guest OS files can be a part of your backup strategy as well.
But about application servers, I think that create backup or snapshot can’t help in some situations like file system (VMFS) corruption and restoration may be takes more time compare to recreating application servers, it can be a waste of resources. The teams must have ability to deploy new application servers by taking backup from configurations/using configuration management tools and not the useless data.
VMware snapshots can be a pain! In production environments they should exist as short as possible.
To reduce snapshot time:
- Veeam offers the hardware integrated snapshot feature to reduce this time massively. VMware snapshots get deleted as soon as array snapshot is created.
To monitor snapshots:
- A vCenter alarm should be created to show warning/alarm at specified snapshot sizes.
- A simple script can be run - for example - once a day to list all snapshots, or even delete them.
To get rid of VMware snapshots
- VMware vVols can be used. With vVols, a snapshot triggered in vSphere is forwarded to the array that creates the snapshot natively.
@BertrandFR Basic task to do is to enable orphaned snapshot reports (comes with Veeam One). It will give you an insight into snapshots in your environment.
However, when Veeam tries to take a snapshot and finish the backup task, it consolidates the current snapshot by default. In some cases it won’t be able to do so but simply taking a snapshot and delete all snapshots after it will delete all snapshot chain.
@BertrandFR Basic task to do is to enable orphaned snapshot reports (comes with Veeam One). It will give you an insight into snapshots in your environment.
However, when Veeam tries to take a snapshot and finish the backup task, it consolidates the current snapshot by default. In some cases it won’t be able to do so but simply taking a snapshot and delete all snapshots after it will delete all snapshot chain.
Something to add, Veeam Snapshot Hunter, only works when the snapshot was made with a Veeam's tag, it does not work for other orphaned snapshots.
Are people still seeing frequent orphaned snapshots after upgrading to vSphere 6.x that has a new snapshot engine?
Are people still seeing frequent orphaned snapshots after upgrading to vSphere 6.x that has a new snapshot engine?
Actually, VMware changed how snapshots gets deleted. Before, a helper-snapshot was necessary to do so. Changed during deletion was written to helper-snapshot. To remove helper-snapshot another helper-snapshot was created and so on. This is done until the current helper was that small to commit with least disruption.
New with vSphere 6: There is no helper-snapshot needed any more. Instead changes are made to original and current snapshot-file simultaneously.
Are people still seeing frequent orphaned snapshots after upgrading to vSphere 6.x that has a new snapshot engine?
Sorry, to answer your question: With current version in 6.x commiting snapshots work quite well. Currently there are issuse when deleting snapshots in vSphere 7.0 U1. CBT can make troubles there on certain servers.
