Scheduled Snapshots of ZVOLs (iSCSI, ESXi)

[Scampicfx]

Dear guys,

my san consists of two zvols:
san\esxi-boot
san\esxi-data0

esxi-boot is a small zvol where ESXi boots from.
esxi-data0 is a large zvol which hosts all VM-files

1.) Let's assume, I configure a simple scheduled snapshot task for esxi-boot. Are there any certain points to consider? This esxi-boot zvol gets used via iSCSI. I'm wondering what happens when e.g. a large file operation on this zvol takes place while the snapshot is being taken (I know, a large file operation on esxi-boot is a bit unrealistic ;) - but I'm wondering if there are any scenarios which could lead to a snapshot which is unusable in case of restore? Do I see it correctly that a snapshot could get triggered exactly during an installation proccess of e.g. a driver? Restoring such a snapshot could lead to an unpredicted behaviour of the ESXI host?).

2.) Initially, I tend to use regular scheduled snapshots for esxi-boot and schedulded vmware-snapshots for esxi-data0.
But: I'm wondering what outcome to expect when not using vmware-snapshots but instead using regular scheduled snapshots for esxi-data0?

Documentation says:
Once this type of snapshot is created, FreeNAS® will automatically snapshot any running VMware virtual machines before taking a scheduled or manual ZFS snapshot of the dataset or zvol backing that VMware datastore. The temporary VMware snapshots are then deleted on the VMware side but still exist in the ZFS snapshot and can be used as stable resurrection points in that snapshot. These coordinated snapshots will be listed in Snapshots.

3.) Okay, do I see it correctly that each VM gets is own snapshot? Is this the difference between regular scheduled snapshots and vmware snapshots?
Regular Scheduled: Complete zvol gets backuped. In case of restore, maybe complete volume has to be restored?
vmware snapshot: individual vms get backuped?

Are there any other certain points to consider? :) Are vmware snapshots possible on ESXi free edition? (maybe I should have asked this question at first... hmm... )

Thank you! :)

EDIT: I found one piece of information which gives some proof about the possible data corruption, when the snapshots are triggered mistimed. https://redmine.ixsystems.com/issues/788 (The report itself is 6 years old, but I think, there is no change from technical point of view how snapshots are being executed).

Quote:
[...] but just apply a big fat warning about potential data corruption (unless one manually pauses the virtual machines).

Any other / additional information are highly appreciated, though!

 

[Ericloewe]

The big point is that VMware snapshots attempt to coordinate things to minimize the chance of the snapshotted data being dirty

Unfortunately, there is no magic universal way of properly doing the coordination, so, even then, results are not perfect.