Introduction
This article is intended to discuss the topic of snapshots as it relates to VMs within the EHIS VMware Infrastructure.
Instructions
From a vendor perspective, the VMware Product does provide the feature of taking snapshots of those VMs configured in a specific manner, however EHIS does not support the use of snapshots on any of the VMs managed by EHIS regardless of whether it is configured to accept snapshots. The reasons are many including but not limited to cost and instability.
1. Snapshots require disk space - not budgeted into current architecture
2. The presence of the snapshots leads to backup failures.
3. If back-end storage fills up due to the snapshots, it will impact all VMs on the shared storage which will cause one or more VMs to become inaccessible and could lead to one or more major incidents and unspecified impact.
In lieu of snapshots, EHIS utilizes our Backup Infrastructure for backup and restore purposes. It is strongly recommended, especially when there is concern about a change, that a Service Request is submitted to the appropriate OS Hosting Team (Windows Hosting Services or Linux Hosting Services) to validate that a good backup exists before the change is implemented. If needed, a one-off backup can be requested closer to the change window to ensure the most recent copy of the server has been obtained in the Backup Infrastructure. If the system has a Database, please open a Service Request with the appropriate Database Support Team (SQL Server Services, Oracle Services or another appropriate team) as the OS Hosting Team's backups do not include database backups.
Note: Unless otherwise requested (special cases), backups for all systems start at 1800 each evening, 7 days a week and are validated for success/failure during business hours each work day, Monday through Friday, except in the case of an organizationally recognized holiday. In the event of a backup failure, steps are taken to remediate the issue, such that the backup will run successfully the following evening, so there is the potential for a backup not to exist for a given date. Once backups start at 1800, they are queued and can run in a randomized order; this is non-configurable (no way to control the order). Backups exist for 60 days or 60 backups, whichever comes first.
If you have any further questions, please open a Service Request with the appropriate OS Hosting Team.