12-01-2020 09:49 PM
VM backup is configured through VMware policy. During Vmware backup storage utilization is so high.
what would be the possible ways to resolve this?
1. Can we change the backup window of VM policies so that it will not start on a same time?
2. we can change the backup window only between the authorized time given by client. But can we run the backup 24 hrs? is it increases the load on server?
i think server should need some idle time to relax and to sit idle. if it works for 24 hrs then may be it will crash or giving bad performance. is it right?
Kindly share your thoughts on this issue
12-02-2020 03:57 AM
High Storage utilization is probably the snapshots taking up space.
Please have a look at Resource Limits:
You can limit the amount of concurrent snapshots per vCenter, max jobs per datastore, etc. etc...
Have a look at the list of resources that you can limit, then discuss with your VMware admin which resources to limit and what the value should be.
If you limit resources, then only a limited amount of backups will go active. When one job completes, another one can go active, as per the Resource Limits that you configure.
Servers are made to be running 24 x 7.
If you do not overload them during the backup (by implementing Resource Limits), you should not see impact on performance.
12-02-2020 05:28 AM
Currently VCenter limit is set to 50, ESX server limit is set to 8 and Datastore limit is also set to 8.
Here, we need to reduce the Vcenter limit or datastore limit?
Snapshot is created on datastore so do we need to reduce the datastore limit or both?
12-02-2020 05:45 AM
And snapshot limit is set to "No limit". Do we also need to limit the snapshot?
12-02-2020 07:08 AM
You really need to discuss parameters with your VMware admin.
Can the VM admin confirm that the snapshots are causing the increase in storage utilization?
I have honestly never seen anyone with concurrent jobs on Datastore set to 8.
I have seen a reccomendation of no higher than 4, with some environments even limiting it to 1 or 2 per datastore.
Again - speak to your VM Admin.
You need to agree on values that work for your unique environment.
When you make changes, monitor and ducument the results before making more changes.
12-02-2020 04:46 PM
These are all valid questions but they are not really relevant to backup as such - the question you need to ask is the I/O profile of the apps and how much space and IOPS they need on the datastore while backup takes place. Remember that these values defined by the application that runs inside those VMs and not NBU or any other backup app. For a fairly static VM and fast flash based datastores you might be able to back them up in dozens simultaneously from the same datastore without much of a problem but for a high-volume transactional VM on a spinning rust even one backup might be too much. All these questions needs to be discussed with your VMware admin in the context of how much I/O shares you are able to allocated to backup and set the limits accordingly.
Do not forget that in NBU limits will be respected by VIPs only. Separately scheduled backups that use individual VMs in VMware policy which is not a VIP will not consider these limits.
12-02-2020 08:57 PM
also backup is running through NBD transport mode. Can we configure them through SAN mode? will this change will help here?