NetBackup Virtualization Resource Limits
Alright everyone, I wanted to share another little gem that isn't really well known in NetBackup:
Resource Limits
This is available for both VMware and Hyper-V, and allows you to limit backups based on various options to prevent you from thrashing your virtualization farms. Sorry Xen, you're still not supported yet. You get to it from Host Properties -> Master Server -> Properties -> Resource Limit
First, let's look at the options available for the 800-lbs gorilla in the room, VMware:
vCenter, snapshot, Cluster, ESXserver, VMXDatastore, Datastore, DatastoreFolder, DatastoreType, VMXDatastoreNFSHost, DatastoreNFSHost and DatastoreCluster. Sounds promising right?
Let's go down the list:
vCenter - Maximum jobs that can run per vCenter - good if you have a small vCenter Server
snapshot - Maximum snapshot creates/deletes that can run at the same time - will not limit backups, only snapshot operations
Cluster - Limits jobs running on a vSphere Cluster
ESXserver - Limits jobs running per host
VMXDatastore - Limits jobs per datastore when VMX files span multiple datastores
Datastore - Limits jobs per datastore - good for preventing thrashing on LUNs with limited spindles
DatastoreFolder - Limits jobs per Datastore Folder
DatastoreType - Limits jobs per Datastore type (VMFS/NFS)
VMXDatastoreNFSHost - Limits backups per Datastore at a NFS Host level - need to look into more
DatastoreNFSHost - Limits backups per Datastore at an NFS Host level
DatastoreCluster - Limits backups per Datastore cluster
Wow, that's a lot, both to ingest and to type, so let me boil down what we do:
We worked with our Storage and VMware Admins to determine the maximum load that can be handled per environment. We then restricted based on Datastore and ESXserver. Resource Limits use the most restrictive method, so if you have 10 Datastores presented to 4 hosts, and you specified the typical setting of 2 each, then that means you will run a maximum of 8 jobs at any given time, two per host and no more than two jobs would come from any datastore, limits host overhead as well as disk thrashing. If you had the same settings but 6 hosts, then the maximum jobs would be 12, as the number hosts would still be the limiting factor.
More information can be found at the following link:
https://www.veritas.com/support/en_US/article.v47183490_nbu_vmware_web
Handy right? Options for Hyper-V are similar:
Active Snapshots Per Server
Active Snapshots Per Cluster
Snapshot Operations Per Server
Snapshot Operations Per Cluster
Active Backups Per Hyper-V Server
Active Backups Per Hyper-V Cluster
If you read through the VMware Section, these should be pretty self explanatory.
More Hyper-V Resource Limit info can be found at the following:
https://www.veritas.com/support/en_US/article.000121543
Hope this help balance your environment and preventing overloading any particular part of your environment
Charles
@whitehattechs
Resource Limits
This is available for both VMware and Hyper-V, and allows you to limit backups based on various options to prevent you from thrashing your virtualization farms. Sorry Xen, you're still not supported yet. You get to it from Host Properties -> Master Server -> Properties -> Resource Limit
First, let's look at the options available for the 800-lbs gorilla in the room, VMware:
vCenter, snapshot, Cluster, ESXserver, VMXDatastore, Datastore, DatastoreFolder, DatastoreType, VMXDatastoreNFSHost, DatastoreNFSHost and DatastoreCluster. Sounds promising right?
Let's go down the list:
vCenter - Maximum jobs that can run per vCenter - good if you have a small vCenter Server
snapshot - Maximum snapshot creates/deletes that can run at the same time - will not limit backups, only snapshot operations
Cluster - Limits jobs running on a vSphere Cluster
ESXserver - Limits jobs running per host
VMXDatastore - Limits jobs per datastore when VMX files span multiple datastores
Datastore - Limits jobs per datastore - good for preventing thrashing on LUNs with limited spindles
DatastoreFolder - Limits jobs per Datastore Folder
DatastoreType - Limits jobs per Datastore type (VMFS/NFS)
VMXDatastoreNFSHost - Limits backups per Datastore at a NFS Host level - need to look into more
DatastoreNFSHost - Limits backups per Datastore at an NFS Host level
DatastoreCluster - Limits backups per Datastore cluster
Wow, that's a lot, both to ingest and to type, so let me boil down what we do:
We worked with our Storage and VMware Admins to determine the maximum load that can be handled per environment. We then restricted based on Datastore and ESXserver. Resource Limits use the most restrictive method, so if you have 10 Datastores presented to 4 hosts, and you specified the typical setting of 2 each, then that means you will run a maximum of 8 jobs at any given time, two per host and no more than two jobs would come from any datastore, limits host overhead as well as disk thrashing. If you had the same settings but 6 hosts, then the maximum jobs would be 12, as the number hosts would still be the limiting factor.
More information can be found at the following link:
https://www.veritas.com/support/en_US/article.v47183490_nbu_vmware_web
Handy right? Options for Hyper-V are similar:
Active Snapshots Per Server
Active Snapshots Per Cluster
Snapshot Operations Per Server
Snapshot Operations Per Cluster
Active Backups Per Hyper-V Server
Active Backups Per Hyper-V Cluster
If you read through the VMware Section, these should be pretty self explanatory.
More Hyper-V Resource Limit info can be found at the following:
https://www.veritas.com/support/en_US/article.000121543
Hope this help balance your environment and preventing overloading any particular part of your environment
Charles
Comments
Post a Comment