We have a large amount of VMs on NetApp, have to backup all CIFS share, VMs, SQL and Exchange and other applications daily. Lately we have some VM performance issues, we doubted the backups, specially VM, big SQL and Exchange backups generated too much read traffic and slow the NetApp performance. can someone share your experience and solutions?
thanks in advance.
Solved! See The Solution
can you share your infrastructure design?
in that way we can help you figured out a possible solution.
there are many example of backup design that can alleviate the impact on the primary storage array.
Hope this help!
thanks for all replays.
Here is our details?
1). NetAPP AFF300
--one pair nodes for CIFS and NFS, 24 disks, one aggregate on each node, partitioned disks
volumes distributed on both aggregates
scheduled snapshot policies on each volumes
--second pair nodes for ESX and other SAN provisioning, 48 disks, one aggregate on each node, partitioned disks
--third pair nodes for an mission critical app, 24 disks
Veritas NetBackup, NDMP backups from nodes for all CIFS, VMs, SQL, Exchanges...
Hard to say with just that. Please follow those KBs to see if ONTAP is responsible.
Did you try what I said?
Do you have AIQUM, OCUM, Nabox?
Have you looked into using QoS on the volume doing the backup to limit how much impact there is to performance?
Also, if that volume is a FlexGroup volumes, consider disabling the share property "changenotify" - that property has been known to cause performance impact on a system.
Hello my friend,
If we're talking about traditional backup, reading blocks and files from the filer, yes, there's a good possibility of it being the culprit.
Backup operations normally puts significant pressure on the array. Of course, some feels it harder when having small aggregates or slow disks. But in my experience, almost 100% of my customers have a turbulent environment when in backup windows.
If you have AIQUM or NAbox, you could easily check if during the backups you latency and disk utilization is putting your VMs in bad shape.
If that is the case suggestions would be (some may not apply, it all depends on your environment):
- reschedule backups so they does not compete for resources
- add more disks
- rethink your volume placement
- you could change your env making snaps in a primary storage and replicating efficiently with snapmirror/vault to another filer.
- the list goes on....
Not sure if 7-mode or CDOT, but try these KBs to start with: https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/How_to_troubleshoot_performance_issues_in_ONTAP_9 and https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/How_to_troubleshoot_performance_issues_in_Data_ONTAP_8_7-mode