I recently discovered and resolved an issue I was having in one of my ONTAP simulator labs. I'm usually running 2, 3 or 4 single-node ONTAP clusters. The performance of my sims was tanking at what seemed random times. I eventually discovered it was happening at 10am each morning (and at some other off-hours times). I eventually discovered the cause was heavy IOPS on vol0 which was being caused by ONTAP's system configuration backup process.
SCOPE: The changes below should only be needed on ONTAP simulators. I don't believe actual ONTAP systems running on decent hardware will ever have this same problem. This is likely only an issue on cramped lab environments like a single ESX node with a single SATA disk datastore (ex: my lab). On a laptop with VMware Workstation or Fusion the symptoms are possibly even worst. However, on an SSD-based laptop you might not even notice it is happening (they are quiet and fast). Just don't schedule demos between 10:00am and about 10:45am. 🙂
SUMMARY: By default, ONTAP systems run regularly scheduled system configuration backup jobs that gather data on each node and then consolidate it for the entire cluster, all the while hammering vol0(s). For a simulator, or worst yet, multiple simulators running on a single hypervisor this can be a strain.
The changes suggested below simply cut down the number of times it runs and reduces the number of copies retained. You can read more about the issue here in the standard ONTAP forum where I originally sought help --- http://community.netapp.com/t5/Data-ONTAP-Discussions/What-are-these-snapshots-that-appear-on-my-vol0-and-system-config-backup/m-p/125871#M27126
CHANGES: The commands below show the settings before, change the settings, and then view the settings afterwards. I've tested these on ONTAP 9 and 8.3.2 but I suspect they are the same for prior releases.
---
#
set advanced # advanced priv level commands follow
#
# View settings before
#
system configuration backup show # shows current config backup files
system configuration backup settings show # shows optional URL and Username
system configuration backup settings show -instance # shows # of files retained, schedules
job schedule cron show # shows schedules (8hour, daily, weekly)
#
# Changes to reduce the impact for a simulator environment
#
job schedule cron modify 8hour -hour 18 -minute 15 # Change to once a day at 6:15pm (up to you)
y
system configuration backup setting modify -numbackups1 1 # Change 1st (8hour) to 1 copy
system configuration backup setting modify -numbackups2 1 # Change 2nd (daily) to 1 copy
system configuration backup setting modify -numbackups3 1 # Change 3rd (weekly) to 1 copy
#
# View settings afterwards
#
system configuration backup show # shows current config backup files
system configuration backup settings show # shows optional URL and Username
system configuration backup settings show -instance # shows # of files retained, schedules
job schedule cron show # shows schedules (8hour, daily, weekly)
#
set admin # return to admin mode
---
The number of retained copies will eventually drop from 2 to 1 as the backup jobs execute and purge older copies. However, if you are anxious to get the space back sooner you can also delete older copies explicitly with this command:
#
system configuration backup delete <backupName>
I could not find a way to completely disable the backups or change them to use some other named ONTAP cron schedules. If anyone figures out a way to do that, let us know.
