High CPU utilization on a Snapmirror Destination Node
Applies to
- ONTAP 9
- Snapmirror DP relationships
Issue
- Latency is seen when high CPU utilization occurs due to Snapmirror job in progress.
-
The top workloads for D blade utlization was WAFL SCAN, wafl cp , WAFL and snapmirror as per perf archive graphs.
-
Output of
wafl scan status
shows scanners like snap create summary update; block ownership calculation and volume deswizzling ( which is for snapmirror) scans happening. -
Since this is the destination for snapmirror, so these scanners are expected.
::>system node run -node * -command wafl scan status
Node: xxxx
Volume xxx:
Scan id Type of scan progress
603754 snap create summary update block 496578 of 676206
604189 block ownership calculation block 469500 of 676205
Volume xxy:
Scan id Type of scan progress
604343 snap create summary update block 6452 of 45690
604265 blocks used summary update block 24304 of 45690
Volume abc:
Scan id Type of scan progress
604291 snap create summary update block 7689 of 59397
604234 blocks used summary update block 13233 of 59397
55 volume deswizzling snap 366, inode 183826 of 33554416. level 1 of normal files. Totals: Normal files: L1:1798785/3219086468 L2:6755/3234086207 L3:3369/1298071658 L4:0/0 Inode file: L0:0/0 L1:0/0 L2:0/0 L3:0/0 L4:0/0