Clients are observing high latency when Vscan is enabled due to the Vscan servers being overloaded
Applies to
- ONTAP 9
- CIFS
- Vscan
Issue
- Clients are observing high latency when Vscan is enabled
- Example: Jobs that normally take 10-12 minutes take over an hour
- High latency is seen in the
Network
column ofqos statistics volume latency show
infrequently- Note: Sometimes no latency is seen and must be checked against
qos statistics volume performance show
to see offbox latency
- Note: Sometimes no latency is seen and must be checked against
Cluster1::> qos statistics volume latency show -volume vol1 -vserver svm1 Workload ID Latency Network Cluster Data Disk QoS NVRAM --------------- ------ ---------- ---------- ---------- ---------- ---------- ---------- ---------- -total- - 5.52ms 4.41ms 15.00us 981.00us 113.00us 0ms 0ms vol1 10210 4.60s 4.59s 6.00us 91.00us 12.00us 0ms 0ms -total- - 85.00us 0ms 0ms 0ms 0ms 0ms 0ms vol1 10210 84.00us 0ms 0ms 0ms 0ms 0ms 0ms
- Errors may be seen like below:
Wed Jul 06 00:20:15 EEST [cluster1-02: kernel: Nblade.vscanBadUserPrivAccess:error]:
For Vserver "svm1", the attempt to connect to the privileged ONTAP_ADMIN$ share
by the client "10.1.2.3" is rejected because its logged-in user "VSCAN-admin"
is not configured in any of the Vserver active scanner pools.
Wed Jul 06 00:20:16 EEST [cluster1-02: kernel: Nblade.vscanConnBackPressure:error]:
For Vserver "svm1", AV server "10.1.2.3" is too busy to accept new scan requests.