Notifications

203 views

Description

When Shazzam launches Credential-less Discovery, it makes multiple copies of the input payload for each IP it scans. When a large number of IPs are alive but not active, the input payload can be large and making multiple copies for each IP can use hundreds of MB of memory causing slow response times.

 

Steps to Reproduce

  1. Scan an IP Range with the Shazzam batch size of OOB 5000 IP Addresses for at least 40 K IP Addresses.
  2. Most of the IP Addresses would lead to Credential-less Discovery being launched, meaning the Shazzam batch size would be sufficiently large, approximately 3 to 5 MB.
  3. Have Shazzam processing slow enough so that all eight worker threads are processing Shazzam inputs: One for each Shazzam input and each input with 5 K IP Addresses.
  4. Once the above condition matches and all eight worker threads are processing concurrently, observe the memory usage for each worker thread.

Workaround

Reduce the Shazzam payload batch size for the schedules that you suspect will use Credential-less Discovery. 

If the Credential-less Discovery feature is not intended to be used, set the property "mid.discovery.credentialless.enable" to false which will disable Credential-less Discovery.


Related Problem: PRB1336879

Seen In

SR - IRM - GRC Profiles - Madrid 2019 Q2
SR - IRM - PA Premium Integration - New York 2019 Q3
SR - ITOM - Cloud Management Google Cloud Connector - v1.0
SR - ITOM - Cloud Management Terraform Connector - v1.1
SR - ITOM - CMDB CI Class Models - 201908
SR - ITOM - Discovery and Service Mapping - 201908

Fixed In

London Patch 9
Madrid Patch 4
New York

Associated Community Threads

There is no data to report.

Article Information

Last Updated:2019-10-06 22:24:23
Published:2019-05-09