SpectroSERVER lost contact with the Archive Manager during the OnLineBackup - post_olb_script

Products

• CA Spectrum
• DX NetOps

Issue/Introduction

Every Sunday, around 12AM PDT we see the attached events in one of our SpectroSERVERs, because of which the servers won't be connected to the primary archive manager and devices won't be reachable for some time. Can you please let us know what might be the reason for this to happen?

Environment

Release : 10.4.1

Component : Spectrum Alarm & Event Management

Cause

The Archive Manager is stopped every time the DDMdb backup is launched via $SPECROOT/custom/CsScript/post_olb_script (which is actually launched by the OnLineBackup - OLB).

There is an incorrect customization in your $SPECROOT/custom/CsScript/post_olb_script file.

You have uncommented the section (from rows 53 to 63) that launches the ddm_backup. This section should be uncommented only when the Spectrum release is below 10.3.1.

The following rows in a red box must be commented out.

The post_olb_script below is from Spectrum 10.4.0 installation. A fresh installation of Spectrum 10.4.1 does not have both sections, it only has the ddm_save.pl section.

As you are in 10.4.1 release, you have to uncomment only the section that launches the ddm_save.pl (from rows 70 to 80) .

Resolution

This is what you should have in the $SPECROOT/custom/CsScript/post_olb_script file:

Rows from 53 to 63 must be commented out.




You have to also follow this KB article to avoid losing the events when the Archive Manager is unavailable:

Title: CA Spectrum Event Storage Best Practices
https://ca-broadcom.wolkenservicedesk.com/external/article?articleId=20967