SQL server probe - mirror_state and mirror_sqlinstance throwing alerts for healthy conditions also
book
Article ID: 127015
calendar_today
Updated On:
Products
DX Infrastructure ManagementNIMSOFT PROBES
Issue/Introduction
We have sqlserver probe version 5.XX Monitoring SQL servers using SQLserver probe from a remote computer. The DB Servers are configured for mirroring. As soon as the probe is deployed, alert triggers for healthy conditions as below. Sample alerts: Profile XXX, instance XXX, status of SQL server instance YYYY hosting the mirror database XXX is 1 (0-unavailable, 1-available) Profile XXX, instance XXX, state of mirroring database YYYY is SYNCHRONIZED
The metric configuration looks correct for: mirror_state mirror_sqlinstance
Can you please let me know why we receive alerts for healthy conditions?
Cause
Client had added unit values to the two check points in their Stardard Archive template they used from distsrv to setup the sqlserver probes.
Environment
UIM 9.02 and ealier sqlserver probe 5.42 and earlier
Resolution
1) there are two checkpoints causing an issue: mirror_state mirror_sqlinstance 2) these were sending alarms as soon as the probe was started. 3) when you install the probe clean these two checkpoints do not have any unit values in them. 4) in your .cfg file we can see the following unit values have been set: <mirror_state> <thresholds> <default> <0> tagid = 0 value = 4 unit = hours sev = major msg = mirror_state_2
<mirror_sqlinstance> <thresholds> <default> <0> tagid = 0 value = 1 unit = days sev = major msg = mirror_sqlinstance_2
5) The unit = should be <blank> for both of these 6) using raw config we changed both of these to be <blank> and restarted the probe and saw that the alarms were no longer generated. 7) on the call we learned that you have a template package in the archive for your sqlserver deployments setup. 8) you will need to edit this package to remove the units so that when this is deployed in the future it does not cause the problem.