Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials

Advanced Alert - HP device poller

After reading a number of conversations regarding UDP and OID's I have tried numerous configurations of both UDP definitions and Advanced Alert triggers to email notifcations on failed disk or failed power supply on HP Proliant servers.

Device pollers configured for OID:

1.3.6.1.4.1.232.3.2.5.1.1.6 - cpqDaPhyDrvStatus

Historical data is not kept, options "raw value" and "get table" are enabled with format: enumeration
I have tried the format text and none, along with using SNMP type GET

Advanced alert is configured with Custom Node Poller Property type:

PollerName equal to cpqDaPhyDrvStatus
Status is greater than or equal to 3

I have tried rate and numeric status

The goal being an email on a disk in any other status than 2 (ok). Thus far any HP proliant the poller is assigned to will send an email regardless of disk health. The alert will identify the correct number and status of disks installed.

I have another poller and alert configured for power supplies, that appears to work fine with the same configuration in place for UDP and in the advanced aert.

1.3.6.1.4.1.232.6.2.9.1 - cpqHePwrSupplyCondition

The alerts send on status other than 2 (ok), and it has been tested to work correctly.

Any advice or tips would be most appreciated. My goal is to consolidate 2 or 3 monitoring systems into one, under our Orion NPM v10.1 umbrella!

Find more posts tagged with

Accepted answers

All comments

smargh

1/ For pollers with enumerated values, "Status" is the text result, i.e. "ok" - so, criteria of "status > 3" won't work. Change to either "rate not equal to 2" or "status not equal to ok"

2/ You said: "The alert will identify the correct number and status of disks installed." I doubt it will, unless you get very crafty with custom ${SQL: ...} within the emailed alert message body.

3/ I use cpqDaPhyDrvStatus rather than cpqDaPhyDrvCondition. I have no idea what the difference is between the two, but cpqDaPhyDrvStatus has so far identified all failed disks. I've added cpqDaPhyDrvCondition and everything is coming back as "ok"

4/ For finding specific things within rows in SNMP tables, you must use "GET TABLE" - otherwise Orion will only look at the first row. For example, in the data (ok,ok,failed,ok) Orion would only look at "ok" in deciding whether to alert or not. [E&OE - unless this has changed in recent versions - I did my trial & error diagnostics in v9.5.1]

These are my alerts for HDD and PSU status for HP servers - they do work fine. Note that I've changed the enumerated text a little bit on my pollers, so change the criteria as appropriate for you:

I use this SQL query to find out the appropriate criteria for triggers, so that I don't get annoyed trying to figure out whether to use status, rate, numeric status or raw status. Change the last line to the name of the custom poller. Maybe there's an easier way, but it's what I use.

SELECT TOP 700 [CustomPollerAssignmentID]
      ,[DateTime]
      ,[Rate]
      ,[Total]
      ,[RawStatus]
      ,[Status]
      ,[RowID]
FROM [CustomPollerStatus]
where CustomPollerAssignmentID IN (SELECT [CustomPollerAssignmentID]
                                   FROM [SolarWinds].[dbo].[CustomPollerAssignment]
                                   WHERE [CustomPollerID] IN (SELECT [CustomPollerID]
                                                              FROM [SolarWinds].[dbo].[CustomPollers]
                                                              WHERE [UniqueName] = 'cpqDaPhyDrvStatus'))

YourVillageIdiot

At risk of smearing mud on my own face, it was a small detail of "Trigger Alert when all of the following apply" on the condition group.

Having tried the configuration with "SNMP get table" and "status is not equal to ok" I had overlooked "any" on the condition group until this morning. Fresh eyes, and reviewing someone elses suggestion provided more focus.

Thanks

LGentile

Here is an example of my disk alert:

And OID:

Basically, anything outside of "online" is what we look for. I always use Status/Numeric status. I have seen some come back with text, so i included the text fields as well for 'failed' although they are rare. We have a lot of different ProLiant models, but this has proven to be effective. I have similar alerts for PSUs, RAID battery, DIMM and the same for Dell servers, which are more complex due to changing OIDs over time, but all my alerts work in the same fashion.

Hope this helps.

edruiz

Thanks for the useful info. Out of curiosity, how did you format your message? I'm not sure as to which variable to show the drive status.

cgregors

I assume you mean in one of the alerts that get sent to the unix support staff. I have an alert defined like this:

Alert Name: OSS : Filesystem Capacity Critical (24x7) > 90%

trigger conditions:

node status = up
volume percent used > 0
volume percent used > 90
volumetype = Fixed Disk
Volume Responding = Y
custom property (support group) = unix team

the email action is as follows:

Subject:

ORION : Volume ${Caption} on ${Node.Caption} utilization > 90% utilization

Body:

Volume name = ${Caption}
Volume % utilization = ${VolumePercentUsed}
Node name = ${Node.Caption}
IP address = ${Node.IP_Address}
Node Contact = ${Node.ContactName}
alert triggered at ${AlertTriggerTime}
Alert generate by orion alert "${AlertName}"

I like adding the AlertName at the end of each email so I can troubleshoot where an incorrect alert came from. After a while there a too many to keep track of.

Chris.

YourVillageIdiot

To update status on our use of custom pollers regarding disk and power supply health. We are using seperate notifications on system down events, and component status. We've kept the information in the email alerts to basic information necessary for our service desk to assign workorders via Infra. The sys admin when a work order is received then has the email alert to review along with direct links to review current status of the server and components.

We are working on the automation process to document such incidents, assignment of work orders, and closure should an incident be acknowledge or resolved.

The HP OID's used are referenced in the trigger conditions:

The email notification:

Subject: NMS2:Component:failure:${NodeName}

Message Body (HTML formatting):

<table cellpadding="2" bgcolor="yellow">
<td height="15" valign="middle" width="300"><center><strong><font face="Arial, Helvetica, sans-serif" color="black" size="3">HP COMPONENT FAILURE</font></strong></center></td>
</table>
<strong><font face="Arial, Helvetica, sans-serif" color="black" size="3">Name: ${NodeName}</font></strong><br />
Status: ${Node.Status}
IP: ${Node.IP_Address}
Location: ${Node.DeviceLocation}

Review current status online:
http://netmon01/Orion/NetPerfMon/NodeDetails.aspx?NetObject=N:${Node.NodeID}

REview current HP component health:
http://${Node.DNS}:2301

CT Primary Contact:   ${Node.CT_Primary_Contact}
CT Secondary Contact: ${Node.CT_Secondary_Contact}
CT Team:              ${Node.CT_Team}
Business Contact:     ${Node.Business_Contact}
Business Group:       ${Node.Business_Group}

Time of detection: ${AlertTriggerTime}
Time of notification: ${ShortDate} ${LongTime}
Alert Name: ${AlertName}

MIB: ${CustomPollers.MIB}
OID: ${CustomPollers.OID}
#

Following link to node status in Orion the Sys Admin can review HP Components: