Kevin Justin's Blog

Azure Log Analytics Service Map Planning and Pre-reqs

My grandfather said two things:

An ounce of prevention is worth a pound of manure

Death and taxes are part of life

Planning out a deployment is a good thing.

My best friend would say “No one plans to fail, they just fail to plan”

This will be a multi-part blog – breaking out the high level steps, and my experience getting the solution set up.

What do we need for Service Map?

Azure connectivity
- Setup Log Analytics workspace on MMA/SCOM agent article
- Troubleshooting onboarding issues KB,
  - Check for Events in Operations Manager event logs blog
Computers in scope for visualization
- What computers (Windows or Linux)
- Pricing FAQ
Dependency agent installed on computers
- Configure
Azure Service Principal
- (think of it as an SSH shared key ID/password for Azure Apps to communicate)
- Docs article

High level steps

Overview blog
Install the MMA agent blog
Install the dependency agent blog
Configure Azure Service Principal blog
Configure Service Map on SCOM blog

Possible SQL issues affecting SCOM performance

Good reasons for a Risk Assessment

SQL RAS runs 800+ queries to check on target SQL servers

Check Best Practice Recommendations (BPR)

May be good opportunity to audit the SQL build for BPR!

Ran across some good examples where SQL settings brought SCOM to a standstill

One was Cardinality Estimation – basically, predicts how many rows a query will return

Part of SQL since 1998 with SQL Server v7.0

Let’s figure out what SQL2016 runs OoB (out of box)

SQL 2016

SELECT ServerProperty(‘ProductVersion’);
GO

SELECT name, value
FROM sys.database_scoped_configurations
WHERE name = ‘LEGACY_CARDINALITY_ESTIMATION’;
GO

The other is CLR Strict Security

SELECT * FROM sys.configurations

WHERE name = ‘clr enabled’

Talking with Shawn Nakhostin – SQL PFE, we discussed opportunities and questions around SQL optimization and best practices.

Shawn gave me the following feedback on customer performance issues:

I’ve found some customers who have had performance issues with SQL based on organizational SQL settings:

Trace flag 9481
CLR Strict Security is by default enabled

Trace flag 9481

Enabling or disabling this TF is not a matter of best practice.

The customer should see what works for them.

Here is the explanation:

Customer started using a new cardinality estimator in SQL Server 2014.

The product team knew that the new CE improved some of the query plans, but not all of them. In other words, they knew that this would improve overall query performance in “some” environments but might have a different impact in other environments.

For this reason, they created TF 9481 so that environments that see query performance degradation after upgrading SQL Server from version 2012 and earlier, they can turn on this trace flag so that the query optimizer uses the old algorithm for CE.

Note:-Trace flag 9481 forces the query optimizer to use version 70 (the SQL Server 2012 version) of the cardinality estimator when creating the query plan.

https://blogs.technet.microsoft.com/dataplatform/2017/03/22/sql-server-2016-new-features-to-deal-with-the-new-ce/

https://support.microsoft.com/en-in/help/2801413/enable-plan-affecting-sql-server-query-optimizer-behavior-that-can-be

CLR Strict Security is by default enabled

This causes all assemblies to be treated as unsafe.

As a result, assemblies will not load.

To get the assemblies to load they can do one of the following:

Sign the assembly. This may work if you have a few assemblies but becomes a huge task if there are many assemblies to sign.
Set the TRUSTWORTHY database property to on.
- This is not recommended because in some form defeats the purpose of using CLR Strict Security.
Add the assembly to the trusted assemblies list.
- This is called whitelisting, which may be a better option than the previous two.

https://docs.microsoft.com/en-us/sql/database-engine/configure-windows/clr-strict-security?view=sql-server-2017

Updated Skype for Business 2015 (premise) Addendum MP

Updated and completed for Company Knowledge!

Many thanks to Nick Wood for his help deciphering user impact for these alerts.

Reach out to Nick on LinkedIn

This has been an arduous effort to complete the Skype alerts and components.

Gallery Download

It’s taken a few steps along the way, to get all the content delivered.

To read the whole Skype Addendum journey, please read these additional blog posts
June blog
January blog

The initial Addendum pack with just service recoveries and Azure overrides

Old pack https://gallery.technet.microsoft.com/Skype-for-Business-2015-b005f49f
This download disabled Azure, set service recovery tasks

The new Gallery download contains the following:

Skype SCOM Alerts.xls
Microsoft.LS.2015.Monitoring.ComponentAndUser.Addendum.xml
Microsoft.SystemCenter.Notifications.Internal.xml
Skype.for.Business.Server.Management.Pack.Alert.Grooming.xml

NOTE The Skype.for.Business.Server.Management.Pack.Install.txt file contains the information as well

#############################################################
#
# Breakdown of files
#
#############################################################

#
# Skype SCOM Alerts.xls
# Skype SCOM Alerts XLS is is MP Export excerpt formatted for XLS workbook
#
# Feel free to search this file

# Column D is ‘Escalate to Who’
# This has values as SCOM Engineering, Messaging Ops, Telephony, Messaging Engineering
# Column E is impact
# This has values: P3-P5, *Email
# Column F is the Display String for the monitor
# Column H has the User Impact, Cause, and Troubleshooting steps

#
# Microsoft.LS.2015.Monitoring.ComponentAndUser.Addendum.xml
# Addendum management pack sets up company knowledge tab for each Skype monitor, with actionable troubleshooting steps.
#

#
# Microsoft.SystemCenter.Notifications.Internal.xml
#

# Backup the current MP first, and merge if you are adding this to your environment!
# Use this pack cautiously, as it will replace existing Channels, subscribers, subscriptions.

# On Management server, open PowerShell window as Admin
#
# cd <path>
# Example

cd $HOME/desktop
Get-SCOMManagementPack -Name *Notification* | export-SCOMmanagementpack -path “C:\”

Copy-item .\Microsoft.SystemCenter.Notifications.Internal.xml .\Original-<CompanyName>-Microsoft.SystemCenter.Notifications.Internal.xml

# Save the bundled Notifications pack to the same path

# YES it’s that important, the file can eliminate any alerts leaving SCOM!
#
# Save file to local drive

# Follow MP Fragment authoring if you need to merge existing Notifications with Skype pack
# https://kevinholman.com/2016/06/04/authoring-management-packs-the-fast-and-easy-way-using-visual-studio/
#

#
# Skype.for.Business.Server.Management.Pack.Alert.Grooming.xml
#
# This file is to set alerts per the XLS, making warnings for P4,P5, email, and P3 for Critical

#
#############################################################

What ID’s is SCOM using

Ever need to audit what ID’s SCOM is using?

Maybe you have to figure out how someone else setup SCOM.

Did they set up SCOM as recommended for best practices with different AD accounts per role?

If the ID’s are not logged during install, it’s a little more difficult to figure out what ID was used.

Domain Account for ALL services,
Enter in the unique DOMAIN\OMAA, DOMAIN\OMDAS, DOMAIN\OMREAD, DOMAIN\OMWRITE

Try these PowerShell commands to find what SCOM is using.

ON MS (from PowerShell (don’t need admin unless you’re restarting services)

$Services = ( Get-WmiObject -Class Win32_Service )

$Services | ? { $_.Name -eq “OMSDK” -OR $_.Name -eq “cshost” -OR $_.Name -eq “HealthService” } |

ft name,Startname,StartMode

ON SCOM DB’s, Reporting (from PowerShell (don’t need admin unless you’re restarting services)

$Services = ( Get-WmiObject -Class Win32_Service )

$Services | ? { $_.DisplayName -like “*SQL*” } | ft name,Startname,StartMode

Source https://blogs.technet.microsoft.com/heyscriptingguy/2012/02/15/the-scripting-wife-uses-powershell-to-find-service-accounts/

Azure UNIX server SCOM agent setup errors with OEL v7.x

Ran into some customers with UNIX agent problems, including Azure Oracle Enterprise Linux servers with SCOM agents.

Basically this error means

Fully-qualified domain name cannot be determined from the UNIX or Linux host itself
The FQDN known to the UNIX/Linux host does not match the FQDN used by the management server to reach the host

Full error message text

Agent verification failed. Error detail: The server certificate on the destination computer (agentname.contoso.net:1270) has the following errors:

The SSL certificate could not be checked for revocation. The server used to check for revocation might be unreachable.

The SSL certificate is signed by an unknown certificate authority.

It is possible that:

The destination certificate is signed by another certificate authority not trusted by the management server.
The destination has an invalid certificate, e.g., its common name (CN) does not match the fully qualified domain name (FQDN) used for the connection. The FQDN used for the connection is: agentname.contoso.net.
The servers in the resource pool have not been configured to trust certificates signed by other servers in the pool.

The server certificate on the destination computer (agentname.contoso.net:1270) has the following errors:

The SSL certificate could not be checked for revocation. The server used to check for revocation might be unreachable.

The SSL certificate is signed by an unknown certificate authority.

It is possible that:

The destination certificate is signed by another certificate authority not trusted by the management server.
The destination has an invalid certificate, e.g., its common name (CN) does not match the fully qualified domain name (FQDN) used for the connection. The FQDN used for the connection is: agentname.contoso.net.
The servers in the resource pool have not been configured to trust certificates signed by other servers in the pool.

Troubleshooting links

Old TechNet article for SCOM 2007R2

Docs site – link for 1801 – Steps haven’t changed, and IMHO, docs site is better documented

Here are some commands to help troubleshoot UNIX agent

ScxAdmin

Check UNIX Agent status

scxadmin -status

Example Output

$ scxadmin -status

scxcimserver: is running

scxcimprovagt: 2 instances running

Set Unix agent to START verbose logging

scxadmin -log-set all verbose

Restart Health Service & tail scx log

scxadmin -restart

cd /var/opt/microsoft/scx/log

tail -f scx.log

To correct a SCOM agent getting a SSL certificate error:

From the Docs site, the SCXsslConfig “tool is useful in correcting issues in which the fully-qualified domain name cannot be determined from the UNIX or Linux host itself, or the FQDN known to the UNIX/Linux host does not match the FQDN used by the management server to reach the host.”

As root:

1. Get the exact hostname of the server with the hostname command

2. Stop the SCOM agent – /opt/microsoft/scx/bin/tools/scxadmin -stop

3. Rebuild the cert – /opt/microsoft/scx/bin/tools/scxsslconfig -v -f -h HOSTNAME -d <FQDN_Here>

4. Start the SCOM agent – /opt/microsoft/scx/bin/tools/scxadmin -start

Additional Configuration topics from the docs site

Configuring SSL Ciphers link

Specifying an alternate Temporary Path for scripts link

Universal Linux – Operating System Name/Version link

Skype for Business 2015 (premise) Addendum MP

Ever try to figure out a Skype alert, for which server in the pool(s) is failing?

While maybe not the clearest to find root cause, the Skype pack brings a bunch of functionality, including synthetic transactions.

I was lucky enough to collaborate with Nick Wood, Skype PFE, to help provide more detail, troubleshooting, impact on what is critical versus warning.

What the addendum pack brings

Do you think 656 monitors can all be critical?

Sets up service restart recovery tasks for all Skype services

Company Knowledge tab for troubleshooting/user impact

Gallery Download

Here is a visual of our Skype efforts for integrating troubleshooting details into SCOM console.

NOTE: Company Knowledge tab would be accessible from the alert as well

Company Knowledge

SCOM Console, Authoring tab, Dispatcher Queue monitor

Highlight monitor, right click, choose properties

Click on ‘Company Knowledge’ tab

Incorporated the XLS into SCOM under Company Knowledge for additional information on user impact, causes, and troubleshooting (under resolutions)

Active Directory 2012-2016 Addendum packs updated

Man time flies!

Thought I’d share some new functionality for AD DS (Active Directory Domain Services)

Ran across some customer errors with AD Event ID 1084, which exists in the old 8321 pack, but not in the v10.x pack.

Well, if you get these errors, your DC isn’t replicating, and most likely will need to be rebuilt.

Gallery download

Broke out the packs to separate the Recovery Tasks in their own pack, versus added functionality in the addendum.

Figured better to send packs NOT sealed, so that meant 2 packs,

WYSIWYG (wizzy-wig acronym)

What this means

v1.0.0.1 pack had just the AD DS Service Recovery Tasks

v1.0.0.2 pack has a Service Recovery Tasks pack, and the Addendum pack

What I think is cool is that the Addendum pack contains 2 rules, simple rule event (enabled by default), and also a PowerShell rule.

Rule Figured out how to simply look for criteria, count it, and alert on it.

We always look for alert suppression, some of the sliding/counting monitors are too much.

Starting with Holman’s alerting rule fragment, we can create more powerful combinations than just a single symptom.

Using Variations of the get-date command, we can actually specify how far back to look, to count for alerts.

Easier method to count events, to figure out an alert threshold.

From the rule in the Addendum pack

# Check blog for more detail https://blogs.technet.microsoft.com/heyscriptingguy/2015/01/21/adding-and-subtracting-dates-with-powershell/
# If you want this in other time increments – AddHours, AddSeconds, AddMilliseconds
#
$LastCheck = (Get-Date).AddMinutes(-65)

[int]$TempCount = (get-eventlog -logName “Directory Service” -Source “NTDS Replication” -InstanceID 1084 -Message “*8451 The replication operation encountered a database error*” -After $LastCheck).Count

IF ($TempCount -ge 1)
{
$Result = “BAD”
$Message = “The number of 1084 Replication Database error events was greater than 1”
}
ELSE
{
$Result = “GOOD”
}