Skype for Business Server – Fabric Logs Consuming Disk Space (Script)

You may notice that your Skype for Business Front End server consumes a lot of disk space and the culprit will either be:

C:\ProgramData\Windows Fabric\Log (SfB 2015)
or
C:\ProgramData\Microsoft\SF\Log (SfB 2019)

This is caused by Windows Fabric log growth and can be rectified by either:

1) Increase disk space
2) Limit how long Fabric Logs are kept (default 3 days)

This post shows how to do option 2 manually in SfB 2015: https://myhybridsetup.org/skype-for-business-2015-limit-fabric-tracing-logs-growth/

The post above is still valid for SfB 2019, except the ‘ClusterManifest.Xml.Template’ is in a slightly different path: C:\Program Files\Skype for Business Server 2019\Server\Core (Quite obvious) and the Fabric Logs have moved to: C:\ProgramData\Microsoft\SF\Log

You can implement the same fix manually, or I have a script a colleague created to do this for you.

Disclaimer: Do not run this on a prod environment without first testing yourself.

$ver = Get-CsServerVersion

# Get Fabric Template Config files
$file = “C:\Program Files\Skype for Business Server $($ver.Substring(26,4))\Server\Core\ClusterManifest.Xml.Template"

# Copy and rename files incase of issues
Copy-Item -Path $file -Destination “$($file).old"

# Replace 3days with 1 day
(Get-Content $file)|foreach {$_ -Replace 'Name="DataDeletionAgeInDays" Value="3"','Name="DataDeletionAgeInDays" Value="1"'}|out-file $file -Force

# Delete logs SfB 2015
If (test-path -Path "C:\ProgramData\Windows Fabric\Log" -ErrorAction SilentlyContinue){Get-ChildItem "C:\ProgramData\Windows Fabric\Log"-Recurse |where {$_.name -like "*.etl" -or  $_.name -like "*.blg"}|Remove-Item -Force -ErrorAction SilentlyContinue}

# Delete logs SfB 2019
elseif (Test-Path -Path "C:\ProgramData\Microsoft\SF\Log" -ErrorAction SilentlyContinue){ Get-ChildItem "C:\ProgramData\Microsoft\SF\Log" -Recurse |where {$_.name -like "*.etl" -or  $_.name -like "*.blg"}|Remove-Item -Force -ErrorAction SilentlyContinue}

Write-Host "Restart Front End Services Manually" -ForegroundColor Yellow

The script is doing a couple of things:

• Get SfB Version to work out where ClusterManifest.xml.Template file & Fabric Logs may be (Assumes SfB has been installed on the C: drive!)
• Take a backup of ClusterManifest.xml.Template
• Perform a find and replace in ClusterManifest.xml.Template to set DataDeletionAgeInDays to 1 rather than 3
• Delete all SfB 2015 Fabric Logs
• Delete all SfB 2019 Fabric Logs
• Prompt you to restart Front End services

When we deploy SfB, we run this script on all servers on an FE pool to keep Fabric Logs/Traces to a minimum.

WARNING – It’s not clear if this is supported by Microsoft or whether future updates/CUs overwrite ClusterManifest.Xml.Template.

Chris Hayward

Technical Architect at Symity