There is no automatic id assignment per record so it is not possible to sort by id and select top from desc.

A workaround is needed as the Top to be expected are last inserted records.

For EDB this means using TryMoveLast and TryMovePrevious

For SQL and MDB another workaround is needed.

When moving data from mdb to edb or sql to edb, the pull server will not see the inserted records.

moving data out of edb to mdb or sql will make it available to the pull server.

Investigate what needs to be done before the pull server will see the records. One thing I see is tat inserts by the pull server code make use of update.SaveAndGotoBookmark(); which is not implemented in the current dscpullserveradmin code for set-eserecord

Hi,
it would be nice if Paramter of ESEPath has a default value like
"C:\Program Files\WindowsPowerShell\DscService\Devices.edb"

Thanks

When I try to pass a credential to New-DSCPullServerAdminConnection, I get the following error:

PS C:> New-DSCPullServerAdminConnection -SQLServer SERVERNAME-Database DSC -Credential (Get-Credential)

cmdlet Get-Credential at command pipeline position 1
Supply values for the following parameters:
Credential
Invoke-DSCPullServerSQLCommand : Exception calling "Open" with "0" argument(s): "Login failed for user 'CORP\mcelreathm'."
At C:\Users\mcelreathm\Documents\WindowsPowerShell\Modules\DSCPullServerAdmin\0.1.5\DSCPullServerAdmin.psm1:1081 char:19

• ... estResult = Invoke-DSCPullServerSQLCommand -Connection $testConnectio ...

•             ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  

  • CategoryInfo : NotSpecified: (:) [Write-Error], MethodInvocationException
  • FullyQualifiedErrorId : SqlException,Invoke-DSCPullServerSQLCommand

I'm using an account that is a sysadmin on the SQL server which is also the account I'm logged in as. If I omit the -Credential parameter, it works fine.

PS C:> $PSVersionTable

Name Value

PSVersion 5.1.15063.1292
PSEdition Desktop
PSCompatibleVersions {1.0, 2.0, 3.0, 4.0...}
BuildVersion 10.0.15063.1292
CLRVersion 4.0.30319.42000
WSManStackVersion 3.0
PSRemotingProtocolVersion 2.3
SerializationVersion 1.1.0.1

Add filtering options to Get-DSCPullServerAdminStatusReport like:

• Newest (default to 1)
• Oldest (default to 1)
• Status (success, failed, ??)
• OperationType (LocalConfigurationManager, Consistency, Initial, ??)
• ...

Running
Get-DSCPullServerAdminReport -ESEPath 'C:\Program Files\WindowsPowerShell\DscService\Devices.edb' | select -Last 1
once works. When i run the same code a second time i get the following message:
`Get-DSCPullServerAdminReport : Instance Name already in use
At line:1 char:1

• Get-DSCPullServerAdminReport -ESEPath 'C:\Program Files\WindowsPowerS ...

•   + CategoryInfo          : NotSpecified: (:) [Get-DSCPullServerAdminReport], EsentInstanceNameInUseException
    + FullyQualifiedErrorId : Microsoft.Isam.Esent.Interop.EsentInstanceNameInUseException,DSCPullServerAdmin.src.CmdLets.GetDSCPullServerAdminReport`
  
  

This only happens if you start the cmdlet and break it while it is running (CTRL+C)

Env: Windows Server 2016
PS: 5.1.14393.693

Hi,
first thanks for this great project.
Since the database Devices.edb is growing and growing it would be nice if we could clean up StatusReports table. A switch how much days back we would like to have would be nice.

Thanks

The Module already is capable of running xPlat when SQL databases are used (MDB and EDB are not supported on non-Windows).

However, the use of RequiredAssemblies for Microsoft.Isam.Esent.Interop in the module manifest makes it impossible to import on Linux / Mac.

To fix this, the module needs another way of importing this assembly (Windows only) and make it clear to a non-Windows user that MDB and EDB are not available

Encountered a case where TryMoveNext resulted in False. TryMoveFirst resulted in true and after this is done, the records can be processed.

Need to adjust logic to compensate for this

I know I mentioned this on Slack but I thought it would probably be best to raise it here instead.

I am using version 0.4.2 on a Server 2012 R2 pull server and have stopped IIS with iisreset /stop before making a copy of the Devices.edb file to a temporary location. Then against this copy of the DB I run the following commands and get the error as shown below.

Import-Module DSCPullServerAdmin
New-DSCPullServerAdminConnection -ESEFilePath '.\TempDevices.edb'
Get-DSCPullServerAdminStatusReport

Get-DSCPullServerESERecord : Cannot open any more tables (cleanup already attempted)
At line:2092 char:17
+                 Get-DSCPullServerESERecord @eseParams 
+                 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ CategoryInfo          : OperationStopped: (:) [Write-Error], EsentTooManyOpenTablesException
+ FullyQualifiedErrorId : Microsoft.Isam.Esent.Interop.EsentTooManyOpenTablesException,Get-DSCPullServerESERecord 

Have you any idea why I am running into this?

As I mentioned today on Slack I had no issues when using the Copy-DSCPullServerAdminData command to create a fresh DB and import all my node registrations across. This temp DB is a copy of the fresh DB so there shouldn't be any old StatusReports hanging around only ones sent since this morning.

Cheers.

To introduce MDB support we need access to System.Data.OleDb.OleDbConnection which is afaik only available in 64 bit when Microsoft Access Database Engine is installed. When this is installed, PowerShell 5.1 is needed to access it.

• First create MDB support
• Research if PowerShell 6 can be supported (the rest of the module work with 6 on Windows). UPDATE Test on type available. If type is not available write error suggesting to use PSv5.1. If PSv6.x based on newer netcore does have the type in the future, the Error won't be written.
• Research if a library can be shipped as part of the module to prevent additional installation requirements. UPDATE: Won't do this. Instead write error with download requirements
• Like behaves different on OleDb. It requires a wildcard to be appended in some cases or it won't work. In other cases, it just doesn't work. Research how to solve this. (Note, behavior is Server Core, Windows 10 works as expected). UPDATE: Decided to exclude wildcard support for MDB connections
• Add tests

Now we have SQL support in Server 1803 (SAC) / 2019 (LTSC), it would be nice if we could get the content out of the EDB and insert it directly in the SQL Database to preserve existing node registrations and reporting data.

Right now my thinking is to have the ability to either create a new SQL DB from scratch or be able to connect up with an existing one.

• Cmdlet/Function to create a new DB
• Cmdlet/Function to connect with existing DB (created by either the Pull Server itself or by using the previously mentioned Cmdlet/Function)
• Cmdlet/Function to convert EDB data to SQL INSERT Statements and execute on DB if not exists

• RegistrationData
  • New
  • Set
  • Remove
• Devices
  • Fix Get BUG
  • New
  • Set
  • Remove
• Reports
  • New
  • Set
  • Remove
• Copy to EDB from *
• Fixup parameter sets
• Update Help
• Update tests
• Reduce code for Get to one function
• Add wrapper functions for ESENT tasks to allow for mocking
• Verify proper EDB when setting up connection

This can cause errors which are terminating when happening in ese mode

• Investigate why this happens
• Handle these errors more gracefully
• Determine if it is caused by moving data into edb by SerializeObjectToColumn (maybe needs additional ConvertTo-Json before Serialize will happen, unsure)

When releasing to master, include the GitHub release (tag with version same as published on the gallery)

Right now, release in GitHub is manual and will trigger a new build, need to filter this as it will publish a new version to the gallery.

Currently there are a lot of tests that are basically the same but just targeting a different database provider.
This results in a lot of duplication in the unit test files that could potentially be resolved by moving stuff to testcases