Cylindric's questions

I am trying to set up an environment for testing a clustered SQL Server upgrade, so I have built a replica here on VMware Workstation 11.

I have provisioned 3 servers for this, two Windows Server 2012R2 for the SQL, and a FreeNAS 9.1 to provide the iSCSI shared volumes.

The iSCSI targets seem to work fine, the Failover Clustering setup accepts them for the cluster, and allows it's use for a Quorum volume.

The problem I have is that when I come to install a SQL Server instance onto the cluster, it gets as far as starting the databases up and then fails, with an error in the event log:

Master.mdf was originally formatted with sector size 4096 and is now on a volume with sector size 3072

That last number, 3072, varies on subsequent teardown/rebuilds, and I've seen it in the high 5-digits.

Any research on the error shows many results for issues with specific physical SAN drivers, or ancient results for products many years and versions behind.

Alternatively, is there a better way of bringing up a fully-virtualised SQL Cluster?

I'm trying to manage a Windows server with chef, and need to create some users, add them to a group, and add that group to another group.

user 'test.user' do
    password    'password'
    action      :create
end

group 'TestGroup' do
    action  :create
    members [ 'test.user' ]
    append  true
end

group 'OtherGroup' do          # This is line 43
    action  :create
    members [ 'TestGroup' ]
    append  true
end

That results in an error:

[2015-02-12T09:54:35+00:00] FATAL: ArgumentError: group[OtherGroup] (users::staff line 43) had an error: ArgumentError: A new member could not be added to a local group because the member has the wrong account type.

It's like Chef can't add a group to a group. Adding users to a group is fine.

Oddly it only seems to affect a custom group. I can add to the built in Users and Administrators groups fine.

group 'Administrators' do
    action  :modify
    members [ 'TestGroup' ]
    append  true
end

I have also tried creating the OtherGroup empty first, and then :modifying it, but the same problem happens.

I am trying to detect an eventlog entry on a server, so am trying to use the CheckEventLog command:

./check_nrpe -H hostname -c CheckEventLog -a MaxWarn=1 MaxCrit=1 "filter=generated gt -3d AND id=23 AND source='mpio'" truncate=800 unique "syntax=%severity%: %source%: %id%: %message% (%count%)"

But, I always get an error back from the client:

Exception processing request: Request command contained illegal metachars!

I have set the option to allow nasty meta chars, but that doesn't seem to help.

[/settings/NRPE/server]
allow arguments = true
allow_nasty_meta_chars = 1

If I take out the source part of the filter, it works, but of course event id's are not unique in themselves.

I am attempting to extract a certain date-range of emails from a mailbox into a PST, but using a DateTime parameter for ContentFilter doesn't seem to work as expected.

This should export everything for the last 5 days:

$endDate = Get-Date "00:00:00"
$startDate = $endDate.AddDays(-5)

Write-Host "Exporting items between $startDate and $endDate..."

New-MailboxExportRequest -ContentFilter {(Received -ge $startDate) -and (Received -lt $endDate)} -Mailbox "EmailLog" -FilePath "\\ReadyNAS\backup\Mailboxes\EmailLog\EmailLog.pst"

But the PST ends up with the full 15+Gb of the mailbox in it.

If I specify the dates manually, it works fine:

New-MailboxExportRequest -ContentFilter {(Received -ge "01-06-2013 00:00:00") -and (Received -lt "06-06-2013 00:00:00")} -Mailbox "EmailLog" -FilePath "\\ReadyNAS\backup\Mailboxes\EmailLog\EmailLog-man.pst"

I'm not sure if this is due to some sort of regional-settings conflict, but I would have thought that passing a typed parameter would avoid that sort of sillyness. The reason I suspect this is because if I print out the dates (the Write-Host in the first example) the dates come out in US format:

[PS] C:\Windows\system32>Write-Host "Exporting items between $startDate and $endDate..."
Exporting items between 06/01/2013 00:00:00 and 06/06/2013 00:00:00...

We're seeing some suspicious network activity, and when I was trying to see if it was one particular server of ours I ran a Wireshark trace. I noted a lot of ARP packets asking who has x.x.x.x, but all being told to tell different addresses. In the past I've only seen the "tell" to be a single host - for example a DHCP server.

As you can see from the screenshot, there are only a few IP's being asked for, but the system to tell varies a lot. It's like all devices on the network are trying to find out who 10.10.0.40 (and a couple of others) is.

We have a mailbox that receives status emails that we need to keep essentially for ever, and they contain potentially large attachments. I need to be able to move these messages to a PST on a regular (daily or weekly) basis so want to automate this.

At the moment I am using a MailboxExportRequest to get them into a PST:

New-MailboxExportRequest -ContentFilter {(Received -ge '01/01/2013') -and (Received -lt '01/02/2013')} -Mailbox "MonitorMailbox" -FilePath "\\Server\backup\EmailLog\MonitorMailbox 2013-01.pst"

And I follow this with a SearchMailbox -DeleteContent:

Get-Mailbox -Identity "MonitorMailbox" | Search-Mailbox -SearchQuery "Received:01/01/2013..01/02/2013" -DeleteContent

The problem is that I'm having to manually update the date parameters each time in four places, and I'm worried that it's not very transactional (if that even makes sense in this context).

I'm trying to create a little PS script to iterate through dates by month and build a set of commands to spin through and export this information, but was wondering if there's a simpler way to "move to PST", or if I need to try and build the above statements dynamically.

Here's a more complete script, as I have it so far, which should export all items for the 10 days preceding the beginning of this month (to speed up testing):

$mailbox = "Cylindric"
$endDate = Get-Date -Day 1 "00:00:00"
$startDate = $endDate.AddDays(-10)

$month = "{0:D2}" -f [int]$startDate.Month
$year = "{0:D4}" -f [int]$startDate.Year


Write-Host -NoNewline "Exporting items between $startDate and $endDate..."
New-MailboxExportRequest -Name "EmailLog$year$month" -ContentFilter {(Received -ge $startDate) -and (Received -lt $endDate)} -Mailbox $mailbox -FilePath "\\ReadyNAS\backup\Mailboxes\EmailLog\EmailLog${year}-${month}.pst"
Write-Host "Done."


Write-Host -NoNewline "Waiting for export to complete..."
While(!(Get-MailboxExportRequest -Mailbox $mailbox -Name "EmailLog$year$month" -Status Completed))
{
    #Sleep for a  few minutes
    Write-Host -NoNewline "."
    Start-Sleep -s 60
}
Write-Host "Done."


Write-Host -NoNewline "Removing Export Request..."
Get-MailboxExportRequest -Mailbox $mailbox -Status Completed -Name "EmailLog$year$month" | Remove-MailboxExportRequest
Write-Host "Done."

And the output:

Exporting items between 05/22/2013 00:00:00 and 06/01/2013 00:00:00...
Name            Mailbox       Status
----            -------       ------
EmailLog201305  Cylindric     Queued
Done.
Waiting for export to complete..........Done.

This seems to be exporting a lot more than these 10 days though. The PST appears to contain ALL emails!

I have a Windows Server 2012 (non-domain) with a folder that's shared out as TestShare.

The share permissions are Everyone has full access, and there is a local user TestUser that has full access to the actual folder.

On GuestServer I can connect and/or map a drive to \HostServer\TestShare, specifying the username and password for TestUser.

NTFS permissions:

Share permissions

Effective Access Report

The problem is that when I try to access the folder, I get an "access denied" message.

On the host server I can see the user connected to the share in the Sessions manager, so the password is correct and being recognised. If I use an incorrect password I don't get the "completed successfully" message, nor the 'open session'.

What else can be blocking access to the shared files, when the share seems to be set, and the folder permissions seem to be set, and the connection seems to be okay?

The network is recognised as "public", and the relevant firewall rules seem to be enabled - even disabling the firewall doesn't help.

We have an Exchange infrastructure going through a migration from 2003 SP2 (call it ExOld) to 2010 (ExNew).

All users are now on the new server, but mail is still being directed to ExOld until testing is complete. ExNew sends emails directly to the internet.

For one particular user, emails don't seem to be being reliably delivered, but the odd thing is that it's not all emails. I can see external emails in his inbox.

• If I send an internal email it works fine.
• If I send an email from Gmail to him it doesn't get through.
• If I telnet from outside to ExOld I can send an email to him.
• If I telnet from outside to ExNew I can send an email to him.

This is a transcript that results in a successful send:

    220 ExOldName Microsoft ESMTP MAIL Service, Version: 6.0.3790.4675 ready at  Mon, 22 Oct 2012 10:55:26 +0100
    EHLO test.com
    500 5.3.3 Unrecognized command
    EHLO test.com
    250-ExOldFQDN Hello [MyTestExternalIp]
    250-TURN
    250-SIZE
    250-ETRN
    250-PIPELINING
    250-DSN
    250-ENHANCEDSTATUSCODES
    250-8bitmime
    250-BINARYMIME
    250-CHUNKING
    250-VRFY
    250-X-EXPS GSSAPI NTLM LOGIN
    250-X-EXPS=LOGIN
    250-AUTH GSSAPI NTLM LOGIN
    250-AUTH=LOGIN
    250-X-LINK2STATE
    250-XEXCH50
    250 OK
    MAIL FROM:[email protected]
    250 2.1.0 [email protected] OK
    RCPT TO:[email protected] notify=success,failure
    250 2.1.5 [email protected] 
    DATA
    354 Start mail input; end with .
    Subject:Test 1056
    Test 10:56
    .
    250 2.6.0  Queued mail for delivery
    quit
    221 2.0.0 ExOldFQDN Service closing transmission channel

Emails go through Symantec Cloud, but their "Track and Trace" shows the messages going through, with a "delivered ok" log entry.

2012-10-22 09:19:56 Connection from: 209.85.212.171 (mail-wi0-f171.google.com)
2012-10-22 09:19:56 Sending server HELO string:mail-wi0-f171.google.com
2012-10-22 09:19:56 Message id:CAE5-_4hzGpY2kXFbzxu7gzEUSj5BAvi+BB5q1Gjb6UUOXOWT3g@mail.gmail.com
2012-10-22 09:19:56 Message reference: 135089759500000177171130001194006
2012-10-22 09:19:56 Sender: [email protected]
2012-10-22 09:19:56 Recipient: [email protected]
2012-10-22 09:20:26 SMTP Status: OK
2012-10-22 09:19:56 Delivery attempt #1 (final)
2012-10-22 09:19:56 Recipient server: ExOldIP (ExOldIP)
2012-10-22 09:19:56 Response: 250 2.6.0 Queued mail for delivery

I've tried to search the Message Tracking Center on ExOld, but I just get the following message:

The tracking database on ExOld.FQQN is not available or the message has left the Exchange organisation. Facility: Microsoft Exchange Management ID no: c1032751 Microsoft Exchange Management

I'm not sure where else to look on the old (or new) server for information as to where the mails are ending up.

I'm trying out Windows Server 2012, specifically Windows Server Backup, as it would be a very useful way to see us through until we can upgrade our 'real' backup system.

I'm backing up to a network share, and a Server 2008 WSB works fine.

On the 2012 server however, I get an error:

Backup of volume \?\Volume{298d1a7d-f80f-11e1-93e8-806e6f6e6963}\ has failed. The version does not support this version of the file format.

It's a VHDX written by WSB, so I'm not sure what version it's complaining about.

I can see a whole bunch of files in the destination, so I don't think it's a simple authentication issue, but I only get about 8Mb of VHDX written.

I've been using MDT for a while to deploy my Win7 desktops, mostly using WDS and boot-from-LAN.

The problem I have now is that the company recently took it upon itself to purchase laptops without wired NICs, so I can't use PXE boot.

Can I use the LiteTouch ISO boot image and connect to a WiFi network to pull the data from the Deployment Share?

I am trying to get Windows Search to work on my Windows Server 2008 SP2 fileserver, so I can search in files for content.

I have added the Windows Search Service role to the server, and using the right-click properties in Explorer set some folders to "Index this location".

The problem is that neither on the server or remotely can I search in the files.

I seem to get some inconsistencies in the GUIs, for example the "Indexing Options" panel shows me just 6 locations indexed, but if I click "Modify" I see nearly everything ticked.

For example, the "SeachTest" folder under "infrastructure" has the "index this location" option ticked, but the "Projects" folder does not. I assume this is why some are grey and some not, but they are all ticked. T The "SearchTest" folder contains some files that have nothing but the text PurpleOrange in it, so I should be able to find those.

So, to summarise:

• Which locations are indexed? The ones in the "Index these locations" list, the ones ticked, or the ones not greyed-out in the list?
• How do I get to the state where I can click in the search box and type PurpleOrange and see the files?

I am trying to install an application (SalesLogix 7.2 Admin Tools and Servers) from an MSI package, but only want to install a particular feature. How can I determine the feature name so I can try using the ADDLOCAL parameter?

I have a WDS server setup for use with MDT2010, and everything was working fine until this morning. Now, whenever I try to boot from LAN, I get an error:

Downloaded WDSNBP...

Architecture: x64
WDSNBP started using DHCP Referral.
Contacting Server: 10.50.10.12 (Gateway: 0.0.0.0)
No response from Windows Deployment Services server.
Launching pxeboot.com

My PXE Response Policy setting on the WDS is set to this:

[ ] Do not respond to any client computer
[ ] Respond only to known client computers
[o] Respond to all (known and unknown) client computers
    [X] For unknown clients, notify administrator and respond after approval

The odd thing is that if I clear the approval option (so any computer gets a response) it works fine.

I have delegated permissions on the AD OU to the computer object WDS is running on, but that doesn't seem to have helped.

As it works without approval, I can only assume my DHCP options are fine and this is some sort of AD permission problem.

(Server is Windows Server 2008 SP2)

I am not sure if I am confused as to the purpose of MDT.

We need to deploy a standard Windows 7 build to about 100 new computers, and more as time goes on.

Having followed a whole bunch of documentation, tutorials and white-papers etc, it seems that MDT will always go through the full processes of

• installing windows
• installing applications

Whereas what we need is a simple "take image of reference machine" followed by "clone image onto target". This should be much faster than the full installation routines followed by MDT.

Should I be doing something else with the captured WIM from MDT? Can I simply put it into WDS somehow as a "ready to copy" image?

We've just started using the Microsoft Deployment Workbench and Windows Deployment Services to roll out a bunch of desktops.

The basics seem to be going well, but I have a question about the first GUI option, which is the "Microsoft Solution Accelerator" screen where you have the option of "Run Deployment Wizard" and "Command Prompt".

There is a drop-down box there for keyboard language, which seems to only be used for authentication to the Deployment Share, but I can't work out how to change this default to UK.

My current Bootstrap.ini on the Deployment Share properties is this:

[Settings]
Priority=Default

[Default]
DeployRoot=\\HELPDESK\DeploymentShare$
InputLocale=0809:00000809
KeyboardLocale=0809:00000809
UserLocale=en-GB
UILanguage=en-GB

We recently had a problem with our Exchange 2010 backups (Backup Exec 2010), so we now only have access to a full VMware guest image of the Exchange Mailbox Server.

I have restored the VM fine, isolated from the network in it's own test network, as the original source machine is still running and needs to keep running.

What is the best way of getting a mailbox (or recovery storage group or whatever) off this 'isolated' server back onto the live server? Is it even possible?

At the moment, I cannot start the Information Store service, as there is no access to the Domain in an isolated network, so I get various event log entries such as:

5000: Unable to initialize the Microsoft Exchange Information Store service. - Error 0xfaf. 1121: Error 0xfaf connecting to Active Directory 2104: Process STORE.EXE. Topology discovery failed due to LDAP_SERVER_DOWN

Is it possible to start Exchange in some sort of recovery mode where it doesn't need access to the AD? Or do I need to also clone a DC into the isolated network?

We are running Exchange 2010 on a VMware infrastructure, and backing up with Backup Exec 2010 using ESX licenses.

I was led to believe that even though we are backing up entire VM guest images, Backup Exec would allow granular recovery of individual mailbox items, from this VM image.

Otherwise, we have to backup the VMDK, which gets all the mailbox data files, and do a traditional backup of the Information Store, essentially duplicating the amount of data backed up.

So the question really is, will I be able to restore an individual mailbox (or even message) from an ESX guest backup?

We have a situation where our backups sometimes don't quite fit on one tape, and there's no budget for a bigger drive.

What we do have is two separate tape drives, so we've hooked them both up to the backup server.

My question is, how should I configure Backup Exec (12.5) correctly to use these drives?

Here's what I've tried

• Drive 1 called HP1
• Drive 2 called Dell1
• A device pool called TapeDrives containing Dell1 and HP1
• Policies have a Duplication task pointing to TapeDrives as their device

The problem I have is that the tape is HP1 fills up, and the job moves onto Dell1, but subsequent jobs try to load HP1, so stall.

Should the "Preferred source device" in my jobs be pointing at the drive I want used first (HP1 by preference), or the pool?

The verification jobs also now fail, because HP1 was ejected due to being full!

How do I configure two devices to be used in BE, so that one device is used by preference, then the second if required?

We're using SQL Server 2008 for a (third-party) web app, and they utilise SSRS for reporting.

Recently, so I'm assuming this is related in some way to an MS patch, whenever we click on the "print" icon in a report we get an error popup:

Message from webpage
Unable to load client print control.

• The web server is Windows Server 2003 R2 x32
• The SSRS server is Windows SQL Server 2008 10.0.2531.0
• The SQL server is Microsoft SQL Server 2008 (SP1) - 10.0.2531.0 (X64)
• The clients are Windows 7 and Windows XP

All the servers are up-to-date according to Windows Update.

The SSRS server also has the "Security Update for Report Viewer Redistributable 2008 Service Pack 1 (KB971119) installed.

We have two Cisco 2960 switches that we need to be able to connect together.

We have four links joining SwitchA to SwitchB, I'd like to know the best way of configuring the switches so these four links are used as a trunk between them.

I guess the avilable options are Etherchannel (is this even available on the 2960?) or some other sort of link aggregation.

Many thanks,

Mark