Max's questions

For analytics purposes, I'm looking at large sets of IP addresses in server log files. I'm trying to perform reverse-DNS lookups to understand where traffic is coming from - e.g. what percentage of IPs resolve to corporations, schools, government, international etc.

Despite a bunch of optimizations, individually reverse-DNS'ing every IP address still appears to be fairly expensive though. So -

is there any way to obtain an entire range of IPs from a reverse-DNS?

If yes, this could greatly reduce the number of actual reverse-DNS lookups.

Example (numbers slightly obfuscated):

• Log file contains a request from an IP 128.151.162.17
• Reverse DNS resolves to 11.142.152.128.in-addr.arpa 21599 IN PTR alamo.ceas.rochester.edu
• (So this is a visitor from Rochester University, rochester.edu)
• Now, would it be safe to assume that all at least all IPs from 128.151.162.* will also resolve to rochester.edu?
• What about 128.151.*.*? Is there a way to get the exact IP range?

Lets' say my IIS8 hosted web site is up and running and happily serving requests. Now, from an elevated Powershell, I run:

Stop-WebSite <MyWebSite>

My assumption until recently was that IIS would stop serving new requests, finish up current ones, and then gracefully stop the web site. However, I suspect that this is not always the case: there's a code path in my app like this:

Log("Begin");
try
{
  DoSomething();  // can take a few seconds
}
finally
{
 Log("End");
}

Occasionally, there would be "Begin" entries in my log without matching "End". Any chance Stop-WebSite forcefully kills the host process?

Is there a way to enable Powershell remoting (Enable-PSRemoting) during an unattended Windows Server 2008 R2 installation? Is there a setting in unattend.xml for this?

After upgrading from SQL2005 to 2008 SP1, we're suddenly observing disproportional expansions of the database log file during large transactions.

Example: inserting 75 million rows into table B, deleting those same 75M rows from table A, both in the same transaction. Total amount of data about 2 times 20GB, but database log file grows to 150GB and beyond.

Never seen this in SQL2005, before upgrading.

I couldn't find this documented as a known regression in SQL2008 anywhere.. anyone seen this, too?

Thanks

Max

BCDEdit allows for editing the bcd store of the current system drive (C: drive). Is there also a way to create and edit the bcd store of a secondary drive, so the system can boot from it later on? (For instance, by physically moving the drive to another machine as the primary disk)

I have a domain-joined test machine running W2K8 R2 with two NICs, one connected to the enterprise network, the other one to a private network with a static IP (192.168.0.1). The machine should act as a DHCP server for the other machines on the private network (not for machines on the enterprise network, though).

After adding the DHCP role, I checked that the service only has a binding to the LAN connection on the private network.

The domain controller won't allow to authorize another DHCP server, and the DHCP service throws an error when starting up (event ID 1046): http://technet.microsoft.com/en-us/library/cc726914(WS.10).aspx

My question is: is there a way to tell the DHCP service that it'll only serve a private network and therefore doesn't have to ask the DC for authorization?

Thanks Max

A Windows service running as "Network Service" talks to services on other machines (here: SQL Server and Analysis Services), using Windows authentication. For authentication, we have to grant permissions to the machine account of the service. E.g. if service runs on server MYSERVER in domain MYDOMAIN, it'll authenticate itself as "MYDOMAIN\MYSERVER$". - Am I correct, so far?

Now here's my question: does this still apply when talking to a service on the SAME machine? Or will it authenticate with something like "NT AUTHORITY\Network Service" instead when connecting to a local service?

And: is there any chance this is a breaking change from Windows 2003 to Windows 2008? We're having an actual issue in our system where the account was able to connect to local services with only the machine account having permissions in W2K3. In W2K8, this doesn't seem to work anymore: authentication to local services now fails, but still works to remote machines.

We have some servers with Windows 2003 and SQL Server 2005 and would like to upgrade them to Windows 2008 and SQL2008 because of the new cool features.

How can that be done? Anything in particular to think about?

When a new Windows server machine joins a domain, AD seems to create a machine account "DOMAIN\MACHINENAME$" for that machine with a SID.

If the machine gets reimaged (with another OS, here: W2K8 instead of W2K3) and then rejoins the domin, will AD re-use the existing domain account with the same SID?

(Reason I'm asking is that we use some machine accounts as logins in SQL2008 databases..)

Thanks Max

Is there a Powershell cmdlet or script to query Active Directory if a given domain account (such as "myDomain\myUser") exists?

Windows HPC Server 2008 provides the option to automatically deploy OS images to new cluster node, using Windows Deployment Services. However, this requires the HPC cluster to be set up with a "private network" network topology.

From HPC Cluster Manager Help: "Network Topology 'All nodes only on an enterprise network': Because all nodes are connected only to the enterprise network, you cannot use Windows Deployment Services to deploy compute node images using the new deployment tools in Windows HPC Server 2008."

When selecting "Compute nodes isolated on a private network" as network topology, HPC network configuration wizard calls out that "2 online network adapters are needed for this topology". No problem for a cluster of physical machines.

My question is: if the compute cluster consists of virtual machines (running on multiple host machines), how can I set up a "virtual network" for all VMs? And, is there a way to set up some kind of virtual network switch on the head node which HPC accepts as a second NIC?

Btw. this is for a test environment, the production cluster consists of physical machines.