I have a Server 2008R2 guest running inside a KVM host. I've enabled various Hyper-V enlightenments through KVM (-cpu host,hv_relaxed,hv_vapic,hv_spinlocks=0x1000), but I'm not sure how I can confirm that the guest is actually making use of them.
How can I list what enlightenments the guest is actually making use of?
I have an environment where there is one DHCP server servicing a number of different hosts/vlans. The switches are configured to forward the DHCP requests over (via ip-helper) and include information about the port (option 82). I'd like to take that information and translate it into an actual lease for the server.
I don't think it's particularly feasible for me to pregenerate a list of available leases, but I should be able to determine an address for a lease as it comes in.
Is there an DHCP server that can execute a script when it receives a request? (Note: I'm looking to assign the IP from the script, not have the DHCP server assign an IP then execute the script)
Edit: So, ultimately I'm trying to provide DHCP/PXE services over a large number of distinct vlans. This is so we can do OS installs via PXE booting without having to have a separate PXE vlan. I've got the switch config down no problem, and I have the DHCP server recognizing option 82.
I need a way to pull DHCP assignments from another system (this other system would know what subnet to use on what vlan), but I do not want to have to pregenerate a list of vlan:DHCP range pairs.
Let's say that I have a machine on my domain that has been compromised by an unknown attacker. If I use my domain credentials to log into that machine remotely, via RDP, does that expose my credentials to the attacker?
I know AD uses kerberos internally, but I'm unclear if the RDP client gets the kerberos ticket directly from the domain controller, or if it sends the login info to the server, which then gets the ticket.
To clarify, the computer running mstsc is Windows 7, and the possibly compromised machine is 2008 R2.
I'm using active directory to manage logins to a large number of servers. We have various admins that will be connecting via RDP to these servers. Is there a way I can force windows to remove their local profiles after they log out?
I found the 'Delete user profiles older then a specified number of days on system restart' and 'Delete cached copies of roaming profiles', however the former doesn't help a whole lot (these servers generally only get restarted for windows updates), and the latter doesn't seem to apply to local profiles.
I might be able to accomplish this with a script that runs on logout, however I'd rather not do that unless absolutely necessary.
I have a number of machines running Win2k3 and 2k8R2 that I would like to join to a domain. All the machines have been imaged from the same image, without using sysprep or anything similar. This means they all have the same SID. Am I going to run into any issues if I try to join all of these to the same domain?
I've already tried this with a couple test machines, and I experienced issues when the domain controllers came from the same image as all the other computers (I was seeing error log messages complaining that the domain controller SID was the same as the machine SID).
I did another test with freshly installed domain controllers (so they had different SIDs), and everything seemed okay.
I've done some research on this, and found some conflicting information. I can find a lot of references to this post, which seems to indicate that the duplicate SIDs are okay. I also find a lot of people saying horrible unspeakable things will happen unless I correct the SIDs.
Are these duplicate SIDs going to cause a problem?
Before anyone suggests any of these:
How would I go about replacing a TTY with a custom script in CentOS 6. CentOS 6 uses upstart instead of the standard SysV init scripts, so I'm having difficulty figuring out how it all works.
Typically (in SysV init scripts), I would edit /etc/inittab and replace one of the mingetty lines with my script, like this:
1:2345:respawn:/root/myscript tty1
However, with Upstart, this file doesn't the relevant lines, so I'm unsure of how to accomplish this.
Is there any standard way to prevent shared web servers from being abused? I run a CPanel box with a few people on it, and I get the occasional person that decides to use it to DoS other websites. My current 'detection' involves looking at Munin graphs for traffic spikes, then poking around on the machine until I find the cause.
Is there any software out there to detect attacks as they happen?
I run the following: LMD - http://www.rfxn.com/projects/linux-malware-detect/ Clanscan
These do an okay job of detecting your standard C99 shell, or commonly used DoS scripts, but it doesn't detect any sort of custom stuff.
How would I go about retrieving 64 bit counters for network traffic on windows (Server 2008 specifically)? I'm seeing too many rollovers on the 32 bit counters for our polling interval, leading to gaps in the data.
We feed the data through RRDTool, which can successfully handle one rollover per polling period, but can't deal with two. 64bit counters would prevent this issue, but I can't seem to find a way to enable them on windows