Darbio's questions

I have a HP ProCurve 1810G-8 which I currently use as a normal switch between 3 servers and a firewall.

2 of the servers are ESXi hosts, and one is a Nexentastor box with 2 iSCSI target LUNs.

As the iSCSI traffic is on the same LAN as all other traffic, I would like to switch this to use a SAN for iSCSI traffic and the LAN for all other traffic.

The Nexentastor box only has 2 NICs, and as such, with a physical arrangement, I presume that one must be plugged into the SAN VLAN and one on the LAN VLAN ports of the switch.

Is there a way to have multiple VLANs over the same port?

e.g. the Nexentsator box has 2 NICs, both plugged into the switch, both ports with access to both of the VLANs?

I have a HP ProCurve switch with 24 ports.

If I set up port 24 to mirror port 1 (which goes to my router), can I use port 24 for normal LAN access as well as to run something like ntop or bandwidthd?

Or does mirroring mean that the port can only be used to read the mirrored packets? Meaning a 2nd NIC is required for LAN access?

_{(source: spiceworks.com)}

Using Putty, I have set up a reverse proxy which allows me to connect on port 8080 of the server at server.tld to port 80 of the machine which initiated the SSH connection.

The server allows me to connect on localhost:8080 and returns the result of initiator:80 when I connect on the server.

This question says to enable GatewayPorts and bind to all addresses.

Using Putty, how can I expose port 8080 on the server so that when a request comes through on the external interface (e.g. a web request) the port is forwarded to the initiator:80?

I will preface this by saying that I am a solution architect, and I have a client who wishes to host an app in the cloud. The app needs to access information from SAP and other systems.

What is the best way to securely connect from a cloud machine to an erp system in a corporate network? I would like to do this without altering the firewall. Ideally I would like to have some kind of tunnel from the app to the SAP server.

I saw a company called RunMyProcess which has a software component which you install on a server behind a firewall, which somehow allows them to access your internal systems. I think they use Google Secure Data Connector.

How does this work?

We recently acquired a company running a very similar domain to ours, and we would like to merge the company into our domain.

The question is, can we provide a domain alias and map that to our domain (e.g. ComapnyA\username -> CompanyB\username) meaning we don't have to implement the new domain on the old machines.

Or do we have to maintain 2 domain forests?

The acquired company uses a username format of JoeB wheras we use JBloggs can we map the users to a specific user on the domain?

I have set up a distributed network as in the figure below. The subnet x.15.x is a remote subnet, connected to the internal LAN via IPsec VPN. The Firewall has 3 interfaces, one connected to the net, one connected to the x.2.x subnet and one connected to x.3.x subnet.

1. When I ping 192.168.3.25 from 192.168.2.10 I get a response.
2. When I try and ping 192.168.3.25 on it's external IP address from 192.168.15.50 I get a response.

However;

3. When I try and ping it from a client on the IPsec connection (e.g. 192.168.15.50) I get a request timed out.

Am I missing a principal of IPsec networking? How can I get it to send data to the x.3.x subnet?

The firewall is a pfSense, and the servers are Windows Server 2008 R2. The tunnel is a IPsec tunnel through the internet.

We currently run pfSense with no problems, however we are looking at TMG as it is included in our partner subscription to MS and allows Windows 7 DirectConnect features to our domain for off-site users.

I have had a google, but there don't seem to be any comparisons of TMG to pfSense.

Anyone have experience/knowledge of this?

Our infrastructure is Windows Server 2008 R2 behind pfSense at the moment.

We are a small consulting firm, with a need for a windows domain based network. Most of our consultants work either at home, or on-site.

What would be the best way to provide a domain centric infrastructure to our staff?

VPN would be used to connect to the infrastructure (hosted COLO), but this could be annoying if the user can't access the VPN due to on-site restrictions etc.

We are looking to move our infrastructure from our office to a COLO.

Currently we run a rack-mount white-box server using commodity hardware, and ESXi 4 as the hypervisor to power 9 VM's for internal development/DC/Exchange etc.

We are looking to use a SAN for storage, and have come up with a network diagram which allows us to use the spare ethernet port on the physical server to attach to another server - which is proposed to be used by the SAN.

The question is, is an ethernet port sufficient for this application? It is a gigabit ethernet port. I have used fibre in the past for this, but not ethernet.

These guys (http://www.datacore.com/) have a method of providing SAN over ethernet.

The proposed physical architecture is as follows:

With the virtual machines looking a bit like this (onbviously the connection between pfSense and eth1 would be removed if the top server was a SAN ):

Using gcc to make mono 2.8 from source on a CentOS 5 VPS with 128mb RAM, and it seems to be hanging. The result of free-m shows all memory has been used.

$ free -m
             total       used       free     shared    buffers     cached
Mem:           128        126          1          0          0          1
-/+ buffers/cache:        124          3
Swap:          127        127          0

What can I do to ensure that mono get's built? Will increasing the swap size help this?

Thanks.

Possible Duplicate:
Can you help me with my software licensing question?

I've got a product which I would like to start marketing as a SaaS product - I'm just a bit at a loss when it comes to licensing.

The product requires .NET web services, ASP.NET, and runs off of a SQL Server DB. I looked into using MySQL and Mono, but it all seems a bit complicated to migrate as the product is half way through production.

What licenses will I need to provide this to a customer as a service, hosted on my server in a data centre. I had a look at SPLA, but it is unclear whether this is the only way, or if I can buy a license of windows 2008 and SQL Server Enterprise, and use that to host the software.

I am a MS Partner for Development, so all's good that end - just doing the research before hanging myself out there!

Is it possible to virtualize windows server 2008 x64 on a server with VT not enabled? I know that using ESXi it is not.

If not, what are the problems associated with running an x64 MS Windows OS as the host, and then running guest machines as VM's off of that?

Current setup is 3 x86 VM's running on ESXi. I need to run Exchange 2010, so need an x64 OS to run it on. The motherboard is x64 capable.

Server = x346 IBM

JD

I have an x346 server with 3 x RAID 5 73GB U320 HDD's running ESXi with 2 VM's on it.

1 VM is an exchange server running IIS and SQL2008. The other is an applications VM running IIS.

I have another 3 x 36GB U320 HDD's and want to put these into the server as more storage, possibly on RAID 5.

What is the best way to add these drives? Is it possible to have 2 RAID 5 disk arrays? I've not had a play yet with putting the drives in.

As the 36GB drives will be used for storage which will be backed up off-site, which RAID option would be the best? I would like redundancy, but as this will be file storage ONLY (no DB, OS etc.) a backup may be sufficient.

Thanks,

JD

Is it possible to have a SAS to SATA.

e.g the server has SAS onboard, and I want to use SATA drives instead of the SAS ones (cheaper and easier to source here in Australia).

They have cables on this website -> http://www.techcable.com/HTML/SAS.htm

Will they work just by plugging in the SAS to the SATA using that cable?

The server is an IBM xSeries 345.

Thanks,

JD

Further to this the SAS is Ultra 320 scsi/sca2/lvd, 80 pin i think.