SvrGuy's questions

We are trying to implement some data logging to CSV files using a Data Collector Set in PerfMon (on a windows Server 2008R2 system).

The issue we are running into is that we (seemingly) can't control the permissions being set on the log files created by perfmon.

What we want is for the log files created by perfmon to have Everyone:F permissions (Full Control for Everyone). So, we have a directory structure setup where all logs go into a folder:

c:\vms\PerfMonLogs\%MACHINENAME% (e.g. c:\vms\PerfMonLogs\EvaluationG2)

In the above example, c:\vms\PerfMonLogs\EvaluationG2 has permissions Everyone:F (below is the icacls for this directory)

EVALUATIONG2/ Everyone:(OI)(CI)(F) NT AUTHORITY\SYSTEM:(OI)(CI)(F) BUILTIN\Administrators:(OI)(CI)(F) BUILTIN\Performance Log Users:(OI)(R)

When the data collector set runs, it creates new sub folders and files within c:\vms\PerfMonLogs\EvaluationG2, e.g. (C:\vms\PerfMonLogs\EVALUATIONG2\M11d26y2012N3)

Each of these directories and files has the following permissions:

M11d26y2012N3 NT AUTHORITY\SYSTEM:(OI)(CI)(F) BUILTIN\Administrators:(OI)(CI)(F) BUILTIN\Performance Log Users:(OI)(R)

So these new folders and not simply inheriting permissions from the parent folder (don't know why).

Now, we tried adding Everyone:F using the security tab on the collector set (No dice).

Any ideas? How do we control the permissions on the log files generated by perfmon data collector set?

We have some windows servers (2008 R2) and we are experiencing a strange network issue on them. The best way that I can describe it is that we are experiencing low throughput (but high bandwidth) only when operating over internet latency (20ms or more) networks.

Symptoms:

1.) The servers in question are located in two tier 1 datacenters with great, wide internet pipes.

2.) If we RDP onto servers in Data Center #1, we can run SpeedTest.NET test and get hundreds of meg down/up

3.) If we download a CentOS ISO image (basically a big file) on a server, we can get between 6 - 20 Megabytes/sec bandwidth

4.) If we connect to a web server we run on a FreeBSD 9.0 server (also in the data center) from another datacenter (so, freebsd in dc #1, windows server in dc #2) and download a big file (100s of MB) we also get great performance. (again, many megabytes/sec)

5.) If we connect from a server in DC #1 to another windows server in DC #1 we get awesome performance (nearly 1gb/sec).

6.) If we connect from a windows server box in DC#1 to a windows server box in DC#2 we get ~450 kilobytes/sec (really poor throughput). but we can open multiple browsers/windows and download as many times as we want (simultaneously) and still get exactly the same 450 kb/sec.

7. We ran NTTTCP and a windows port of netperf between windows servers in DC#1 and DC#2 and, in every case we can find, we get the same throughput (~450 kb/sec) on each download.

8.) We ran a whole bunch of other tests that allow us to conclusively eliminate firewalls etc. as the source of our issue (also, see test 4, as the FreeBSD server is communicating though the same [hardware] stack as the windows servers...)

We tried disabling the TCP window scaling on a pair of windows servers to no effect.

So ServerFault community: it seems like it must be a windows TCP setting of some sort -- what causes this sort of issue?

EDIT

1.) We installed VMWare/CentOS on one of the windows server 2008R2 boxes having the issue, with bridged networking, guess what great throughput! So, VMWare image, on windows host, greate performance, go figure!

We are using rsync to transfer some (millions) files from a Windows (NTFS/CYGWIN) server to a Linux (RHEL) server. We would like to force all file and directory names on the Linux box to be lower case.

Is there a way to make rsync automagically convert all file and directory names to lower case? For example, lets say the source file system had a file named:

/foo/BAR.gziP

Rsync would create (on the destination system)

/foo/bar.gzip

Obviously, with NTFS being a case insensitive file system there can not be any conflicts...

Failing the availability of an rsync option, is there an enhanced build or some other way to achieve this effect? Perhaps a mount option on CYGWIN? Perhaps a similar mount option on Linux?

Its RHEL, in case that matters.

I work for a large web site design and hosting firm. We are planning to introduce a suite of SEO services.

I have been given an interesting task as part of the development effort. The goal is to be able to clearly demonstrate to our clients (who number in the thousands) how many phone calls are being placed into their business by people who visit their web site.

The process by which we hope to achieve this is as follows: we will issue clients (again, thousands of them) a new, toll free number that they will only display on their web site. This number, when dialed by a consumer, will forward the call to the store’s existing, legacy number, but our servers will monitor the length of the call etc.

To illustrate by example: our company might assign a store, “ABC Discount” a new 1-800 number, “1-800-111-1111”. Now, for many years, ABC Discount has had a 1-800 number they have been using in their business, “1-800-999-9999”. A customer goes to ABC discount’s web site, where they see their web only phone number (1-800-111-1111). The customer is interested in ABC Discount’s products, so they dial the 1-800 number on the site (1-800-111-1111). Our server systems (somehow) pick up the call to 1-800-111-1111. The servers know to forward the call to 1-800-999-9999 (which happens silently, in the background, with minimal delay with Caller ID info etc. preserved). Our servers note that the call was made, record the duration etc. and (possibly) even have the capability of recording and saving the audio of the call.

Later, the owner of ABC Discount can log into our web based content management system and see how many calls he received from his web site today, how long each call lasted and even listen to the audio of a call or two to see what was said. This lets us show him how hard his web site is working for him/her. (It is only one of many analytics tools we will have available, BTW) I know that the technology building blocks to implement this sort of solution are out there. There are hundreds (if not thousands) of companies that offer virtual PBX services, all of which (I presume) are built on top of some sort of industry standard equipment etc. It looks to me lie we are basically trying to setup our own (very stripped down) virtual PBX solution (where every business is configured identically, essentially).

No amount of Googling, however, has revealed to me the software/hardware virtual PBX systems/providers are built on. I just don’t know where to start.

We would love to outsource the “guts” of the system to a competent host, if anyone can point us in the right direction, that would be great. It would have to be a real “commercial” grade solution, because we would be managing thousands of numbers and need API level access to provision accounts, track minutes used etc. (IE: don’t tell us to go talk to virtualpbx.com, tell us to go talk to the telephony system hosting provider that virtualpbx.com uses to host its solution)

We are also willing to build our own solution and have the budget to do it (i.e. buying a 30k, 50k or even 100k worth of telephony hardware and collocating it etc. does not scare us in the least…). Anyone know what sort of hardware/connectivity we would want to use? We DON’T want to do any VOIP or anything like that. It needs to be POTS for the most part.

Any ideas on where to start on this project? (Even the contact info of suitable companies/consultants would be useful)...

We are thinking of evaluating Nexenta for a Tier 2 storage solution. It looks great, but I have doubts over its future after the demise of OpenSolaris.

I know that Nexenta uses an OpenSolaris kernel and ZFS. Now that OpenSolaris is no more, will Nexenta die a slow and painful death?

I have scoured the net, and there does not seem to be any discussion on this topic out there. Anyone in the server fault community know if Nexenta has a future with OpenSolaris dead?

Also, anyone have any "in production" Nexenta experiences to share?

Conceptually, and I don't think I am alone here, I love the concepts of ZFS. You can create one big disk, everything is stored on multiple drives, and you can grow and shrink the pool and drive at any time. Plus, the disks don't need to be same etc.

What I hate about ZFS is that, best I can tell, its only truly stable form on Solaris running on Oracle (formerly Sun) hardware. Sure, OpenSolaris exists, and FreeBSD has ZFS support, but reliability is generally said to be poor when on FreeBSD/OpenSolairs/Solaris on non oracle hardware.

We run (exclusively at the moment) Window Server 2008 R2 servers. What is the "next best" DAS storage option for this OS? I can't find a ZFS implementation for windows, so that's out. There are a myrad of RAID cards out there: can someone recommend a setup that approaches ZFS flexibility? A setup where you can just add a disk to the pool and bam more storage without having to take the server down for rebuild?

In Sum: if you love ZFS but want to run Windows Server 2008 R2 whats your best/coolest option(s)?

PS: This is for in production systems, budget is on the order of 10k per system.

We use PFSense firewalls at three installations with the following LAN networks:

1.) Datacenter #1: 10.0.0.0/16 2.) Datacenter #2: 10.1.0.0/16 3.) HQ: 10.2.0.0/16

All of these locations are linked via an IPSEC tunnel that works properly. Hosts in any of the above networks can communicate with hosts in any other of the above networks.

Now, for our laptops etc. we established a road warrior network 10.3.0.0/16 and have implemented OpenVPN to link the laptops etc. to Datacenter #1. This works great too, so our laptops can connect and communicate with any host in Datacenter #1 (anything on 10.0.0.0/16)

The problem is the laptops can't communicate with any hosts that Datacenter #1 can reach by its IPSEC tunnel to Datacenter #2 (and/or the HQ for that matter).

Does anyone know what to do configuration wise on the PFSense box in Datacenter #1 to configure to route packets received on the OpenVPN tunnel to Datacenter #2 over the IPSEC tunnel? It could be a setting on the OpenVPN or some sort of static route or some such.

Any ideas?

On Windows Server 2008 R2, one must specify a user for each scheduled task to run under.

We have a policy whereby each domain user must change their password every 30 days.

This creates a problem, as we have many scheduled tasks that need to run on a large number of servers. When the user associated with a given scheduled task's changes his/her password (as required every 30 days), the tasks all stop working (until you log in and change the password on the task).

We tried to run the scheduled tasks under the SYSTEM account but this did not work.

I have thought of creating a new account (SCHED_TASK) that can not log in interactively, is an administrator, and is exempted from the password change requirements.

This seems very much a "hack" to me. Is there a better way or a built in account of some sort that we should use?

What is the best practice in this instance?

With a recent version of CYGWIN (installed yesterday) on Windows 7, we are having a weird problem getting rxvt to work.

We used to use the following batch file:

@echo off

C:
chdir C:\cygwin\bin

rxvt -fn 18x18 -bg black -fg white -g 85x55 -sr -e bash --login -i

To launch rxvt. On our new machines, rxvt starts and immediately terminates.

Using the supplied cygwin shortcut (to lauch bash) and then executing the command:

rxvt -fn 18x18 -bg black -fg white -g 85x55 -sr -e bash --login -i

produces the right result, however.

Any ideas how to make a batch file that will launch rxvt properly? What has changed that is leading to this script (which we have been using for years) to suddenly stop working?

We are running Active directory on Windows Server 2008 R2. All servers and desktops are Windows 7 or Server 2008 R2.

Is it possible to create a policy using GPO such that every account in an OU will have a certain set of Folder Option settings (Windows Explorer) set automatically and consistently whenever users in the OU log into any computer on the domain?

Basically, I want "Show Hidden files, folder and drives" enabled, fire extensions and hidden files visible on all machines automatically on all machines and we want to achieve this without using roaming profiles

Any ideas?

We have a ProCurve Switch 2810-48G (J9022A).

We need to disable auto negotiation on two ports and manually configure them to be full duplex gige ports.

From the web GUI, Configuration Tab, Port Configuration sub tab, I am only presented with the option to configure the port as Auto - 1000. I take this to mean, auto negotiate duplex, manually configure the speed to be gige.

From the CLI when I try to set 1000-full I get the following error: Value 1000-full is not applicable to port 39 (or whatever port I try) The exact commands I have entered are:

config interface 39 speed-duplex 1000-full

BTW: speed-duplex auto-1000 works ( I also tried full-1000 and that did not work either)

How do I manually configure the port such that it is manually configured to use full duplex, 1000 mbs?

On any computer that has a web browser that supports flash, one can use SpeedTest.net as a crude way to measure network performance.

It gives tree very important numbers. Ping time, sustained upload speed and sustained download speed to a menu of different hosts around the world.

Is there a method to get the same information given that you have access to only one *nix box?

I know that one can run ttcp or wget if you have access to two machines (say, one in NYC and another in LA or whatever). I know you can run ping to get the ping (obviously) but is there a way to do something like speedtest.net against a list of reference servers?

Best I can come up with is running wget or curl against a large file...

We are building some firewalls that run PFSense. The systems use the SuperMicro Super X8DTN+ motherboard, which includes a dual link intel gigabit network card (Intel 82576 controller)

I have tried this out on several different identical boxes -- always the same result.

Anyway, if I configure PFSense to use the IGB driver for the intel 82576 and I setup one port of the dual port controller to be WAN and the other to be LAN I get terribly slow speeds (about 10 kB/sec). This with wan=igb0, lan=igb1

If I add an additional, separate PCI-X network card (using the EM driver) i get phenomenal speeds (wan=em0, lan=igb0). Kicker is -- the igb driver is still in use!

Any ideas why this might be or how to go about fixing it?

We currently operate three clusters of collocated machines in different data centers.

Recently, I became aware that our newest data center will offer to cross connect us to a bandwidth provider free of charge.

In the past, I never really investigated a cross connect for bandwidth because I figured that the rates would be similar to what we are paying the colo now and that it would reduce our resiliency (because we would only be using one or two carriers for IP, where as the colo uses, say 8 different providers).

Then I saw an ad for hurricane electric internet services (http://he.net/cgi-bin/ip_transit_quote) that gave a price for IP transit at $1/Mbs, which is much better than the $30/Mb we pay for the blended bandwidth.

What are people out there typically paying for bandwith via cross connect and how hard is to setup? Is my understanding that what you do is open agreemetns with two or three ISPs, cross connect to them and then configure your top of rack router on their network. Can you really get IP transit down to a couple of dollars per megabit per month just by doing the routing yourself?

Or, is my understanding of cross connection fundamentally wrong?

Does anyone know the details as to the differences between the TCP/IP stack in Windows Server 2008 (Standard) vs Windows 7 Ultimate?

This question pertains to TCP/IP ONLY.

I know that Windows 7 and Server 2008 have a unified driver model -- so the driver portion of the TCP/IP stack is, theoretically the same.

On precisely the same hardware, supporting a heavily loaded TCP/IP app, will the Server 2008 TCP/IP stack offer any advantages over a out of the box Windows 7 Ultimate box?

I have heard many (vauge) insinuations as to advantages of the server's stack, but have never seen anything concrete like: Server supports a XYZ megabyte flux capacitor buffer, while Windows 7 only supports a 2 mb flux capacitor buffer...

A link to a tech doc or site outlining the differences would be awesome.

We use a lot of SuperMicro gear, and we just noticed that SuperMicro now has a Layer 3 managed switch. I am interested in using it in a top of the rack application.

As with most of SuperMicro stuff, the price is right.

I can't seem to find anyone online who has tested/reviewed/used on of these switches. Before plunking down the money, I wanted to find a review or hear someone's good experiences with this gear.

Anyone used a SuperMicro layer 3 switch (particularly SSE-G48-TG4) and have anything good or bad to say about them?

I have isolated a problem down to this:

On Windows XP Pro, SP2 if one does an experiment where one tries to open TCP/IP sockets in a loop (bascially, listen port 7000, listen port 7001, etc.)

After approx 649 open sockets, one will start getting errors: No buffer space available (maximum connections reached?)

The question is, what registry settings etc. does one edit to increase the number of listening sockets one can open before receiving this error?

Much is made of the fact that VMWare's ESXi hypervisor is "free"

As best I can tell, you can install the hypervisor on a host for "free".

Because ESXi does not have a built in management console, you need a program, of some sort, to connect to the ESXi hosts to "manage" them. By "manage" I mean, start, stop, install, reboot and backup vms.

If you install the free ESXi on a host and connect to it via a web browser, you are prompted to download vSphere to manage the host. OK, but vSphere is, as best I can tell, not free. When you install it you are continuously reminded that you have only 60 days to evaluate vSphere.

My question is this: Is there a completely free management tool for ESXi hosts that enables one to:

• Create VMs
• Modify VMs settings (memory etc.)
• Power VMs on and off
• Backup the VM (via any means)
• Resore a VM from a backup

Failing that, without licensing something from VMWare, is there any tool that will let you manage your hosts after the 60 day evaluation period of vSphere ends?

I have not found a straightforward explanation of this on VMWare's web site. Does anyone out there know the answer (even better if you can point me to a clear explanation on VMWare's website...)

Assume that you have 1 rack in a reliable colo facility. The colo (obviously) has advanced chassis switches and can provide any reasonable manner of drops specified (but a limited number of drops). That is to say, you can specify two GB cat6 drops configured such that (specify additional config here).

Also assume that you have N (say 10) "servers" each with 2 GB ethernet ports. Each server needs to have one always accessible, routable ip address. That is to say, each server has an IP address WWW.XXX.YYY.ZZZ that should be pingable from any properly configured host on the internet.

What is the simplest logical and phsical network topology you can install top of rack such that there is no one single point of failure leading to ip connectivity issues between the servers and the gateway provided by the colo?

By simple, I mean, generally speaking, cheapest to implement using Cisco networking gear. That is a rough definition, but I think it should correlate well with the answer I am after.

Our company runs e-commerce web sites (thousands) on two clusters in two separate data centers.

Basically, all we require to operate is rack mountable server nodes. Each node needs:

1.) 4 or 8 cores 2.) 32 gb ram 3.) 1 250 gb sata disk 4.) 2 Port, Gigabit, Ethernet Adapters 5.) Ability to boot Windows XP Pro

That it. We run about 40 such nodes in a fully redundant, always up (hopefully!) cluster (we wrote the clustering part ourselves)

Previously, we bought our systems whiteboxed (basically had a small shop custom build our servers (supermicro) to our specs).

This scheme was working well up to our last round of node purchases. Out of the last round of node purchases have had a super, super high failure rate (30% failed in 6 mo.) No one reason, bad PSU, bad memory, mobo fried, etc.

My questions are these:

Will we have more consistent reliability if we purchase from a name brand vendor (IBM/DELL/HP) or are we basically in the same crap shoot of reliability we were in before? Remember, these are low end servers. We are not going to transition to a mainframe or anything exotic.

Will our reliability vary with the form factor of the servers? That is to say, will 2u servers be any more reliable than high density, 2 nodes in a 1 u box server?

Anybody out there transitioned from white box servers to name brand servers (or changed form factors) and have a tale to tell?