gregmac's questions

I'm trying to run Register-ScheduledJob from a script executing as SYSTEM (from an external deployment tool), but I get an error. This must be initially invoked as SYSTEM due to the way the deployment tool works.

To reproduce this problem (requires powershell.exe running as SYSTEM), I created another scheduled job as follows:

• Run as user account: SYSTEM
• Action:
  • Run: powershell.exe
  • Arguments: Register-ScheduledJob -name testjob -filepath c:\target.ps1 > C:\testjob.txt 2>&1

Note: if you want to reproduce this yourself, you'll also have to create c:\target.ps1 as an empty file.

This job can then be executed from Task Scheduler UI, and you can see the output in c:\testjob.txt.

When the script runs Register-ScheduledJob, it shows the following error:

Register-ScheduledJob : An error occurred while registering scheduled job 
definition testjob to the Windows Task Scheduler.  The Task Scheduler error is: 
(32,4):UserId:.
At line:1 char:1
+ Register-ScheduledJob -name testjob -filepath c:\target.ps1 > C:\testjob.txt ...
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
    + CategoryInfo          : InvalidOperation: (Microsoft.Power...edJobDefini 
   tion:ScheduledJobDefinition) [Register-ScheduledJob], ScheduledJobExceptio  
  n
    + FullyQualifiedErrorId : CantRegisterScheduledJobDefinition,Microsoft.Pow 
   erShell.ScheduledJob.RegisterScheduledJobCommand

As far as I can tell this has something to do with not being able to create a job as the SYSTEM user. I'm not really concerned with what target.ps1 actually runs as (eg it doesn't have to be SYSTEM), so long as it's an account that I don't have to manage a password for.

My script has to be initially invoked as SYSTEM (from a service which is running as SYSTEM), so how can I use it to create a scheduled job?

I'm experimenting with CloudFormation right now. I deployed a stack that included an AutoScalingGroup (and LaunchConfig and CloudFormation::Init), LoadBalancer, a single EC2 instance (and CloudFormation::Init), and a couple security groups.

If I terminate EC2 instances created by the AutoScalingGroup, they get recreated as expected. This is actually quite useful for testing changes to my LaunchConfiguration and CloudFormation::Init settings (basically making sure it works from scratch).

I also have an EC2 instance (named DnsServer1) that has a public IP. While working on its CloudFormation::Init data, I also manually terminated its instance. I know CloudFormation doesn't restore it like an AutoScalingGroup does, but I was expecting it would be recreated when I ran aws cloudformation update-stack.

Instead, I get the error: AWS::EC2::Instance DnsServer1 i-014eee8720c4fb542 does not exist.

Full update log:

11:03:40 UTC-0500   UPDATE_ROLLBACK_COMPLETE    AWS::CloudFormation::Stack  GregCFTest1 
11:03:38 UTC-0500   UPDATE_ROLLBACK_COMPLETE_CLEANUP_IN_PROGRESS    AWS::CloudFormation::Stack  GregCFTest1 
11:03:36 UTC-0500   UPDATE_COMPLETE AWS::AutoScaling::LaunchConfiguration   WebServerLaunchConfig   
11:03:36 UTC-0500   UPDATE_COMPLETE AWS::EC2::Instance  DnsServer1  
11:03:13 UTC-0500   UPDATE_ROLLBACK_IN_PROGRESS AWS::CloudFormation::Stack  GregCFTest1 The following resource(s) failed to update: [DnsServer1].
11:03:11 UTC-0500   UPDATE_COMPLETE AWS::AutoScaling::LaunchConfiguration   WebServerLaunchConfig   
11:03:10 UTC-0500   UPDATE_FAILED   AWS::EC2::Instance  DnsServer1  i-014eee8720c4fb542 does not exist
11:03:00 UTC-0500   UPDATE_IN_PROGRESS  AWS::CloudFormation::Stack  GregCFTest1 User Initiated

What can I do to get CloudFormation to recreate this resource, short of deleting the stack and recreating it?

I'm attempting to deploy a .NET application to ElasticBeanstalk using the AWS-cli tools.

First upload then create version using:

aws elasticbeanstalk create-application-version 
   --application-name "MyApp - DEV" 
   --version-label "0.1.165" 
   --source-bundle S3Bucket="xxx",S3Key="MyApp.0.1.165.zip" 
   --process 

This works, so next step:

aws elasticbeanstalk update-environment 
  --environment-name "myapp-dev-env" 
  --version-label "0.1.165" 
  --application-name "MyApp - DEV" 

This sometimes exits with code 255 and outputs:

A client error (InvalidParameterValue) occurred when calling the UpdateEnvironment operation: Application version is unusable and cannot be used with an environment

What is frustrating is these exact steps (run by my continuous integration server) sometimes work, and if I go and deploy this version using the AWS console, it also works. The zip I'm uploading doesn't appear any different, and in fact, if I rebuild from the same source revision (using a build that previously deployed successfully, which gives it a new version number), it seems to (usually) run into this problem.

I can't find any detail about what this error means, so where else can I look to figure out what's happening?

I'm trying to control apache based on cookie values, but I can't seem to get SetEnvIf to work with HTTP_COOKIE. I've boiled this down to some simple logic to isolate the issue and be easy to test.

Apache 2.2.22 on Ubuntu 12.04.1 LTS.

What I'm using is:

     Header set Set-Cookie "cookie1=1"
     SetEnvIf HTTP_COOKIE "cookie1=1" is_cookie1
     Header set Set-Cookie "cookie2=2" env=is_cookie1

Using Chrome's resources tab, I am inspecting the cookies for the page. What I expect to see is:

• First page load, cookie1=1 exists
• Second (and subsequent) page loads, cookie1=1 and cookie2=2 exist.

Instead, all I ever get is cookie1:

If I add the line:

     SetEnvIf Remote_Addr ^192\.168\. is_cookie1

Then cookie2 is set immediately, as I'd expect, so the last Header ... env=is_cookie1 line appears to be fine.

I also tried to verify that HTTP_COOKIE was being set correctly:

     RewriteRule ^/test/$ /test/%{HTTP_COOKIE} [R=302,L]

Now going to /test/ immediately redirects to /test/cookie1=1%3b%20cookie2=2 as I was expecting, and so HTTP_COOKIE seems to be set properly.

I've also tried a bunch of variations of SetEnvIf and nothing seems to work:

     SetEnvIf HTTP_COOKIE "^cookie1=1$" is_cookie1
     SetEnvIf HTTP_COOKIE ^cookie1=1$ is_cookie1
     SetEnvIf HTTP_COOKIE "^.+$" is_cookie1
     SetEnvIf HTTP_COOKIE ^.+$ is_cookie1

..although

     SetEnvIf HTTP_COOKIE ^.*$ is_cookie1

sets cookie2 immediately (on first load) in any situation (which... is not useful at all. but at least it tells me this line does something).

What am I doing wrong?

Windows (going back to at least IE6) by default has WPAD turned on (Internet Options > Connections > LAN Settings > Automatically Detect Settings).

When you don't have a proxy or WPAD server, browsers initially take several seconds saying something to the effect of "Discovering proxy server.." before timing out and loading the page directly.

I can find lots of stuff about how to set up your proxy information via WPAD, but not how to configure WPAD to tell clients you have no proxy.

WPAD works via both DHCP and DNS, however Firefox only supports the DNS method, so a complete answer ideally provides configuration of both DHCP and DNS.

Proxies were around before NAT, and so early on in the internet, they were popular ways to get an entire network online with only a single public IP. NAT initially required complex software and/or hardware to get it working, and so understandably some organizations still used proxy servers to provide access.

In the last decade however, NAT has become immensely popular, and is available in virtually every router on the market. The hardware is so fast and cheap that unless you are trying to use a $50 home router to run say, a 5000 person office, you're not going to have issues. Firewalls are also smart enough to do complex authentication, easily allowing different users to access different services at different times of day (even the most basic home routers often have this functionality built-in). In fact, NAT is likely easier to set up and manage than a proxy server.

In my job I still often deal with clients using proxy servers to provide internet access, and the ensuing issues that arise from authentication, what account services are running as, etc. This includes some companies that are very big and old (and so I guess got their network installed over a decade ago, and just kept it that way) as well as relatively new organizations that have only a few years old.

So my question is, why in the heck are there still so many places using proxy servers for this purpose?

I have a new system running under VMware, and I'm just trying to make sense of the memory use.

In vSphere client, it shows me active using 335,544 KB:

However, linux shows me using 3,146,148 KB:

root@PFDEV-SVN:~# free -k
             total       used       free     shared    buffers     cached
Mem:       4118636    3146148     972488          0     477216    1268364
-/+ buffers/cache:    1400568    2718068
Swap:      6040400          0    6040400

Why is this so different? Am I just reading this incorrectly?

This is Ubuntu 9.10, without vmware tools installed, running on an ESXi server.

I have a bunch of files that were exported (from my trac wiki) that are named like ParentPage%2FSubPage but they actually should be like ParentPage/SubPage. Anyone have a quick and dirty way to rename and organize them, so that ParentPage will be a directory, and then SubPage will be a file inside that?

Slightly modified from @Christopher Karel's answer, which is what I used:

for FILE in $(ls | grep "%2F")
do
  CONVERTED=$(echo $FILE | sed -e 's/%2F/\//g')
  DIRNAME=$(dirname $CONVERTED)
  if [ -f $DIRNAME ]; then  mv $DIRNAME $DIRNAME.page; fi
  mkdir -p $DIRNAME
  if [ -f $DIRNAME.page ]; then  mv $DIRNAME.page $DIRNAME/$(basename $DIRNAME); fi
  mv $FILE $CONVERTED
done

and the test data I used (in an empty dir), before running the above:

touch Test; touch Test%2FA%2F1; touch Test%2FA%2F2; touch Test%2FB

I'm in the process of deploying some new PCs (there's only 8), all identical hardware. What I'd like to do is install Windows 7 (64bit), join to domain etc, install a bunch of other software, and then clone that drive to multiple other machines. I'd also like to be able to use it as a backup image, so the machine can be restored back to that image at some future date.

I understand this involves at least sysprep, but I am confused after reading some tutorials that talk about using Windows Automated Installation Kit, or hacks with the registry and custom-build batch files. This process seems overly complex to me: I did something similar 10+ years ago, and and don't remember it being this bad. Surely things have improved in a decade?

There's also some products that involve having network servers running deployment software, network boot, etc etc.. this is way more than I want to set up.

My systems are all identical hardware. Is there a simplified way to clone PCs?

Preferably (since I'm a lazy developer, and not an IT admin) I'd like to find some off-the-shelf product that I can run after I get the machine setup, that will spit out a bootable DVD I can run on all the other systems, which will boot up, ask for a computer name, join it to the domain, and that's it. Does such as product exist?

Does anyone know of an app to manage a wall-mounted display? I'm looking for something that can show a couple web pages side-by-side, and periodically refresh them, or flip back and forth between a couple of pages. It should show in full-screen on start up, and have all this configuration saved so if the system reboots someone doesn't have to fiddle with it.

This would display things like build status, network traffic, performance stats, service availability, etc.

I've looked around, and there are "digital signage" application for high-end markets (eg, running airport flight info screens), and "video display wall controllers" that can do some of this but are really designed to run a matrix of 16 big displays, often with networking capabilities, and the price tag to match.

On the low-end all solutions I've found seem to be one-offs: people are just using screens that some application happens to provide, and rarely, using a web browser in a kiosk mode on startup to display the page (seems like most people don't automate this). This just seems like a primitive way to go - surely someone has written this app? Am I using the wrong terms?

(And in case it isn't obvious, I'm on the low-end market here, I just have a single big display I want to show stuff on).

I'm trying to set up a SOCKS proxy using dante for testing purposes. However, I can't even get it to work with a web browser, after looking at several tutorials on how to do that.

I've tried in both IE and Firefox, in both cases, using "Manual proxy configuration", leave everything blank except for SOCKS host, and then put in the IP of my proxy and the port number (1080). I just get "Server not found" / "Problems loading this page" and don't see anything in danted, even running in debug mode.

If I do a "telnet 10.0.0.40 1080" I do see the connection open in danted debug output, so I know that much is working.

Here's my config:

logoutput: stdout /var/log/danted/danted.log
internal: eth0 port = 1080
external: eth0
method: username none #rfc931

user.privileged: proxy
user.notprivileged: nobody
user.libwrap: nobody

connecttimeout: 30   # on a lan, this should be enough if method is "none".

client pass {
        from: 10.0.0.0/8 port 1-65535 to: 0.0.0.0/0
}

client pass {
        from: 127.0.0.0/8 port 1-65535 to: 0.0.0.0/0
}

client block {
        from: 0.0.0.0/0 to: 0.0.0.0/0
        log: connect error
}

block {
        from: 0.0.0.0/0 to: 127.0.0.0/8
        log: connect error
}

pass {
        from: 10.0.0.0/8 to: 0.0.0.0/0
        protocol: tcp udp
}

pass {
        from: 127.0.0.0/8 to: 0.0.0.0/0
        protocol: tcp udp
}

block {
        from: 0.0.0.0/0 to: 0.0.0.0/0
        log: connect error
}

I'm sure I'm probably missing something simple, but I'm lost. I haven't even thought about SOCKS since the late 90's.

I want to create a rule that forwards all mail with a specific sender to another email address.

From what I can tell, this might be possible with sender_bcc_maps but I can't find any examples using a specific address, only domain matches.. additionally, the user does not exist in postfix, this is simply an outbound relay server (on a static IP, in a datacenter) for another Exchange server (on a dynamic ip, in an office).

I need to forward (equivalent of BCC) all messages to/from a given user to another user on Exchange server 2003. Is this possible?

I can do it for inbound using How to add an email forwarding in Exchange 2003 but it doesn't apply to outbound messages.

(I also realize if the user is sending directly to another SMTP server, this would not be possible - I just need it to work when they are sending through Exchange).