Dixon-Ratzlaff Family Blogs

Back when I started this job and wanted network monitoring tools, I was already familiar with MRTG and I soon found Nagios, two great open source projects that did exactly what I needed: give me traffic graphs and let me know when something goes down. The only problem with both of those packages was that configuration was a manual process, editing configuration files. As our network grew, I started looking for alternatives.

I found OpenNMS and used that for quite a while.  It combined the functions of my MRTG and Nagios, and seemed a little friendlier to configure.  Then when we were looking for help desk software, we found Spiceworks (not open-source, but free!), which also does a good job of inventorying all our equipment and also notifying us when something is awry.

Over time, our OpenNMS installation grew slower and slower and finally started locking up at various times.  I found I didn’t really use it as much as I used to since we had Spiceworks, so I just turned it off.

But recently I found myself wanting the traffic graphs again.  Some performance issues had cropped up on the network, and I needed to see what was going on.  I actually considered setting up MRTG again, until I found Observium.

I followed a step-by-step walkthrough on how to get it set up on CentOS 6.  One of the last steps was “add your first host manually” via the command-line.  I chose our core Cisco router and let fly, but was amazed at how much data was flashing across my screen.  Observium was intelligent enough about Cisco devices that it accessed CDP information and other data to discover all the rest of my routers and managed switches I had on the network.  Most of the system ended up getting set up in that one command.  The few devices it missed were easy to add via the web interface.

I haven’t had good luck adding Windows servers yet, mostly because it seems that Windows has so many “interfaces” defined that really don’t mean anything to me.  It was too much work turning off monitoring to all the interfaces I didn’t care about.  Also, Spiceworks does a decent enough job monitoring CPU and RAM usage on Windows that I didn’t see the need to have the data twice.

In addition to logging traffic on network interfaces, Observium collects data on whatever the device will provide.  For example, from our Cisco 7206 VXR router Observium collects CPU and memory utilization, temperatures, voltages, power supply status, and all manner of statistics from the network interfaces.

Right now, I’m just using Observium as a kind of Super MRTG on Steroids, but it can also send out notifications if something goes down, and I’m interested in setting it up to receive interface status traps from my routers to get even quicker notifications when a link goes down.

I like it.  It’s pretty, it offers a lot of useful information, and it’s easy to configure.  And it’s free and Open Source.

One of our servers hadn’t been getting updates.  Once I finally stripped some of our group policies off the box to allow me to check for updates manually, Windows helpfully reported “Code 80244022 – Windows Update encountered an unknown error”.

Searching Google for this error code returned a Microsoft article which was also less than helpful.  But the bit about checking proxy settings got me to thinking.

Windows Update uses winhttp and its proxy settings rather than Internet Explorer’s proxy settings.  Sure enough, when I checked I found a proxy set for winhttp and no exceptions listed.  Our WSUS server is on the same LAN as the problem server, and that server has no need to access the Internet, so I simply cleared the proxy settings:

C:\Windows\system32>netsh winhttp reset proxy

Current WinHTTP proxy settings:
    Direct access (no proxy server).

Then after restarting the Windows Update service, everything worked correctly.

I came across a TechRepublic article today about how DropBox could be used as a conduit to drop malware on a corporate computer.  Long story short – DropBox wasn’t the security hole, just the delivery mechanism.  An attacker was able to utilize a compromised laptop that synced to a corporate desktop via DropBox to deliver malware files to the corporate desktop.  Since the default location for the DropBox files were in the user profile, when the malware ran, there were no UAC warnings.

Many of the comments on the article had to do with, “What if DropBox wasn’t installed in the user profile?”  But that’s not the real problem in my mind.  I wondered, “Why do users have access to execute programs in their profile at all?”

In our network, we have AppLocker group policies in place to disallow execution from anywhere but trusted program locations.  Our users do not have access to write in those trusted locations, and the AppLocker policies do not allow execution from anywhere our users do have access to write files. (such as their user profile and removable drives)  This helps prevent our staff from installing or running software they download or bring from home (which is against our company policy), and it also mitigates the types of attacks described in the TechRepublic article.  If it can’t be executed, it can’t infect.

Of course, AppLocker policies can only be enforced on workstations running Windows 7 Enterprise, Windows 7 Ultimate, or Windows 8, or on servers running 2008 R2 or 2012.

Bing took over Yahoo Search a while back, in order to compete with Google. Why then is Google still the big frontrunner in Internet search? From my perspective, because they’re responsive and accurate.

We recently redesigned our website. It really needed it, we still had a frame-based layout with almost no search engine optimization, and quite a bit of the information was outdated. After the redesign, all the pages had new URLs, and the old links were no longer good.

I set the web server to generate user-friendly 404 errors, and set up 301 redirects for our “domain.com” to point to our “www.domain.com”, and did all the standard SEO stuff. We have a sitemap XML file that generates dynamically from our website structure, and robots.txt mentions it. We also have a news page that generates an RSS feed that’s published through FeedBurner.

I submitted the new sitemap url to both Google and Bing (and also through the old Yahoo SiteExplorer) 10 days ago. So far, all of our new URLs are showing up in Google search results, including our news articles. But Bing still returns all of our old URLs, very few of the new ones, and none of the news articles.

When I check on Bing’s Webmaster tools page, it still says my sitemap submission is “pending”, and when I check my web server logs, Bing’s robot has never fetched our new sitemap, even though it has fetched our new robots.txt file several times.

After seeing the inaccurate results that Bing returns for our site (pages that have been dead for 10+ days), I certainly wouldn’t be using it for my searches anytime soon. In my opinion, Bing will need to me a lot more responsive to changes in the web and be more accurate before they can ever hope to compete with Google in search.

First of all, when we first decided that Windows Backup was not going to be sufficient to back up our servers anymore, (specifically since 2008′s version doesn’t support tape drives anymore) we looked at several packages. Backup Exec (whoever owns it now) was the software I was most familiar with, but it’s almost the priciest package out there. Then we found NovaBACKUP from NovaStor.

NovaBACKUP did everything we needed – it backs up open files, SQL databases, and Exchange server. Then we learned about NovaStor’s storage server technology, which allows you to back up to a remote server in another location. Their Remote Workforce product incorporates all these features into a single solution. And it’s priced much lower than anything else I found.

So, we’ve been running this package for almost a year now. The “cloud” backups work using a technology NovaStor calls “FastBIT” which generates a patch file containing only the changes since the last backup, thereby minimizing the bandwidth required to perform the new backup. This worked great for us, as we wanted to locate our storage server in another location connected via T-1. The drawback to this technology is it requires quite a bit of free space to generate the patch files, and the process (especially for large files) is not exactly speedy. Backing up one of our large databases took just over an hour to complete.

With their new 12.1 version, they’ve included a new FastBIT algorithm which requires less disk space and works much more quickly. After performing the upgrade, that same database backup now only takes 20 minutes!

If you’re looking for a backup solution that supports local tape devices as well as several different online backup options, check out NovaStor.

We’ve had an ASP application for some time now that generates PDF files from database data and outputs it to the browser. When we first wrote this app, we purchased the CutePDF Form SDK. It works very well, making it easy to enter values into our PDF forms programmatically and output the finished document as a stream to the browser. And it’s reasonably priced at around $250 for a single server.

Now that we’re converting our app over to ASP.NET, I decided to look around for other options. CutePDF’s Form SDK has a COM interface, and I wanted to see if I could find a library that was .NET native.

I found ABCpdf from WebSupergoo Software. It’s written in .NET, but it also has a COM interface for backwards compatibility with ASP. It does everything we were using CutePDF Forms SDK for and more. In addition, at least in my testing, the ABCpdf component runs slightly faster than CutePDF’s.

What really made this a clincher for us is WebSupergoo is giving away licenses of this component for free. All you have to do is link back to their website.

We’ve had an ASP application for some time now that generates PDF files from database data and outputs it to the browser. When we first wrote this app, we purchased the CutePDF Form SDK. It works very well, making it easy to enter values into our PDF forms programmatically and output the finished document as a stream to the browser. And it’s reasonably priced at around $250 for a single server.

Now that we’re converting our app over to ASP.NET, I decided to look around for other options. CutePDF’s Form SDK has a COM interface, and I wanted to see if I could find a library that was .NET native.

I found ABCpdf from WebSupergoo Software. It’s written in .NET, but it also has a COM interface for backwards compatibility with ASP. It does everything we were using CutePDF Forms SDK for and more. In addition, at least in my testing, the ABCpdf component runs slightly faster than CutePDF’s.

What really made this a clincher for us is WebSupergoo is giving away licenses of this component for free. All you have to do is link back to their website.

We’ve had an ASP application for some time now that generates PDF files from database data and outputs it to the browser. When we first wrote this app, we purchased the CutePDF Form SDK. It works very well, making it easy to enter values into our PDF forms programmatically and output the finished document as a stream to the browser. And it’s reasonably priced at around $250 for a single server.

Now that we’re converting our app over to ASP.NET, I decided to look around for other options. CutePDF’s Form SDK has a COM interface, and I wanted to see if I could find a library that was .NET native.

I found ABCpdf from WebSupergoo Software. It’s written in .NET, but it also has a COM interface for backwards compatibility with ASP. It does everything we were using CutePDF Forms SDK for and more. In addition, at least in my testing, the ABCpdf component runs slightly faster than CutePDF’s.

What really made this a clincher for us is WebSupergoo is giving away licenses of this component for free. All you have to do is link back to their website.

We’ve had an ASP application for some time now that generates PDF files from database data and outputs it to the browser. When we first wrote this app, we purchased the CutePDF Form SDK. It works very well, making it easy to enter values into our PDF forms programmatically and output the finished document as a stream to the browser. And it’s reasonably priced at around $250 for a single server.

Now that we’re converting our app over to ASP.NET, I decided to look around for other options. CutePDF’s Form SDK has a COM interface, and I wanted to see if I could find a library that was .NET native.

I found ABCpdf from WebSupergoo Software. It’s written in .NET, but it also has a COM interface for backwards compatibility with ASP. It does everything we were using CutePDF Forms SDK for and more. In addition, at least in my testing, the ABCpdf component runs slightly faster than CutePDF’s.

What really made this a clincher for us is WebSupergoo is giving away licenses of this component for free. All you have to do is link back to their website.

We’ve had an ASP application for some time now that generates PDF files from database data and outputs it to the browser. When we first wrote this app, we purchased the CutePDF Form SDK. It works very well, making it easy to enter values into our PDF forms programmatically and output the finished document as a stream to the browser. And it’s reasonably priced at around $250 for a single server.

Now that we’re converting our app over to ASP.NET, I decided to look around for other options. CutePDF’s Form SDK has a COM interface, and I wanted to see if I could find a library that was .NET native.

I found ABCpdf from WebSupergoo Software. It’s written in .NET, but it also has a COM interface for backwards compatibility with ASP. It does everything we were using CutePDF Forms SDK for and more. In addition, at least in my testing, the ABCpdf component runs slightly faster than CutePDF’s.

What really made this a clincher for us is WebSupergoo is giving away licenses of this component for free. All you have to do is link back to their website.