I was redoing our "news" section of www.bitshop.com to pull from a RSS feed of our blogs and ran into several more complications than I expected to have.

First let me preface with I did this as a quick measure, I feel custom code to render the blog news IN HTML (not in javascript) is useful, that way the search engines pick it up and see the main page (or all pages) of your site are changing regularly.

When you do the same in javascript it APPEARS the same to the end users, but to search engines it isn't displaying your blog at all. Realize this.

Back to the issue:

I decided to use Google's AJAX Feed API, their example is pretty simple and shows the rss item's title, however it doesn't show a link. Adding a link you wouldn't think is hard, yet the way it's coded they add TEXT, so I can't put in < a href> to get a link.

NOTE: your script src including your google key, something like http://www.google.com/jsapi?key=.....

Unfortunately the software on this site is messing up the below code, feel free to email me and I'll send you a copy of this code if this doesn't answer your questions.

 < / script >



google.load("feeds", "1"); 

function initialize() 
 { 
 var feed = new google.feeds.Feed("http://www.bitshop.com/Blogs/tabid/95/rssid/7/Default.aspx"); 
 feed.load(function(result) 
   { 
   if (!result.error) 
      { 
      var container = document.getElementById("feed"); 
      container.innerHTML = '';
      for (var i = 0; i < result.feed.entries.length; i++) 
         {
         var entry = result.feed.entries[i];
         var newSpan = document.createElement("span"); 
         newSpan.innerHTML = '<a href="' + entry.link + '">' + entry.title + '</a>< br>';
         container.appendChild(newSpan); 
         }
      }
    }
  );
 }

 google.setOnLoadCallback(initialize); 

< / script >

Loading BitShop News...
 

You'll notice I add a span and set the innerHTML of the span, then append it to the container (The container is the div id feed).

This code I didn't find online easily and was surprised this wasn't easy to find.

References:

[+]

---

Permalink (referenced by: 164 posts / references: 0 posts) 0 comments [view comments]

I'll confess to using a few laptops, however the one I carry most and use the most is a Lenovo Thinkpad X61 which is a decent tablet PC.

I've had about a dozen times the system won't boot saying drive failed, usually if I throw it on the ground and pick it up it'll work (ok, I'm exaggerating..).. But it's obviously something "loose".

Today I couldn't get it to boot and pulled out the disk drive cover screw (on bottom under the right shift key there's a Phillips screw). The disk drive was installed, the cover on, the padding there - yet the disk drive wasn't connected to the laptop at all. I simply pushed it back in, put back on the cover, and everything worked.

This almost seems like either a) my padding around the disk is worn, b) a piece of padding was left off at the factory.

I was in a rush to do something on it so booted and did my work, however I'm curious if others have had this same problem. A quick search didn't turn up any stories like this so I am posting this blog entry.

Hopefully this helps someone get their laptop to boot and avoids a service call / data loss for a simple "fit" issue.

Steve Radich
BitShop, Inc.

---

Permalink (referenced by: 0 posts / references: 0 posts) 0 comments [view comments]

So here I am, Memorial Day weekend, doing some testing of a new NAS which is in the process of going into production after passing most of our tests (the ones it hasn't passed we're considering non-critical at this point). The system is OpenSolaris based, using ZFS which IMHO is the best storage pool / file system around right now - at least at a reasonable price point. I confess I'm not overly familiar with the higher priced ones as we *REALLY* value something we can dig into ourselves if there's a problem (i.e. open source is good for that). It would be nice if Sun officially supported the components we need, but they're only in OpenSolaris as of today.

Now, let's take a simple task - Add 6 drives to a SAS enclosure using an Adaptec storage controller (latest HBA "Performance" series, not the value series controllers). What happens? 3 drives recognized as added, 3 not.

Now right here let me stop and say we've already had to waste about 20 hours with Adaptec to get the controller to work - When we added SATA drives it locked up until flashing the controller. There's no mention that anyone pointed us to saying this is a known issue, no mention of it on the knowledge base that we found, etc. That put us NOT very happy with the quality of this controller previously. We've worked through all those problems, but I'm not very happy with the Adaptec controller including a few features that are "missing" like ability to export unassigned drives as JBODs to the host OS (3ware and areca both do this I'm 99% sure, I should verify but don't have time to right now).

So 3 drives are recognized, 3 aren't. Let's unplug the 3 that aren't working and try re-plugging them. Nope, no dice. Controller reports drives removed (first thought, good, it recognized them)... oh %(*&! those were drives that are in production that it says were removed - they weren't removed though!

So here we are, a production server that's supposed to handle 250+ drives being attached, and Adaptec can't figure out what drives are being plugged in and what are being removed.

Let's try a bunch of things, find the drivse didn't seem to actually disappear (although the ones that disappeared luckily were only hooked up for a performance test, so I honestly did little about it saying they were removed - Instead I concentrated on the new drives). So far I'd say everything sounds like an enclosure problem, not controller, but...

I get a notice for a SMART warning on a new SATA drive (don't laugh, but about 4 of 7 new drives had issues in the first 72 hours. When i say new I mean right out of the box. Generally we test drives before going into production but for this box we didn't). So SMART error - There's a spare drive that I've inserted so let's tell Solaris to move the mirror to the new drive - Sounds easy, right? zpool replace ... starts off fine.

Now let's go back to getting drives to hot plug, the move / remirror above is running in the background (it's a 1TB SATA drive, so takes a while). swap this, that, toss in a kitchen sink (hard to get it to fit in the 3.5" bay, but we tried).. So far we're just spinning in circles...

Suddenly the NEW 1tb SATA drive volume has been removed according to the Adaptec Storage Manager - I look at the drive light and low and behold it's flashing still for activity. iostat shows the drive writing about 60 megabytes per second - Imprsesive for a drive that's been removed, I didn't realize wireless SATA was available

Tell Adaptec Storage Manager to rescan the bus, eventually the drive shows back up. The entire time it's writing according to the OS fine, never removed.

So at this point I'm thinking more Adaptec bugs, let's call them. Of course there's no weekend support. Frustrated with the entire set of problems I decide to see what LSI's support policy is - It's not too much trouble to replcae the controller (ok, it WOULD be a lot of trouble, but I'm really unhappy with Adaptec so far).

Check LSI's web site: Sounds good, 24/7/365 support, just what I wanted. Cnosider ordering LSI controller (already have seen fairly close model and was considering it during earlier problems). Just for fun let's call the number and see if I had this problem with a LSI controller would I have support today?

Let's examine the wording on their site:

Access support for Host Bus Adapters (HBAs), Internal RAID (MegaRAID), 
and External RAID (Engenio) products, 
24 hours a day, 365 days a year.

So here we go: Call LSI.. Answered almost immediately by a human (impressive so far). Explain I'd like to find out if I'd have support for this issue if I had a LSI controller in the box right now.. What's the answer: No, there's not support for that on weekends. WHAT? The web site says.. I'll let you fill in the rest of the conversation. The person was extremely professional, in Kansas (USA), and very friendly about it. He also said it was another group that handled the controller, but that it would be Tuesday before I'd have an answer unless i wanted to do their premium hourly at something like $300/hour. Let's see, 24/7 support, it's Sunday around midnight on a holiday weekend, expect a call back in 60 hours or so.. Or maybe the more famous support answer: You are currently #1 in the queue, the expected wait time is approximately 3600 minutes. Lovely 24/7 support. I'll conclude with he was very pleasant and if I had a server down with their controller I may be fuming, but honestly before you discredit them TOO much how many people would ADMIT that if you didn't have a serial # you were calling about? Probably less than you think.

So let's try Adaptec's paid support. Call, get a big long recording, finally a human. I'm 95% sure she said $149 per call - My main question at this point (since my production drives are working) is: "If this is a BUG do I pay?" I guess I'm used to Microsoft's professionalism on finding legit bugs with their products - You don't pay. If you did something then yes, pay you must :-) But if you discover a problem they're quite good about refunding their charges. I ask if this is Adaptec's policy - She isn't sure, she tells me she can page someone from that group who will take the call, the credit card info, and let me know if it's chargable. So far so good.. 90 minute call back window (not too good for a potentially production storage device to be down 90 minutes, but I'll live).

Pleasant surprise, about the time it takes me to log into my blog and start to type my cell rings, it's Adaptec support. I explain briefly the problem and ask if it's a bug what's the policy? You pay, you pay regardless of what is found, you pay if your controller needs to be RMAd because it was a defect. You simply pay.

Now let me say, how much you pay was reasonable - However given that, like I said above, we're already 20+ hours of debugging problems with their controller not having the capabilities they claim I'm not real happy to pay to get it to do the ONLY thing it's supposed to do - Handle drives being hot plugged.

So here I am, about 4 hours into adding 6 drives to a box and thinking we probably need a different controller vendor than Adaptec.

In all this I've learned a LOT more about SAS, I realized before it was far superior to SATA especially (and SCSI too), however I didn't realize how much. It's easily expandable by dropping in another enclosure, which Adaptec says is hot pluggable - the whole enclosure full of drives they say can be hot plugged.. Of course can you trust that? The drives are supposed to be hot pluggable and THAT simple task doesn't work, can you imagine dropping in 24 drives with one cable being connected?

I even more strongly agree with Sun's philosophy that the controller just needs to be stupid and export the disks to the operating system as JBODs.

I need to also blog sometime about what happens if you try to pull a failed drive from the controller - It says no, you can't do that, it MAY have data on it. I'll get around to writing that up sometime - but basically since the Adaptec won't export the drives to the OS we had to create simple volumes for each drive - The simple volume though we can't remove to take out a drive, argh.

The joy of being on the bleeding edge of technology.

---

Permalink (referenced by: 13 posts / references: 0 posts) 0 comments [view comments]

Let me preface with I've been loyal to Microsoft technologies for almost 20 years, but on some areas they are loosing touch with at least my needs if not plenty of others.

So I decided to go check out what Sun had to offer, I decided Friday night to get a airline ticket - now Monday I'm in San Fransisco, California (I'm from the DC area, other end of the US).. Pretty spure of the moment, but their ZFS file system and some of the virtualization technologies interested me quite a bit and not knowing really anything about Sun I figured I'd learn something..

What did I think? I have to say, they're impressive, until you try to IMPLEMENT their "available" technologies.

Let's take some examples:

• VirtualBox / Xen / whatever they call it (since they don't really seem to know themselves).. Well, the presenter says the binary download is identical to the open source - I saw a few things I'd want to change to even consider using it on our network, so I start downloading at the presentation.. The open source edition is crippled and lacks several features - They JUST said that's NOT the case - Which is it? Crippled or full product?

• ZFS timeouts on disk i/o - I tried installing Solaris on a somewhat flaky box, I ended up locking up the system with no errors during an install; and during running ZFS I could corrupt the pool. Ask a question and I get the typical "good" speakers response of take it offline afterwards - That's a sign of one of two things: 1) The speaker really wants to talk technical, 2) You've caught him not knowing how to respond and he wants to blow you off. Talk after - basically the discussion is speaker has to go to another presentation, adios.

So I've pretty well wasted a day on technologies that sound great; Sun's got some FANTASTIC ideas and technologies they're working on - And I even understand getting people excited before they are products / ready - But don't tell me you have something that you don't, especially when I take a day off work, fly across the country, just to find the things I wanted to hear about don't really exist in a useable fashion.

So I'm frustrated, but I'm also going to end with ZFS makes Solaris *THE* file storage operating system to use. We'll be dumping a lot of hours into ZFS, even if it means coding it to make it meet our needs. The base is there, although perhaps buggy - Windows has nothing to compare today, and I know of no rumors of anything close. The management ease and expandability of ZFS makes it worth some headache; if we don't put the headache into it we'll just have the headache managing our current SAN and NAS environments.

Hopefully a few people read this and consider it, in summary:

• Microsoft: I feel you are falling behind in file storage and volume management IMHO. Look out, Sun may just make a comeback.

• Sun: I hear plenty of bloggers say you don't know how to manage/market what you have, I've just witnessed that first hand. I also don't know even if you released everything to open source today how you'd profit from your investment, not good for investors - Look out.

• Admins: Times are changing in storage, a revolution started a while ago with NAS and SAN, but the bottom is dropping out of the pricing for those technologies and many new ones are worth looking at.

Live from San Fransisco a few more minute before I leave for the airport being nobody from Sun seems to want to discuss the issues I have with rolling out their technologies, or their hardware questions I hoped to address while here.

---

Permalink (referenced by: 40 posts / references: 0 posts) 0 comments [view comments]

If you're like some others (myself included) you may get a Vista computer which can't open media files, for example Windows Media Videos. I don't have the details on what causes this, but I do have a work around:

You can watch a quick video at bitshop_asp_videos

The steps are:

• Launch IE
• Go to the Tools Menu, then Internet Options
• Click "Programs"
• Click "Set Programs"
• Click "Set program access and computer defaults"
• NOTE: Allow the action which requires permission
• Click "Custom"
• Scroll down to "Choose a default media player:"
• Select "Windows Media Player"
• Click OK, X, etc. to close the windows..

This should work around the problem.

Happy media watching :-)

---

Permalink (referenced by: 126 posts / references: 0 posts) 0 comments [view comments]

Unfortunately I didn't have time to blog much last night or this morning, but I'm at MIX06 and interested in meeting anyone while here who is interested in meeting. I'm also not going back until Thursday evening (extra day) so have some time to meet then.

So what is MIX06? Well, it's most basically showing what's coming on the web in what many refer to as Web 2.0 - AJAX, Microsoft Atlas, Windows Presentation Foundation, and much MUCH more.

Everything so far is great, It's very obvious the browser wars are back in full force, as well as so much is changing on the desktop we may as well say a desktop war is being started. LOTS of new featuers in IE7, some are a bit of a hassle (like pop ups warning me a site may be unsafe if it's based on an IP instead of a domain name), but overall I'd recommend IE 7 be quickly adopted. The UI is much cleaner.

I'll be blogging more tonight (time permitting) and trying to do a few videos about these technologies here over the next few days.

Steve Radich
BitShop, Inc.

---

Permalink (referenced by: 113 posts / references: 0 posts) 0 comments [view comments]

You may receive the following error if your permissions on your X.509 certificate are wrong:

System.Security.Cryptography.CryptographicException: Cryptography_CSP_NoPrivateKey

If this happens then you want to launch the X.509 Certificate Tool, select your certificate and press View Private Key File Properties - Then the security tab.

At this point the exact permissions depend on your app pool, impersonation, etc.

Most likely if you don't know what to set then you can try the local ASPNET user and the local NETWORK SERVICE user. They will need access to this file for the certificate to be used.

Hope this helps.

You can always contact www.BitShop.com for your hosting and WSE hosting needs and we'll make sure you have this set up right and working reliably, from a single shared server running a WSE based app to a cluster of dedicated web services servers.

---

Permalink (referenced by: 22 posts / references: 0 posts) 0 comments [view comments]

Image Storage: Is SQL Server better or static files on the web server?

If you, like many web sites, are managing lots of images for a site then you may consider moving the images into SQL Server for easier management.

There are plenty of reasons to do this, and plenty of reasons not to. If you spent enough time searching old archives you'd find several old discussions from about 6-8 years ago on this topic that I wrote (back when I wrote more regularly). My opinion then was never do it due to the load on sql server and the extra cpu load on the web server. My opinion has changed though, due to the management cost of disk based storage more than anything - However let me explain a few things you should know about this to assist in your decision on the topic.

First, I'll start with the bad (because I want to end on a positive note :-)

• Disk load on SQL Server: Scaling a SQL Server is FAR pricier than scaling the web farm. Replication costs, dbas, etc. make the cost higher. A Web Server disk you can (most simply) just xcopy everything to another server and have 2 boxes, or use a virtual directory mapped to another drive (even a network drive is possible). NOTE: There are many instances this may not be so simple to xcopy on the web server.
• CPU load on SQL Server: There's not a lot of CPU time spent to retrieve the image from disk, but it's some, and your SQL server may not have any to spare, especially as your business grows. My concern is more if you are out of CPU cycles do you want it wasting time on your images? Usually this isn’t much concern though.
• Fragmentation - Realize you can have multiple "types" of fragmentation in your SQL database.
• NTFS Fragmentation: First, the underlying files that SQL stores it's data in (.MDF and .LDF files) can get fragmented by the operating system - These files CAN"T be defragmented since they are in use unless you are going to take the database offline and defragment.
• SQL Pages can get fragmented: If you are doing a lot of updates, especially on images, then you will likely fragment your pages pretty badly due to the size of the images changing continually.
• SQL Index Pages Fragmentation: The index pages can fragment (this happens on everything, not just images).
• The good news: The SQL fragmentation your dba should have a maintenance job scheduled to reorganize the pages weekly or so which will clean this up.
• Realize that if you never delete the images you should have far less fragmentation in the database, if you rarely delete, or want historical data, then consider the cost of disk space versus the overhead of defragmenting (or searching if fragmented).
• Web Server potential downsides of storing in SQL:
• CPU Load: Processing the ADO.Net commands and streaming it to the browser has some associated overhead.
• Network Load: If your network connection to SQL is any concern then realize you are doubling your network i/o by having to read from SQL then output to the client. Generally this won’t be a concern.
• Caching: Generally file cache on the web server is going to be more efficient than anything you can do in the code, including the new caching in ado.net in ASP.Net 2.0 (Whitbey).

Advantages to reading from files on the web server:

• Static file I/O is very fast in IIS – It’s FAR less CPU overhead than reading the ADO.Net data and streaming it to the client. IIS can stream static files off the disk VERY quickly with very low CPU load.
• Caching is efficient, see above.
• Backup is fairly simple, scaling is simple.
• The KIS principle (Keep it Simple) - Especially if you are doing ths for a single server deployment (i.e. not a web farm).

The downside of keeping things on the web server:

• Generally administering anything in a sql database is easier than files
• Easier multi-user administrator scenarios (don’t have to update file permissions / ftp rights, you probably already have some kind of admin interface to take care of the image meta data)
• Easier granular control / integration with existing administrative interfaces
• Web farms (multiple web front end servers) can pull from one database server easier than you can keep files replicated
• Easy ability to cross reference images to some other table, since most likely they relate (either images to a page, customer, or whatever the relationship is for your business). Ability to find orphaned images (images not used anywhere) if there is no relationship.
• Ability to enforce foreign key relationships to avoid accidentally deleting an image that is still referenced anywhere.

Best practices if you are going to put your images in SQL Server (I'll expand on this later, this is just the first few things that come to mind):

• Store image meta data in either your primary database or a separate meta-images database. Why? Your backup strategy for this may differ from your actual images, since it will be smaller and, depending on your business, may be updated more often (especially if you have counters on the images, or frequently updating comments / descriptions). The backup of this database should be small and very quick.
• Put the images in their own database so this database could be moved without much trouble to another disk subsystem or another SQL Server. By putting it in its own database you can also back it up separately, which, as the database grows you may want to have a different backup schedule than your meta database (not as often to avoid excess disk i/o reading the whole db).
• Your database files should start (be created) at the size you expect them to grow to - This will allow, assuming NTFS has the unfragmented space available, the .MDF file to be continuous on the file system. If possible you may want to do a defragment on the disk partition of the SQL Server before creating the database - Since most files are in use you'll generally have a pretty fast defrag - Obviously this will be a peformance hit on the server so defrag with some caution.
• If your images are a certain size, let's say 64k or smaller (an actual business requirement that they be this) then you could consider using several 8000 byte fixed length columns to store the image and breaking it into parts in your ASP.Net code. Of course if you do this and the image is larger it will fail to be stored. This will allow replacement images to occupy the same amount of disk space, reducing fragmentation. If you do this then consider wasting extra disk space and making most of your fields fixed sizes so the entire row is a fixed size. Realize there is a trade off, at some point, on fragmentation versus reading data you didn’t need from the disk. If you do this then also consider sql page sizes, index space utilization, etc. This isn’t a tip a beginner should do, if you don’t understand it then just don’t do it.

I hope the above information helps. I will say that on most smaller sites that I now design I use storage of images inside the database – I find it simply more convenient than having to worry about file permissions for uploading images.

---

Permalink (referenced by: 149 posts / references: 0 posts) 1 comments [view comments]

As you can see from my blog posts, I usually write up full pages on the site instead of blog entries, however today I decided to do a quick note since I don't have time to write this up in full.

I was troubleshooting a DHCP server that wouldn't assign network addresses to clients. It had all the "usual" things already checked and working fine so this took a while to find.

DHCP uses UDP traffic to send reqeusts, we ended up finding the problem by using a packet sniffer to see exactly what was happening. We found the DHCP requests came in, but the server never sent any traffic back.

First, let me go over the "usual" things to check:

1) The scope has to be activated, right click the scope and select activate.
2) Windows DHCP servers (since Windows 2000) require the domain to authorize them to start processing requests. This is done because a user could accidentally start a dhcp server on a computer and wreak quite a bit of havok on the network (clients would get wrong IPs, wouldn't find servers, etc). To authorize the server you simply right click the server in the MMC and it's an option there.
3) Check the event log for any errors - At this point there should be an event saying the server is authorized and starting to process client requests.
4) Check the server is on the same LAN segment, DHCP is a broadcast, and short of having DHCP bridged or DHCP Relay Clients (windows component) or set your routers to handle the broadcasts.
5) Check the IPSec policy on the server and client. This is what I ran across - The server's ipsec rules blocked ports 67 and 68 on UDP (which are what are used by DHCP / BOOTP protocol). This resulted in the packets being dropped that were outbound from the server before they made it onto the NIC / Network. The result was the server APPEARED to never respond to DHCP requests. To resolve the domain security policy needed updating to remove those blocks.

Steve Radich – http://www.ASPDeveloper.Net – www.VirtualServerFAQ.com
BitShop, Inc. - http://www.bitshop.com - Managed Servers, Colocation,
Troubleshooting, Development and Perf Monitoring (C#,VB,ASP,SQL)
Dedicated Windows 2003 VMs Starting at $49/month

---

Permalink (referenced by: 338 posts / references: 0 posts) 0 comments [view comments]

I decided to throw together a quick walk through of using ASP.Net 2.0 Master Pages as a movie, instead of a text document to see how people like it.

It’s fairly basic, but I think is very easy for someone with a little asp.net experience to follow.

It is done in Visual Studio 2005 Web Edition Express, which for those of you that aren’t playing with ASP 2.0 that’s a free beta download.

If you aren’t already working with ASP.Net 2.0 then I would recommend you install it in a virtual PC / Virtual server to isolate it from your work computer. The betas are usually fairly stable but .net 2.0 has a lot of components and I’d really recommend having a second box / virtual box for it even if you usually venture into betas on your production development servers / desktops.

http://www.aspdeveloper.net/tiki-index.php?page=bitshop_asp_videos

For those of you that haven’t used Virtual PC or Virtual Server you may want to view some of our coverage of that product (also covered on http://www.aspdeveloper.net). We have sveeral presentation dates available for presenting to users groups on that topic if you have an interest in learning more about it in a relaxed group setting. For those interested in more "formal" training we have "quick starts" which we are putting together which are customized to your exact requirements, not a boiler plate training seminar. These are not just customized to your requirements, but we actually work the servers you are considering migrating into the course, instead of talking about a generic example we redo all examples using your environment before the class starts. These will prove extremely cost effective for those who can take advantage of virtualization in their data centers.

Please let me know what you think of this tutorial, good, bad, boring, loved it, etc. Also we’re taking requests for other topics to write or video about.

I look forward to your feedback,

Steve Radich - http://www.aspdeveloper.net - Virtual Server FAQ - ASP.Net Information - FREE
BitShop, Inc. - http://www.bitshop.com - Managed Servers, Colocation, .Net Development (C#,VB,ASP)
FREE ASP.NET 2.0 BETA HOSTING ACCOUNTS AVAILABLE

---

Permalink (referenced by: 0 posts / references: 0 posts) 0 comments [view comments]

By default SMTPSVC accepts authentication attempts to relay, so a hacker could attempt to compromise your domain accounts via SMTP attacks on your web server (if you have SMTP enabled on the web server).

The error in the log may look like:

Type: Warning
User: N/A
Source: SMTPSVC
Category: None
Event ID: 100

Description:

The server was unable to logon the Windows NT account 'connect' due to the following error: Logon failure: unknown user name or bad password. The data is the error code.
For additional information specific to this message please visit the Microsoft Online Support site located at: http://www.microsoft.com/contentredirect.asp.

Data: 0000: 2e 05 00 00 (Bytes view)
Data: 0000: 0000052e (Word view)

In order to disable the ability for users to try to authenticate via SMTPSVC you can:

1) Launch IIS Manager (Start-Programs-Administrative Tools-IIS Manager)
2) If necessary connect to the computer with SMTPSVC on it
3) Go to Default SMTP Virtual Server - Right Click - Properties
4) Under "Access" Click "Authentication"
5) Uncheck the boxes Basic Authentication, Windows Security Package
6) Hit OK, and save..

Alternatively you could just disable SMTPSVC if you don't have a requirement of sending mail from the server through the local SMTP Service.

---

Permalink (referenced by: 177 posts / references: 0 posts) 0 comments [view comments]

ASP.Net Beta 2 is now available for download from Microsoft, including Visual Studio 2005 and the Express versions.

BitShop (the owner of this site) will be providing FREE hosting for ASP.Net Beta 2 developers, see http://www.bitshop.com for details.

We'll be updating our tutorials and adding more tutorials to our site for your reference.

Please send me some feedback about what you like, or dislike, about this site to suggest@aspdeveloper.net

Thank you for visiting,

---

Permalink (referenced by: 0 posts / references: 0 posts) 0 comments [view comments]

I'm sure most people reading my blog have either seen lists of ports used by various Windows services, or have their own lists. However today in searching for something for Exchange Server I ran into a Microsoft article listing ports used by Windows 2003.

This list is 'straight from the horse's mouth' is why I mention it. It also descrbies a little some of the reasons for random ports in Exchange, etc.

I found this a better description than most others I've seen and thought I'd share it.

Service overview and network port requirements for the Windows Server system

---

Permalink (referenced by: 118 posts / references: 0 posts) 0 comments [view comments]

This is a shorter than normal post as I want to document an error I ran into today and don't really have time to fully write it up as detailed as I usually would.

If anyone else runs into this and wants me to fill in more info let me know and I will.

A very simple IPSec policy change (allowing vpn clients) needed to be rolled out to a group in active directory - This OU was assigned a new GPO, in that GPO was a new IPSec policy. This IPSec policy was there, seemed to work fine, etc.

Some servers however weren't getting the new IPSec policy, which didn't really make sense. It was there, assigned correctly, etc.

Some of the tests that were run included:
* gpresult - shows everything ok
* gpotool - shows policies are ok

I then looked under C:\WINDOWS\Debug\UserMode\gptext.log and found the error:


WriteDirectoryPolicyToWMI failed: 0x80041005

0x80041005 means WBEM_E_TYPE_MISMATCH / Type mismatch occurred.

I didn't find anything particularly "wrong" with the IPsec policy, however I edited each setting, changed the description (name) and saved it. That cleared up the problem. Apparently something was stored corrupted in AD.

I also ran across this utility, but didn't try it, Sys Pro - http://www.sysprosoft.com/index.php?ref=activedir&f=policyreporter.shtml - It just cleans up some of the logs to be more readable and gives you a simple "turn on/off" debugging. Like I said, I didn't use this.

Since there are no google / altavista / yahoo / msdn matches for this error I figured I'd post this in the hopes someone someday finds it helpful to them.

---

Permalink (referenced by: 96 posts / references: 0 posts) 0 comments [view comments]

I imagine many of you that are reading this, like me, either need, or prefer, to be online 24/7 without lugging around a large laptop.

Every year or so I compare phones, PDAs, wireless internet access cards, etc. and have found Cingular and Verizon to be my two current choices of best devices and data connections. I have chosen Verizon for the extra speed, and due to GSM networks not being available in some places I travel..

Let me start with what is a GREAT device, but is horribly slow. T-Mobile has treated me well for the past 1.5 years with a Sierra Wireless Aircard - Speed is slow (< 56k), but service has been reliable and their mapping of where they have coverage and don't have coverage has been accurate. By adding an external antenna I have coverage in many places they claim no service to be available.

T-Mobile now has the HP Ipaq h6315 Pocket PC phone/PDA available. This has a built in camera, Windows Mobile 2003 OS, Wi-Fi, Bluetooth, and more. The most impressive feature is it automatically switches to 802.11 Wi-Fi if available instead of using GSM/GPRS. The downside of this device is that it's not upgradeable to EDGE (or any other faster network T-Mobile plans to roll out in the future). This means you are buying something that is 2G wireless coverage, and no way to upgrade it in the future. This device however it quite impressive if you can tolerate the slower speed. The price is also slightly lower than some of the other devices, and the service plan is FAR lower than some of the competitors.

T-Mobile Sidekick II - If you ONLY need email (POP3 or IMAP), IM, and SSH/Telnet access (no remote desktop) then you may love this little device. Not only is it priced good but service is relatively cheap for it. The keyboard is nice and it simply does everything well that it can do. It's not very expandable however, which is a deal breaker for me.

palmOne Treo - I've got nothing bad to say about this device, other than it's not Windows (since it is Palm). Some of my main requirements are VPN and Remote Desktop, last time I actually tried a Palm based device these things were add ons and there were some questions on the terminal server access licenses with these devices. I haven't looked back at Palm based devices since this, however they are quite good.

RIM / Research In Motion BlackBerry devices: 7100t and others: BlackBerry's are quite popular, sometimes I wonder why. They give you a LOT of PDA type features, and connect great to mail servers, etc. However they aren't quite as general purpose as a Palm or a Pocket PC PDA - Therefore I strongly prefer the other devices. This won't be the case for everyone, and if these do what you want then you may want to look further at them. Most of the data plans for BlackBerrys are cheaper and/or unlimited, saving you from having to watch how many megabytes you transfer. I don't believe you can tether your laptop to any blackberry for access to the net, for some people this will be a deal breaker.

Sprint Audiovox 6600 - The Sprint offering is the same device as Verizon, the reason I didn't choose Sprint is their lack of coverage in a few areas I frequently travel - One of those areas Verizon has coverage - Otherwise Sprint service will soon be basically the same as Verizon (currently Sprint isn't EVDO so their data performance is slower).

Verizon Audiovox 6600 - Verizon offers their EVDO network in many major cities, with nationwide access promised. How fast is EVDO / 3G coverage? It varries with the number of people on the tower, distance from tower, signal strength, etc. However I've seen pretty consistently 400+k/second, with many times 700k+/second transfers. Yes, that's as fast as many people's wired DSL lines, yet I can roam around anywhere with it. Or can I? Not West Virginia, not Western Maryland, not most of North Carolina, and various other places. Why? Verizon and US Cellular can't come to an agreement for data roaming - This is my single biggest gripe with Verizon's service. They let me roam for voice calls, but I have to carry another data device to have coverage in those areas, so I keep my Nextel i90 and laptop cable kit ready if I plan to travel to those areas .

The Audiovox 6600 is my device of choice, it offers me Remote Desktop client, PPTP/IPSec VPN client, POP3/IMAP/Exchange server email accounts, SD card for expansion (camera, wifi, or just storage of MP3s), and much more. It covers everything the other Windows Mobile PDAs cover, although it does them all average to slightly below average. As a PDA you won't love it, you'll "deal with it", but combined with the 3G EVDO data coverage you won't leave home without it :-)

Since it is a Windows device I can easily write code in the .Net Compact Framework and handle tasks that I need written - All fairly easily. If enough people request it I may write up a tutorial on this, otherwise I'll assume you don't need to write custom code for your Pocket PC.

What will all this cost you? Well the Audiovox 6600 is about $650, the unlimited data plan is $45 with a Verizon voice plan, or $49 without. Personally, since I have to carry another phone when I travel to WV and NC I don't have Verizon voice - But I get a phone number, can be called, and can place calls - It's just billed at $0.25 per minute, which isn't a bad rate.

The device can be tethered to your laptop (i.e. be used as a modem for your laptop) via bluetooth (fairly painless) or via the USB cable (a little more painful). This gives you high speed access from your laptop wihle on the road.

Some of you may ask what about VoIP on these devices with unlimited data? I'd love to tell you it works, I've even seen a Verizon independant rep selling a packaged unlimited data with unlimited VoIP using your data plan - The reality is it doesn't work well. I've tried it with a SD 802.11 wireless card in the phone, and EVDO turned off, and the results were horrible. This is a bit disappointing as Nextel, Sprint, Verizon, and T-Mobile all have poor coverage inside our data center, and it would be handy to be able to walk around with a wifi voip phone tied into our office phones, however that currently doesn't work.

I hope you find this information both useful and helpful, please feel free to comment or email me with questions.

Please remember, don't IM and drive at the same time with your new device :-)

---

Permalink (referenced by: 0 posts / references: 0 posts) 1 comments [view comments]

I will be writing a tutorial in the near future about using masterpages to have a common layout for your pages.

Probably most of you are familiar with include files for your header/footer, think of this as the opposite, the content pages are more the "include" to the "template" page. It doesn't work like that, but that seems to be the right "mind set" for how it works.

I only have notes on my tutorial so far, hoewver I'm experiencing an interesting error and wanted to post it to see if anyone happens to have an idea how to correct (if I'm doing something wrong), or incase others are having a similar problem I'll share my (ugly) work around.

The error output is as follows:

Compilation Error
Description: An error occurred during the compilation of a resource required to service this request. Please review the following specific error details and modify your source code appropriately.

Compiler Error Message: BC30560: 'default_aspx' is ambiguous in the namespace 'ASP'.

Source Error:



Line 45: Dim dependencies() As String
Line 46: AppRelativeVirtualPath = "~/pages/default.aspx"
Line 47: If (Global.ASP.default_aspx.__initialized = false) Then
Line 48: dependencies = New String(2) {}
Line 49: dependencies(0) = "~/pages/default.aspx"


Source File: C:\WINNT\Microsoft.NET\Framework\v2.0.40607\Temporary ASP.NET Files\masterpages\622bc8e8\94d050bc\k71ketcz.0.vb Line: 47

The detailed compiler output is as follows:

Microsoft (R) Visual Basic .NET Compiler version 8.0.40607.42
for Microsoft (R) .NET Framework version 2.0.40607.42
Copyright (C) Microsoft Corporation 1987-2003. All rights reserved.

C:\WINNT\Microsoft.NET\Framework\v2.0.40607\Temporary ASP.NET Files\masterpages\622bc8e8\94d050bc\k71ketcz.0.vb(47) : error BC30560: 'default_aspx' is ambiguous in the namespace 'ASP'.

If (Global.ASP.default_aspx.__initialized = false) Then
\WINNT\Microsoft.NET\Framework\v2.0.40607\Temporary ASP.NET Files\masterpages\622bc8e8\94d050bc\k71ketcz.0.vb(52) : error BC30560: 'default_aspx' is ambiguous in the namespace 'ASP'.

Global.ASP.default_aspx.__fileDependencies = Me.GetWrappedFileDependencies(dependencies)
\WINNT\Microsoft.NET\Framework\v2.0.40607\Temporary ASP.NET Files\masterpages\622bc8e8\94d050bc\k71ketcz.0.vb(53) : error BC30560: 'default_aspx' is ambiguous in the namespace 'ASP'.

Global.ASP.default_aspx.__initialized = true
\WINNT\Microsoft.NET\Framework\v2.0.40607\Temporary ASP.NET Files\masterpages\622bc8e8\94d050bc\k71ketcz.0.vb(100) : error BC30560: 'default_aspx' is ambiguous in the namespace 'ASP'.

Private Sub BuildControlTree(ByVal ctrl As default_aspx)
\WINNT\Microsoft.NET\Framework\v2.0.40607\Temporary ASP.NET Files\masterpages\622bc8e8\94d050bc\k71ketcz.0.vb(132) : error BC30560: 'default_aspx' is ambiguous in the namespace 'ASP'.

Me.AddWrappedFileDependencies(Global.ASP.default_aspx.__fileDependencies)
\WINNT\Microsoft.NET\Framework\v2.0.40607\Temporary ASP.NET Files\masterpages\622bc8e8\94d050bc\k71ketcz.1.vb(26) : error BC30560: 'default_aspx' is ambiguous in the namespace 'ASP'.

Return New ASP.default_aspx
~~~~~~~~~~~~~~

---

The interesting thing however is:

If I rename the file to ANYTHING other than default.aspx it works fine (with on other changes).

There are no other extensions / user controls involved in this, just a simple default.aspx file.

Anyone with a tip as to what's going on please email me / leave a comment.

---

Permalink (referenced by: 81 posts / references: 0 posts) 2 comments [view comments]

This is off topic from what I usually write about, however I wanted to share a few thoughts and tools from roughly a year of mobile internet access via an IPAQ with a PCMCIA sleeve + T-Mobile wireless internet access card (Sierra aircard 750).

The reason I chose this solution, instead of some of the other “smaller” (CF) solutions is that I can pull the PCMCIA card out and slip it in my laptop easily if necessary. In fact I more often than not prefer to use my laptop than the PocketPC for internet access.

This solution works decent, not great, but decent. I have however had repeated problems with the connect via PDA. Using the same card in a laptop, at basically the same time (i.e. pull it out, put it in laptop) it works fine. The PDA software seems buggy from Sierra and I gave up for several months on using it. I'm back to using it (and considering one of the new Verizon Wireless Audiovox 6600s, which I'll write here in the future if I try one out, the promise of high speed 500k+ wireless, instead of 40-50k via T-Mobile is very promising).

Even with the problems and headaches nothing really beats the convenience of having internet access in the "palm" of your hand. Not a WAP browser, but a fairly powerful scaled down IE.

What I use this for isn’t browsing but is instead attempting to manage servers via Remote Desktop both when walking around our data center, and when I am outside of our office and need to check out a question / problem. Generally I've found I don't use the IPAQ though (interface is too clumsy for remote desktop) and instead walk to wherever my laptop is and use it. I'm trying to force myself to change this habit and use the IPAQ more, I believe once in the habit the solution will be much more natural and is definitely more convenient.

What I find is the VPN, and mapping server names (DNS), etc. isn't very easy with the built in tools. Instead I find myself wondering if it's connecting via the VPN, or via the internet, having DNS issues sometimes (it seems to decide to use the wrong DNS servers sometimes, i.e. the sierra wireless DNS settings instead of the VPN DNS servers), etc. I also find the VPN disconnecting when I want it to stay connected.

To work around this I've found the following tools quite useful:

Pocket Hosts
http://www.zimac.de/cestuff.htm

VXUtils - Traceroute
http://www.cam.com/vxutil_pers.html

I have no relation to these companies, I am simply providing the links to them for your convenience.

Pocket Hosts will allow you to set up static DNS mappings in your Pocket PC. This assists in working around any DNS problems. By using a short name (not a fully qualified domain name with .companyname.com) the PocketPC uses the VPN always. This is quite helpful.

VSUtils includes a traceroute utility which allows me to trace to a host and make sure the VPN is being used instead of the internet.

One thing I must say I *HATE* about the VPN in the PocketPC is it seems to disconnect as soon as I go to a page that doesn't require the VPN. So for example if I'm trying to work on something, and while doing that want to pop over to IE and search something, I end up disconnected from the VPN.

Some of the other features I find handy are that I can sync my Outlook to Exchange 2003 without connecting to the desktop computer. This works great for email, however the tasks aren't synced (It seems very dumb that they aren't as the tasks are also stored in Exchange). Personally I use the tasks probably as much as I use email (due to using CRM systems for most work related email).

I hope you find this article helpful if you have a PocketPC, whether it's an IPAQ, Audiovox 6600, T-Mobile h6315, or any other device. If you like this subject please comment, or write to suggest@aspdeveloper.net, so we know this is of interest and I will write more about my experience using PocketPC with a full time internet connection.

Happy Surfing!

---

Permalink (referenced by: 0 posts / references: 0 posts) 0 comments [view comments]

As many exchange admins do, I from time to time use exmerge for quick / simple backup routines of a mailbox.

I was trying to export a mailbox that had a log entry going to it every few seconds and had a total of 120k messages in it. I decided the easiest way to let someone deal with deleting these was to stop the new messages, export them to a pst, and let them delete the ones that should be deleted on their local box. For some reason it seemed very slow deleting on exchange (possibly exchange doesn't deal well with 120k messages in a single folder, inbox).

The exmerge completed fine, just with some warnings that went to the log file.

The warnings were: Error copying message with subject ... in folder '\Inbox'. (MAPI_W_PARTIAL_COMPLETION)

The .PST file had 16383 or 16384 (I don't recall which) items in the inbox, however it didn't have the 120,000 I expected.

This seems to be a limit either in the PST file format, on in exmerge.

I haven't had time yet to confirm if this is a known issue or a bug, however some searches didn't turn up any known issues.

Steve Radich - Troubleshooting, Colocation, Development and more..
BitShop, Inc. - http://www.bitshop.com

---

Permalink (referenced by: 0 posts / references: 0 posts) 0 comments [view comments]

Let me first say sorry, this is going to be a very terse post as I want to get it posted and will later try to either answer comments, or modify this post with more details.

If you are using WSE and are getting an error of Cannot find X.509 certificate it could be you have installed the certificate in the wrong certificate store.

Windows 2000/2003 has various places you can store your certificate, usually the certificates would go under

To manage certificates launch mmc - then Console - Add/Remove PlugIns - Click Add - Select Certificates - Click Add - Select My Computer - Local Computer - Click the close button and the OK button to get back to the MMC console.

You will now see varoius certificate stores, the "Personal" folder is usually where you will add certificates you buy / test with.

If you have added the certificate somewhere else, other than the Personal / Certificates folder then you can drop and drop the certificate to the personal certificates folder.

The personal certificates are reffered to as "MyStore" in the WSE code.

---

Permalink (referenced by: 0 posts / references: 0 posts) 0 comments [view comments]

I was reading a question from a user who wasn't getting an IP address inside a virtual server guest operating system. I wanted to give a few pieces of advice to assist in troubleshooting this.

Usually if you don't get an IP address you'll end up with either a) 169.254.###.### or b) your default IP address defined in NIC properties, TCP/IP, alternate configuration.

You can check a few things to see why you didn't get an IP address:

1) Go to a command prompt and type "ipconfig" - Verify your IP address isn't one you expect.
2) Try "ipconfig /renew" which tells Windows to renew it's IP from DHCP
See if you get an error like "unable to contact your DHCP server. Request has timed out"
3) Assuming you got an error you may want to make sure your NIC is working in the guest operating system. Usually your network admin (or wireless router) has some reserved IP addresses that aren't auto assigned by DHCP. You could try manually putting one of those on your NIC.
For example, let's say your wireless router (or network admin) uses 192.168.0.* as the internal subnet, and DHCP uses 100-199.
You could assign 192.168.0.230 (just picking a random number outside of 100-199) to your machine statically
Set the default gateway and subnet mask
Now run ipconfig and you should see the correct IP Address
4) Once you have an IP on there statically can you ping the router? Can you ping another server? And, if Virtual Server Guest, can you ping the Virutal Server machine's IP?

If you can ping everything then the problem is likely with the DHCP server - you may want to make sure you aren't out of DHCP addresses in the pool, or have any kind of security settings preventing this new client from getting an IP.

If you can't ping anything then is your nic supported?

If you are running Virtual Server and the NIC works in the host, but not in the guest, then you may have a NIC that isn't compatible. For example USB nics may not work since the USB driver starts after the virtual server service. I haven't experienced this myself, but you may be able to make the Virtual Server service rely on the USB Nic service, however you should really have a built in nic in whatever you are doing.

I hope this helps you troubleshoot, plese feel free to post any comments here for additional questions and I, or one of my assistants, will try to answer you.

Steve Radich
BitShop, Inc. - Spam Filtering, Colocation, Troubleshooting - http://www.bitshop.com

---

Permalink (referenced by: 8 posts / references: 0 posts) 0 comments [view comments]

I'm writing this entry in reference to a newsgroup question about the error "Virtual Server could not start the VMRC server. Another server may be using the selected TCP/IP port."

This information applies to IIS, Virtual Server, as well as ANY other progam using a port on the server.

To display the listening ports on a server you can do (From a command prompt):
netstat -ano | find ""

For example the newsgroup post is about virtual server which uses port 5900.  You could do:
netstat -ano | find "5900"
This would return something like:
TCP    0.0.0.0:5900           0.0.0.0:0              LISTENING       320

You can then launch Task Manager and find that task ID. If Task Manager doesn't show the process id column then you can (in task manager) go to View - Select Columns - PID (Process Identifier).

Once you have identified which process is bound to the port you can easily figure out if you have something else running which conflicts or what is happening.

Hopefully this helps you easily resolve the port conflict!

---

Permalink (referenced by: 200 posts / references: 0 posts) 1 comments [view comments]

If you are looking for low volume way to host your site, or a development platform to test your web server then Virtual Server 2005 may give you what you need.

Of course if you want hosting of it you can talk to us www.bitshop.com

I've seen people consider putting their low volume site on Virtual Server to be able to quickly restore from a hack. This isn't particularly a good plan though. Let me explain:

1) The virtual server (not just the guest os) would be open to attack unless you have properly firewalled it. In this scenario you likely couldn't protect your virtual server from outside attacks as tightly as you normally would. The guest os web server would, as normal for web servers, be open to public attacks.

2) If the virtual server was successfully attacked then potentially your guest os backup, and the guest os (your web server os) could be compromised at the same time, resulting in twice the work to restore both.

2) The guest os would need to be shutdown to take snapshots of the disk images, resulting in some downtime (Of course you could do a standard backup, but that may not give you the advantage(s) you are seeking).

3) The disk images will most likely be quite large (let's say at a minimum a few gigabytes) resulting in files that take sveeral minutes to backup.

This said, this solution does give you the ability to have a second virtual server configuration for the backup and immediately be able to save state and turn on the backup. In some scenarioes this may be fantastic, in others it may not be.

Of course if you are doing this to avoid being hacked then you have a problem that needs fixing (like updating your security, firewall, etc.). I'd recommend you spend the time figuring out why you are being hacked and close whatever is open to attack before you do this.

For security you may also want to check out ISA Server (Internet Security and Accelerator Server). This provides a high level of security for your servers by acting as a proxy. We always recommend you run a firewall and a local security policy including ipsec for keeping most ports closed on the servers.

As a development platform this gives you an easy way to load a server operating system on your machine and run your tests, instead of testing under Windows XP's web server, or instead of developing on your web server os. For this Virtual Server can be very handy.

Steve Radich - Server colo, hosting, troubleshooting and development services
BitShop, Inc. - http://www.bitshop.com

Have comments about this article, post them below...

---

Permalink (referenced by: 129 posts / references: 0 posts) 0 comments [view comments]

This, as you can see, is my first blog post on aspdeveloper.net - The topic isn't even about asp. However a very interesting technology, at least to me, if Virtual Server 2005. As such we will be developing some content about Virtual Server 2005 either here or on a different site. So for now content is here.

I've seen some questions about how to create ISO images that Virtual PC or Virtual Server 2005 can attach to. There are variuos tools out there to capture ISO images, one that is very easy to use and works flawlessly is www.undisker.com.

For example if you need to perform an install you may want to capture your operating system CD and then boot the ISO image in virtual server to perform an install as you normally do.

These iso images can also be easily mounted for installing additional components later, etc.

---

Permalink (referenced by: 119 posts / references: 0 posts) 0 comments [view comments]