This was a really odd situation in my VMware ESXi environment involving a virtual disk that needed resizing. I needed to add a few gigabytes to a virtual hard disk, which I have done scores of times and have my process pretty well down pat. This one was a little different and I had to do some research and obtain some new tools.
My normal process is this:
- Add space to the .vmdk using the vSphere Client / vCenter Console.
- Use the diskpart command on the server to select the disk and volume in question, then issue the extend command.
- If the command cannot complete (because the disk is in use in a certain way by the server), then I have to shut down that server and add the hard disk to another virtual server (as it's D: drive, for example). I then perform #2 above from that server's console, detach the disk, and then power up the original server.
It's one of my favorite little admin tasks to do in VMware, because I know what it would take for me to add disk space to a server in the "real world", and I giggle a little bit every time I do it.
Today when I tried to use diskpart I could not see the volume when I ran the list vol command. I knew then that something was wrong. I check Disk Management (it's a Windows Server 2003 machine) and saw that the disk was a dynamic disk and was "Invalid". I had never seen that before. I could still access the disk just fine -- I could still see and manipulate all the files on the disk... I just couldn't make any changes to the size. I also did not recall ever converting the disk to a dynamic disk.
Now I'll fast forward to the end... there are a few resources out there for handling this situation. The best, most concise one that I found is here. The explanation is that a byte on the disk that identifies it as dynamic or basic has become corrupt somehow. The HxD disk editor can be used to see the value and even make a change to it. Be careful, though, because you can really mess things up if you aren't careful.
According to the above instructions the trick is to look at sector 0 location 1C2. For dynamic disks this value should be 42. For basic disks the value should be 07. In my case the value was 07 but Disk Management was showing the disk type as dynamic. The instructions in the link above say to only make a change if the value in 1C2 is 42 (change it to 07), so this did not fit my situation. I did notice that the value just below it in the HxD window was 42 (location 1D2). I decided to give it a try, so I changed the 1D2 location from 42 to 07 (to match the value in 1C2) and then my server booted just fine.
So, based on my experience I'll add that it's important for the 1C2 and 1D2 values to match. Of course it's possible that this isn't applicable in all cases.
I am sometimes asked to work on technology stuff around the house. Surprise, surprise. Today my wife asked me to help her transfer her blog from Typepad to Blogger/Blogspot. I haven't fooled with blogging at all except for once before when there was some trouble with getting her Typepad blog started. I only recall that is was a lot harder than I thought I should have been. Lots of scripting, Linux/Web server stuff... I don't know how non-IT types get it done, as it was a real challenge for me. Well, today I had another challenge on my hands.
She is adventurous enough to have taken the task as far as she could take it. She was at a point where she had exported her Typepad blog posts (in an .mht file) and was trying to download the Google blog converter to convert that file to a Moveable Type file (.mt) so that the blog posts could be imported into Blogger. Seems pretty simple, but there were several challenges.
The download file from Google's blog converter was a .gz file, and the computer where my wife blogs had no program to handle that type of archive. I downloaded a 40-day trial of WinRAR to handle this task, and then extracted the folder structure from the Google download file and stored it in a temporary location on the computer. I then set about trying to use the converter.
The converter has sparse documentation but I did find a .\lib folder that has some batch files in it. I used the one called movabletype2blogger.bat but it had errors. Turns out I had to download and install Python (I used version 2.7). I still received errors because the paths in the bat file were all wrong. I pretty much hardcoded all the variable paths that the file's author had tried to make. Here is the batch file that finally worked for me. Note that I had installed Python into its default folder, \python27.
set PYTHONPATH=c:\python27
set BASEPATH=c:\temp\google-blog-converters-r89
set PYTHONPATH=%PYTHONPATH%;%BASEPATH%\lib
c:\python27\python "%BASEPATH%\src\movabletype2blogger\mt2b.py" %1
I continued to receive errors until I renamed the Typepad file from .mht to .mt.
The command that I finally used was movabletype2blogger typepad-export-file.mt. run form a Command prompt, of course. (I had to navigate to the C:\temp\google-blog-converters-r89\bin folder, first).
This produced several seconds worth of scrolling text down the computer screen because the script output was to "stdout", so I simply redirected the output to a file, like this:
movabletype2blogger typepad-export-file.mt > import-to-google.xml
A quick look at the .xml file showed that the file contained blog posts. Success! Seems like I'd almost be finished at this point, but no... the .xml file produced some mysterious error message with no explanation when I tried to import it into Blogger. After some trial and error I was able to pare down the .xml file to a single post and did import it successfully. To do so I opened a copy of the .xml file in notepad and searched for "ns0:entry" (for the beginning of a post) and "/ns0:entry" (for the end of a post). I had to leave "</ns0:feed>" at the bottom, but I found the first "/ns0:entry" and deleted everything between it and the the "/ns0:feed" tag. I then proceeded to use the same process to import additional posts, one at a time at first -- then several at a time. At one point near the end of the 120+ posts the imports began failing. I finally determined that the export from Typepad or the conversion process had placed all the comments (where people had commented on the blog posts) at the end of the .xml file, and these would not import.
So after all that my wife's blog was up and running, but no comments from her Typepad blog were brought over. The blog's pictures still pointed to her Typepad blog, so she had to re-point those to a different location for the pictures before she took down that blog. There were also some minor formatting quirks that she had to work on (extra line breaks and such), but those were simply aesthetic.
So, with all that blog "piping" experience I thought I'd start up a blog and write about the experience. Let me know if this helps you!
I'm a technology guy. I sometimes tackle challenges or learn things that might be beneficial to others. This blog is my attempt to share some of my experiences with you, in case you might have run into the same issues. I'll keep the information here to my professional / technology life.
I'm on LinkedIn.
