I have a love-hate, approach-avoidance relationship with the Windows Home Server on the Centrale SOHO LAN here.  I love that it is always on and providing backups and shared storage of all of our computers and permanent information.  I cringe for the next time WHS will fail in some way.  This has me be wary and also quite aware that the WHS, a failed Microsoft product, is in need of a reliable backup and recovery strategy for itself.  It has become the household’s dominate single-point of computer failure concern.

I’ve been in denial about the precautions I must perfect.

This all came back to me when WHS recently reported that a small set of shared files, very old shared files, were reported as damaged or inconsistent.  The remedy is to remote-console into the WHS and run about 12 hours worth of chkdsk.  Before I did that, I did what I could to (unsuccessfully) grab the folder having the defective files just in case they were lost as part of the disk checking and repair.  In truth, I did not need the particular files any longer, but the same folder did have valuable content.

The chkdsk and a couple of reboots determined that WHS was restored to good health.  Watching the chkdsk reports, it appears that the problem was five index records needing to be created.

This reminded me that there another place where I was extremely complacent.  I have avoided backing up the WHS-shared files to USB drives because that also takes 8-12 hours.  I know these are USB 2.0 drives, but it is unbelievable how long it takes to backup about 300GB of information.

I had a USB backup drive already connected so I decided to perform a backup of all the shares.  I’m sure the backup on the drive was at least 12 months out-of-date. 

One thing I learned, while the lengthy USB backup was underway.  Computer backups still work.  That is, there is enough bandwidth and performance on the WHS that it can be doing incremental backups of computer images while backing up the separate shared folders to the USB drive.

I had to leave the USB backup running overnight.  In the morning, I learned that the backup was “incomplete.”  That’s all I knew.  There was no explanation.

Thinking that perhaps my 750GB USB drive didn’t have enough capacity for the full backup, or there were other problems, I removed the current USB backup and hooked-up a fresh 500 GB drive.  This time, I was careful to keep my system connected and watch the running log of the backup activity. 

Keeping my eye on the second backup revealed the problem: I had material in shared folders whose full path and filenames were too long to be recorded on the USB backup.  It was a small amount of material, but the backup process made several tries at backing up those after everything else was backed up.  I’m not sure why it kept trying.  I went to the shares of those files to see if I could do some renaming, but that was not possible because the files were in use by another program.  Guess who?

I ended the backup, satisfied that everything critical had been backup up.  Afterwards, I went into WHS and drastically cleaned up the two folders that had the problems.  They will be caught in the next backup to the USB drive.

It nagged at me that there was material that was all right on the server but couldn’t be backed-up because of path length and file-name length problems.   I mounted my backup USB on my desktop machine and immediately saw the answer:

Each incremental backup on the USB drive is in its own folder.  The folders from the WHS are two levels deeper.  There is a folder for the date and time of the backup (2012-11-13_1821, above) and another folder (shares, above) that then has the backup of the names shared-folder sets.   That was enough to make it not possible to backup some files deep in the Users folder.

The nice part is files that have not changed are not duplicated.  Instead, links are used to the earlier backup.  In this way, the backup drive does not fill so rapidly and each dated set appears as a complete snapshot.  The part that wasn’t obvious, until it happened, was the hazard of extending the file hierarchy that put some materials over the file-path-length limits of the file system.

I’ve just become one of the authors for the newly-created Apache OpenOffice.org blog. 

As I write this, the blog’s main page is blank.  I didn’t intend to have it be that way any longer.

Today’s lesson: Never ever do initial creation, pondering, and linking of a blog post using a browser.   At some point, I will do some fat-fingered fumble that causes the in-progress post window to close and be lost forever.

I will now collect myself, pour another cup of coffee, and check to see what is going at Apache OpenOffice.org email lists (not necessarily in that order). 

Cooled down, I’ll  maybe take another shot at creating off-line text that I can paste into the browser-based blog editing window without mishap.  The Apache Software Foundation blogs are housed by Apache Roller.   I have no idea whether there is a way to use Windows Live Writer to author for it.  Absent that, I suspect plain old text editing (via jEdit in my case) will be sufficient.  That seems to be a common foundation for the various ways of producing content for Apache projects.  It has something to do with document-management of everything via Apache Subversion and oversight by viewing change-commit logs.  This is a serious dog-food operation, and I haven’t quite got the taste for it yet.  (And folks think Microsoft NIH is excessive!).  I’ve concluded that tool-crafters, and I’m one, are a dangerous breed.

So, When Will Orcmid Get the Lesson?

Now, I already know to author wiki articles this way.  But it seems that this lesson is one that I will need to relearn every time I think I have a new way of intruding myself into cyberspace.  (Don’t ask about Google+, that is just too confusing.)

Fortunately, I was gifted with this lesson in my first attempt and I didn’t lose too many of my beautifully-crafted paragraphs.  (Sob.)

PS: I don’t propose to blog about the soap-opera around OpenOffice.org ending up in the incubator at the Apache Software Foundation.  You can get a taste for that in the related articles, below.  I intend for that to be the final mention I will make of that.

As a newly-minted Initial Committer on the Apache OpenOffice.org Incubator Podling, there is a *nix account and hosted web pages beckoning to me.  Not there yet?  Well, I am about repairing that.

It is natural for me to want to use my existing deployment model for the new site.

Not So Fast, Sparky

There are two important differences. 

First, the new site serves up as a folder of http://people.apache.org/ so I don’t know how much site substructure there is under the public_html/ directory of the ~orcmid account (and I am not that certain of the URL, for that matter).

Secondly, I have to deploy from my new desktop system, Astraendo.  That is because my developer access to ~orcmid and ~orcmid/public_html/ are via Secure Shell (SSH).  I will use PuTTY and its companion SFTP utility to deploy to the account and from there to the Internet.  This works better if I deploy from a working folder on Astraendo rather than on the (to be migrated someday soon) development server that my other sites are deployed from.

Baby Steps

The first step is to produce a single default page and have it visible over the web.  I will do that this way:

1. Setup http://compagno/orcmidApache.  This is a new sub-web on my local development server.  It is where I will make my pages using FrontPage or other web-authoring tools.
2. Create Visual SourceSafe project $/orcmidApache and have it bound to the compagno/orcmidApache development sub-web as its source control system.  I can then make a couple of starter pages and also create an images subfolder.
3. Create Visual SourceSafe project $/apacheOrcmid/web.  This is a mirror (by sharing) of content in $/orcmidApache.  This is where a current set of authored pages are found for publishing. 
4. Create computer folder C:\publicca\ApacheOrcmid\web.  This is where a current version of the web content is staged.   It is the assigned working folder (on Astraendo) for $/apacheOrcmid/web.  It is refreshed by Get Latest Version from $/apacheOrcmid/web only when I want to refresh the public site.  Staging here is decoupled from subsequent development until another Get Latest Version is pulled over.
5. Newer pages are published by SFTP transfer from C:\publicca\ApacheOrcmid\web to the ~orcmid/public_html folder at people.apache.org. 

This chain seems lengthy only the first time.  After that, everything just flows.  This structure also allows for backup from the site to the development source-control system, although I don’t expect to have need for that.

Although one can recover blog posts by scraping pages from a browser window into Live Writer, retrieving posts as drafts is more reliable and more complete with regard to formatting and other features.

Retrieving the Live Hideout Posts

Using the Windows Live Writer that I had configured for authoring Live Hideout (among other blogs), I could obtain a directory of all existing posts via the Open from: Orcmid’s Live Hideout dialog.  I opened each one, individually, bringing it back into Windows Live Writer for editing.

[Note: This is the Open dialog for Windows LiveWriter 14.0, the last version usable with Windows XP.  The preservation was accomplished using Scampo, my Media Center PC, before it was retired.]

Retaining Local Drafts

I transferred all of the posts in the list one by one.  Once opened in Windows Live Writer, the posts can be saved into the folder of local drafts with a single click of the “Save draft” button.  As long as that folder (under [My] Documents\My Weblog Posts) is preserved, there is a source of the Live Hideout (or any other) posts for repurposing, including migration onto a different blog.

An easy way to have a chronological list of only those drafts from Orcmid’s Live Hideout is by using the index page that is provided when the blog is downloaded.  Alternatively, one can use screen captures of the monthly archives on the blog:

Opening the Drafts Later

There are two ways to open one of the local drafts for repurposing or simply reposting to a different blog.  The first method is to use the Open dialog of Windows Live Writer:

This dialog will offer an abbreviated selection.  There are many drafts as a result of the preservation of all of the Orcmid’s Live Hideout posts as well as drafts from other activities.

The other way to re-open a draft is to go to the full collection of them in the My Weblog Posts folder:

Double-clicking on one of these drafts will open it in a new Windows Live Writer window.

[Note: The collection of drafts was performed using a now-retired Windows XP PC.  The folder was backed-up and transferred to a new Windows 7 PC and drafts opened with the newer Windows Live Writer 15.4 version.  The unavailability of new versions on Windows XP was one of the “inducements” to upgrade to a Windows 7 PC.  Fortunately, the draft format has not changed in any important way.]

Reposting a Preserved Draft

By selecting a destination blog account, the styling will be adjusted automatically.  Here I provided a category related to the new location on nfoCentrale Status.   I also specified the original 2007-11-12T21:30 date and time so that the location in time would also be preserved in relationship to other posts on the same destination blog. 

I could have made further changes, as I did for the 2007-08-25 post that was preserved by browser scraping,  In this case I limited myself to addition of a simple entry in the note on updates at the end of the post.

The reposted version is here on nfoCentrale Status.

Related Posts:

• 2011-03-05 Hideout: Preservation by Scraping into Live Writer
• 2011-02-17 Orcmid’s Live Hideout: Benefits and Problems Backing Up the Windows Live Space
• 2010-10-11 Preparing to Migrate Orcmid’s Live Hideout
• 2010-09-27 Woops, More Obsolescence: Windows Live Spaces
• 2010-06-14 Current Blog Status
• 2007-08-25 One Laptop Per Child: I’ll Take Two (reposted update 2011-03-08)
• 2007-08-25 Live Hideout: They Gave Me a Number, Took Away My Name (reposted update 2011-01-23)

Scraping Live Spaces Blog Posts

I chose to scrape my first Orcmid’s Live Hideout post because I wasn’t certain that I could use my Live Writer republishing technique to preserve one that old.   Since this was a post that made sense to preserve here on nfoCentrale Status, I could demonstrate the preservation before I have restored any of my other blogs. 

I used the Live Hideout archive to go to August 2007, open the first post there, and copy the body text to the clipboard:

This was done using Windows XP SP3 with Internet Explorer 8.  Fortunately, the formatting of the blog was such that I could select the body of the post without dragging along other material from the page.

I had no idea whether there’d be anything on the clipboard at all and whether it preserved formatting of the post.

Creating the Derivative Post

To recreate the post elsewhere, I pasted the clipboard into the body for a new post opened-up in Windows Live Writer.  I was delighted to see that all of the formatting of the original text was preserved. 

I used Windows Live Writer to set the new post to the same date as the original Live Hideout post.  I also added the new title and additional explanation to the reposting:

The version of Windows Live Writer used for this was version 14.0, that latest one usable on Windows XP SP3.  All of the formatting, including numbered lists, came across just fine.  I added supplemental information about where this post came from and how it managed to end up on nfoCentrale Status under the Orcmid’s Live Hideout category.

Successful Posting

Although I had no idea why it would not work, I published the scraped post with a certain amount of anticipation. 

The post did not appear on my blog’s front page because it was too old.  But it appeared in the archive in a freshly-created August 2007 folder and it appeared under the Orcmid’s Live Hideout category archive as the first historical entry.  The post is fully available here.

Delayed Reposting

I scraped this particular post because I was somewhat concerned that I might have to use that technique.  I also had a place to repost the article here as historical status of Orcmid’s Live Hideout.  If I simply wanted to preserve the material and figure out what to do with it later, I could simply have saved the pasted-up Windows Live Writer post as a local draft. 

Which leads to my preferred way of cross-posting from one blog to another. …

Previous Posts

• 2011-02-17 Orcmid’s Live Hideout: Benefits and Problems Backing Up the Windows Live Space
• 2010-10-11 Preparing to Migrate Orcmid’s Live Hideout
• 2010-09-27 Woops, More Obsolescence: Windows Live Spaces
• 2010-06-14 Current Blog Status
• 2007-08-25 Live Hideout: They Gave Me a Number, Took Away My Name (repost update 2011-01-23)

In case your Windows Live Spaces blog has not yet been closed or moved, here is some information about what you get by downloading the blog before you do anything else.

Downloading the Blog

Limitations of the Download

Examining the download of Orcmid’s Live Hideout, I concluded that the download is inadequate on two counts: spam glut and missing images.

1. There is an incredible amount of spam in the comments preserved in the download.  I hadn’t noticed these, but in the download they are seriously in your face:
      
      
     
2. Not all images are preserved.  Although there is a directory of images, not all of them are preserved.  I’m not sure why this is.  It may be because I uploaded the images to my Live Spaces separately and then included them in the blog post by reference.  Whatever the reason, images are missing for some of the posts:
      
     

As part of any retention or migration, the spam will need to be cleaned away and the missing images will have to be found and preserved.

What Was Lost

Here’s the post as it appears on Orcmid’s Live Hideout where it is still online until March 2011 unless I delete it or convert it to WordPress before then:

Download Details

The offer for a download is straightforward:

The blog is packaged into a single Zip file and offered as a download to your computer:

You will need to choose a place to store the file.  It’s useful to choose a more-descriptive name for the file while it is being saved locally.  For a recent Windows system, the Zip file will appear as a special sort of folder, LiveHideOut-space.zip in my case (showing after already having been saved once):

The names of files for the blog pages and for the images (in their own folder) are inscrutable.   That is no problem if we extract the entire Zip into a regular folder using the right-click option menu:

The Extraction request brings up a wizard for specifying what you want:

Specify a folder where you want the material to be extracted.  There will be many oddly-named files and an image folder, so it is best to name an empty or not-yet-created folder in a convenient location.  I chose a folder named LiveHideout in the same location (that is, next to) the downloaded Zip file:

At the end of the extraction, you can ask to view the folder with the extracted content:

Notice that there is an img subfolder, an index.html, and a style.css that is used for formatting.  If you double-click on the index.html file your browser will launch and present the folder as a little web site.

It is necessary to have an extracted folder for the browser to be able to access and navigate the downloaded blog.  When you don’t need to view the blog, you can delete the folder that you used for the extracted files.  It can always be extracted again from the more-compact downloaded Zip file.

Tip: Using the browser and the index.html will help you match up the blog page as you know it to the cryptically-named file that carries its content.

[Aside: You might notice that I have used both Windows XP and Windows 7 to prepare these screen shots.  The newer ones are from Astraendo, now operating in stealth mode until I complete my backlog on the retirement of Scampo.]

Cellular Pre-History

I recall land-line party lines, along with rotary dials of various kinds, ending with the Princess night-stand model and, for a time, the exotic rotary-dialing mispronounced “Erika phone” that worked on my landline until services went to mixed-rotary/pulse dialing or some other incompatibility intervened.

Push-button phones were more revolutionary for the telephone system than for the user, who had to deal with switch to all-numeric numbers and then the gross distortion of that concept, telephone numbers that were published as words that required a dial or a labeled touch-pad to figure out.

In this era, long distance service was also not usually flat-rate and there was a lot of reluctance to run up the telephone bill.  So I had not become what you would call a telephone person.  I harken back to the days when a long-distance call was cause for worried anticipation, almost as much as receiving a telegram had been in earlier times.

This is odd, in one sense, since when I was a freshman at Caltech in 1957, one of the students (who had a summer job with some part of the Bell System) knew about the coming of Electronic Switching Service (ESS) and how in-band control tones were used to access long-distance routing.  So while others were building what became known as Blue Boxes (or learning to whistle the proper frequencies into pay telephones with out-of-pattern numbers), I did not overcome my own risk-averse nature. 

1990 Era Just a (Cellular) Telephone

While I was still on dial-up, single-computer dial-up modem service (using the original form of MSN), I obtained my first cellular phone for a reason that now escapes me.  I had the idea that it would be useful for receiving (and perhaps making) emergency calls, but I didn’t think of it as satisfying any yearn for 24/7 connectivity.

It was just a phone.  The carrier was GTE and the bill was from some sort of green eyeshade era.  I usually had the phone turned off and I left it places.  My worst misadventure with the phone was when I left it in my car while it was in the shop and I couldn’t pick up the car until after returning from some business travel.  Somehow, my bill suddenly had lots of long-distance minutes to numbers in Mexico.  Fortunately, I was able to arrange removal of those charges from my bill.  I don’t know if the phone itself was missing (I think it was), but I closed the account and that ended my first adventures with cellular services.

2000 Era Nokia Communicators

My limited analysis of the cellular communications approach in the US led to believing in Europe’s GSM as the proper approach.  My interest turned, in 1999, to the Nokia Communicator 9000i.  When I retired from Xerox in December 1998, my colleagues collected a cash gift with the idea that I could buy two wireless walky-talkies.  I used it toward the purchase of a contract-free Nokia Communicator for developer purposes.  It had no SIM card, but was considerably less expensive than the over-$800 contract price from PacBell. 

In August 1999, the day we arrived ahead of the movers to begin our semi-retired life in the Puget Sound country of Washington State, our first stop was at the VoiceStream store in the Tukwila, Washington, Southcenter Mall.  I initiated a VoiceStream (now part of T-Mobile) contract and obtained a SIM card for the Communicator.  We obtained a Nokia 8290 for Vicki at the same time.  Of course, VoiceStream/T-Mobile did not provide their Internet connectivity in a way that was usable with the 9000.

The 9000i had a low-power Intel processor with GeoWorks, the DOS version of which I had used on my Heath-Zenith PC/XT clone ten years before.  I had found using early Windows versions as a shell and desktop for launching DOS applications far more convenient than GeoWorks and lost interest in it.  I was interested enough to download the Communicator software development kit and tools, but I didn’t go beyond that.

Although I took advantage of the Communicator’s synchronization with Microsoft Outlook, I found that my main interest was in simply sharing my contacts.  There were difficulties in maintaining synchronization and I did not go farther with the Communicator as a software or application platform.

One important characteristic of the 9000i was the infra-red adapter by which I could employ the Nokia Communicator as a modem for my 1998 Dell Inspiron 7000 (aka Compagno).  This became very handy at an AIIM Conference in Atlanta, Georgia, when my hotel’s telephone system went out of service.  I was able to access the Internet and e-mail via MSN dial-up service through the Communicator and the laptop infra-red port.

My second Nokia Communicator was an Italian-language Nokia 9110.  When in Italy for six weeks in 2000, I purchased the phone in Florence and used it the entire time we were in Livorno.  The use of prepaid additions to the Spazio Omnitel SIM was challenging because the operator instructions for adding Lire to the account were in Italian.  I had to hear the recording several times before I could figure out the keyboard sequence required for recharging the phone allowance.

The availability of the 9110 as a modem was also critical during that visit to Italy.  Internet connectivity via the carrier was terrible, but I could use the 9110 as a modem.  Fortunately, calling within the EU was inexpensive enough that I could call into a London dial-up connection to MSN and the Internet.  This allowed me to deal with a small crisis that required rapid analysis and reporting on a bug attributed to the ODMA Connection Manager that I supported.

The 9110 was not adapted to the GSM service used in North America, so the phone was reserved for visits to Italy.  It fell into disuse after Italy converted to the Euro.  I didn’t bother to see how to convert the SIM or whether the phone would continue to operate.

The Nokia 9210 Communicator was my third and last one, replacing the 9000i.  I bought it at a good price as the phone was being obsoleted by the pricey 9300.  The Outlook synchronization was better for a time, but I didn’t keep at it.  The 9210 was my standard phone on our family plan, using the original SIM that I acquired for the 9000i.  I lost the phone at some point in my 2008 attendance at the Microsoft Office Developer Conference in San Jose, California.  Fortunately, I tended to keep the phone turned off and both the phone and the SIM were password protected. 

Intermezzo: Mobile PC Feature Phone

Having lost the 9210, I needed a quick replacement.  I had the idea that a Windows Mobile phone would serve my need for Outlook synchronization more reliably and I obtained a T-Mobile Dash feature phone.  I held onto the Dash long past the contract requirement, expecting that smartphones would become more available and competitive.  I was waiting for something that appealed enough to overcome my reluctance to go for either an iPhone or an Android-based phone, especially at T-Mobile limited availability.

My happiness over synchronization of the Dash with Outlook was also brief.  An preparation for a trip to London, I had moved my Outlook 2003 data from my desktop PC to Outlook 2007 on my Tablet PC.   On bringing the data files back to Outlook 2003 on my desktop system, I discovered that the profile was no longer recognized by ActiveSync.  Unfortunately, I could not break the connection from the Dash side of the arrangement, because synchronization was required to eliminate the synchronization.  Stuck in that impasse, I waited for the day in which (1) I would end up repaving the phone anyhow as part of an OS upgrade or (2) I abandoned the Dash for a better solution.

2010 Phones as Ecosystem Appliances

In November 2010, I abandoned my record as a late-adopter stance and pre-ordered a Windows Phone 7 as my first smartphone.  I paid cash, which may have been a waste of money: I added a second Windows Phone 7 under two-year contract for Vicki, so I am basically contracted on the family plan service for the same period and I doubt that I will want to replace my HTC HD7 before that.

Meanwhile, as part of the Nokia-Microsoft agreement on Windows Phone 7, we are now told that the competitive battle is not among devices but ecosystems.  We’ll see how that goes.  I am more heartened by the situation than many of the commentaries I have been reading.