I’ve made a couple of posts about my Plex server and recently how I’ve been trying to clean up the files and metadata, specifically, the album art.  Using a self-written utility to audit the artwork dimensions and a downloading utility to find better versions, plus the effort of scanning and cleaning artwork that has no high quality option, I’ve done pretty well.

The next step was to clean up – actually add – the artwork for artists.  It’s a smaller effort, only 600-some items instead of 1600 items.  However, I wanted to get through it easily.  This process requires a bit of explanation on how Plex handles things.  Some of this I learned painfully as I went.

Plex gets a lot of its data from Last.fm, which is really cool and pretty effective.  However, when you don’t want to use that data, you have to disable that “agent” for your library.  Because I was manually managing the artwork on my library, because I wanted hi-res art, and I also wanted the specific original cover instead of the cover on the latest remaster of the album, because of all this, I set everything in plex to manual.  No last.fm anywhere.  That was good.

However, that also meant I had no artist images or bios.  Since I had cleaned up my album art, I wanted to now download artists from last.fm.  I wasn’t going to be picky about images there.  So I re-enabled the last.fm agent and stumbled around trying to figure out how to get plex to refresh the artist data from last.fm.  It’s important to know that the command to use in the artist menu is “Match”.  That will allow you to pick an artist from Last.fm for which to get metadata.

It’s also somewhat important to not use the command, “refresh all metadata”, because that caused Plex to utilize the newly-enabled last.fm agent and download all the data for my albums again.  That wiped out over half of my 1000px covers with 300×300 images.  When I saw this, I was devastated.  The worst part is that even if I used my auditing utility and worked my way through those hundreds of albums, the image picker in Plex gives no indication of which image is small and which is large.  It shows them equally at 150×150.  It would literally be a blind guess between 2 and 4 different images for 900 albums.  It’s not reasonable to accomplish.

I resolved to rebuild my Plex database from scratch.  That would wipe out all the effort I made on the artist artwork, but at least I could preserve my work on album artwork.  I would also lose hours I spent building my playlists.  And I figured if I was going to wipe out the database, I would try some more aggressive actions against it.

My auditing utility has only ever done database reads.  It doesn’t do any updates.  But I spent some time restructuring the code and adding a new feature, which I descriptively named, “replace art with largest version.”  See, Plex never gets rid of things, it just accumulates.  My data folder used to have 44k files in it, but after the last.fm influx, it now had 124k files.  All of my old artwork was there, too.  I just needed to scan and compare sizes, which is what I built my utility to do.

And within a day, I ran my new code and updated all the records to use the large artwork I had added previously.  It worked just as planned.  So now, I have my large artwork back, my playlists are still in place, and the artists artwork is still there.  It seems that anger is an excellent motivator for progress.

This isn’t really a “biggest and bloatedest” post, but it is in the same kind of vein, since it involves leaving behind a company that I once really enjoyed.  This time it’s the great remote-access utility, Teamviewer.  I was first introduced to TV long ago when I was doing remote computer assistance (an Uber of computer helpdesks – way before its time).

As time went on, TV got more and more advanced.  They added many new features specifically for providing helpdesk services, none of which were really useful to me.  All I needed was remote desktop access and sometimes file transfer.  I didn’t need chat, or ticket logging, or video capture, or lots of other things.  So I guess in a way, TV did become big and bloated.

But the software itself was impeccable.  Very well-written code and always in touch with current Microsoft security and coding practices.  It’s software that I would buy, but unfortunately, it wasn’t really for sale.  TV’s business was business users and consumers were trusted to use the software for free for personal use.  That sounds really good and fair, right?  It is, but I did feel a little guilty about it.  Not because I was using it for business use, but just that I used it SO much.  I would love to buy a license, but the cheapest you could get was a $50/mo subscription.  Ohhh, I hate subscriptions.  And $50/mo is not really reasonable (to me) for personal use.

So I kept using it for free, until one day I started getting notices that TV detected I was using their software for business use.  I don’t know exactly what they noticed that seemed suspicious.  I have a couple ideas, but I don’t know for sure.  If I choose to really think about it, it’s worrisome that the software is actively watching what I do to catch me doing business operations.  Anyway, I ignored the warnings, because they weren’t applicable.  When I rebuilt my computer and connected to it remotely, I got a much more severe warning.  It declared that I was using the software for commercial use and my connection would be terminated within minutes.  Subsequent connections also got cut off as well. 

I filled out an online form to appeal their judgement, which they said would be responded to in about a week.  But I pretty much knew that my time with Teamviewer was over.  It was time to find another remote access utility.  And the one I eventually chose was the free one built into Windows – Terminal Services or Remote Desktop Connection.  I was a little hesitant to implement it because of the reputation RDC has for being vulnerable to attack.  However, taking my time and considering the risks made me more comfortable in the choice.

For most RDC breaches, attacks are made using common account names and weak passwords.  Neither could be true in my case.  In fact, my configuration is more secure than Teamviewer.  With TV, an attacker needs two pieces of data: the computer ID and a password.  To attack me, you need an IP address, a custom port number, my username and my password.  That’s twice as many elements needed, and the potential values are vast.

So, that’s how I now waste the day away when I’m at work.  I’ve disabled Teamviewer, so that’s one less attack vector for my computer, which to be honest, always spooked me.  TV has not had any account breaches that I know of, but their user database would be a goldmine for hackers.

A title like that is pretty ominous considering I just rebuilt my computer, but the damage is not system wide, I’m just dealing with one application.  But as applications go, it’s a big one.  I have to rebuild my Plex library, which means redoing all my playlists and the extensive metadata improvements I had made over time.

After I had gotten all my files moved from my old drives to the new ones, I knew I had to figure out how to restore my Plex database.  I knew there was a backup that ran every 3 days, but I didn’t know where that backup was or how to restore it.  From research on the Plex website, the only thing I really got out of my questions for restoring was to copy the entire data folder from the old computer to the new one.  Well, ok.

And not unlike my amazement at Lightroom’s massive file structure, Plex was the same way.  Tens of thousands of files got copied to my new hard drive, then I reinstalled the server application and fired it up.  No initial errors, which is great, but when I went to the home page and the music library page, all I got was an error message.

In my mind, I sort of expected it.  The reason for that was because I hadn’t installed any application updates in some time.  Why should I?  The features added didn’t pertain to the music portion of Plex, and there weren’t any security fixes, so I never needed them.  Why potentially introduce bugs with a new version if what I had was working fine?

Well, this came back to bite me when I tried to restore a database for an application version much older than what I installed.  And I had no idea what version I had been running before, so there wasn’t an option to install a matching older version.  So there you go.  A mismatched application and database equals a non-functional application.  Time to start over with a new database matching the current application version.

Rebuilding the playlists will be tedious, but I have a pretty good idea of what I want in them, and to be honest, I really only played three or four of them, so I can start there.  The metadata updates will be tedious as well.  I had decided that I was going to use the plex server as my definitive source of collection info.  But now, maybe I will start using Discogs, which can get me info on my collection in greater detail and more expediently, using their mobile app.

I’ve taken a “no rush” attitude with most everything in my life now, so I know eventually it will all work out, when I’m ready for it to happen.

A few posts ago, I talked about how I was going to be upgrading my computer to take advantage of new storage.  And that involved swapping out all the important parts of my computer: motherboard, CPU, RAM, and drives.  Well, all the parts came in and sat for a few days, because I know, if I’m going to start a project of this size, it best be in the morning, so I have all day to recover from screwups.  I decided Sunday would be the day and I woke up fairly early to get a start.

The first thing I did was get one last backup of the old computer and break the mirrored drive set.  Then I got to work replacing the guts inside the case.  I spread everything out on the kitchen table, old and new components everywhere.  It all went pretty smoothly and even with taking my time, I was done within an hour.  With everything back in place and ready to go, I brought it back to my office and plugged it in for a smoke test.  All the drives in the computer were blank, so there was nothing to really boot up.

But, to my dismay, I got nothing.  Fans spun up, but no video and no POST beep.  Sigh.  My prior experience says that this is usually a short somewhere.  And what has always been a good troubleshooting step for me is isolation.  So I took everything out of the computer except for the bare minimum.  On startup, no change.  Big sigh.  I took the computer back to the kitchen for some reassembly and inspection.  I didn’t see anything wrong with the mounting posts or anything that could short anything else out.

When in doubt, read the manual, right?  And while I was paging through it looking for some connector I missed or hooked up backwards, I noticed that yes, I did screw something up.  I put the RAM into the wrong slots.  What an idiot.  Ok, fix that up and put everything back together again.  Take it back out to the office and start it up.  Still no video and no beep.  What the fuck.  Back to the kitchen and tear everything back out.  I then decide to consult the Internet, which I know is going to be useless since my symptoms are so vague.  How can I determine if the board or CPU or RAM is bad?  I don’t have extra parts to swap out and isolate the differences.

The first result I clicked on for “no video, no beep” had a confident first response of, “bad power supply.”  Really?  I just changed out this power supply a year ago.  And the old computer was working fine with the same power supply.  BUT, I had no other ideas and I did have a backup power supply (that old post means something now).  And when I put in the new power supply, well, you probably know what happened.  It started up.  I don’t fucking get it.

Now.  Hardware complete, now software.  I had my new M2 SSD drive and my two 8TB drive in the computer ready to go and I had the latest build of Windows 10 on a USB drive.  As has been the case with Windows for some time now, the installation process was quick and easy.  When I went to set up my two big drives to hold all my data, I lost my shit.  Windows installed a recovery partition on one of the big drives.  That totally fucks up my mirroring plan for the drives.  Yeah, it’s time for lunch now.

After a nice calming lunch, I came back and reinstalled Windows without the two 8TB drives attached.  This time, the recovery partition was placed on the proper drive and I was ready to move on.  Before getting into the mass data transfer, I ran Windows update over and over again until I was fully patched.  The last thing I needed was a restart in the middle of my data copy.  It’s around 2:00 right now; about 4 hours of effort so far.

I started creating the virtual drives to hold the files from my old drives and I found out that creating a 1TB file is not a quick process.  It’s something like a hour or more of waiting.  Then the file copy is nothing speedy either.  I got the 50GB drive copied over and started working on the first 1TB drive – the first of three.  While that’s going, I installed software from the Internet that I didn’t have on my archive drive.  That’s one thing about this process is that there’s dependencies, where I can’t do this until this other thing is done first.  And one of the big ones is that I can’t install a lot of software until I get all the data copied over.  So I must work and wait.

To be continued….