Copying/rearranging EHDs

Strega

SatelliteGuys Family
Original poster
Aug 6, 2023
44
77
West Coast
A while back, I had to call Dish to update my credit card expiration date, because the website wouldn’t work. Even before I got to an agent, a recording told me that my ancient (722K) would need to be replaced and the agent would help with that. So, I have an appointment to have my 722K replaced by a Hopper 3 in a few days. (Such a simple thing became such a big thing.)

Since the Hopper 3 can handle 7TB disks, I was thinking it would be nice to copy/rearrange my saved programs from a motley collection of .5TB to 1.5TB disks on to a larger drive or two. But I’m wondering how easy this will be. Is it as easy as: Mount new disk. Format as ext3. (All a single partition?) Mount an old disk. Copy files. Mount next new disk. Copy more files… etc.? Or are there some kind of index files that will be all confused if I do something like that? Also, are the file names easily understandable or more like UUIDs? (In case I want to segment them onto two drives in some logical way.)
 
A while back, I had to call Dish to update my credit card expiration date, because the website wouldn’t work. Even before I got to an agent, a recording told me that my ancient (722K) would need to be replaced and the agent would help with that. So, I have an appointment to have my 722K replaced by a Hopper 3 in a few days. (Such a simple thing became such a big thing.)

Since the Hopper 3 can handle 7TB disks, I was thinking it would be nice to copy/rearrange my saved programs from a motley collection of .5TB to 1.5TB disks on to a larger drive or two. But I’m wondering how easy this will be. Is it as easy as: Mount new disk. Format as ext3. (All a single partition?) Mount an old disk. Copy files. Mount next new disk. Copy more files… etc.? Or are there some kind of index files that will be all confused if I do something like that? Also, are the file names easily understandable or more like UUIDs? (In case I want to segment them onto two drives in some logical way.)
To copy one drive to another, I have used the following procedure. After you get the first one copied (preferably the largest one), you will need to copy the files from another EHD to the Hopper, then copy those files to the EHD. I don't know of another way. Make sure you read the complete procedure.

 
  • Like
Reactions: charlesrshell
Direct connect between old and new STBs is recommended by some. I’ve always moved STB recordings to EHDs. My EHDs have always been recognized by the new STB.
 
  • Like
Reactions: charlesrshell
Thanks everyone! It seems like my plan to use a Linux PC to copy files from multiple source disks to destination disk(s) assumed more knowledge than we have about the file structures. I’ll do it llokey’s way.
 
  • Like
Reactions: charlesrshell
I have been transferring EHD files to the H3 for the last three nights. No interruption, no problem.

I’m going thru my many EHDs to ease dropping Dish later this year, if it comes to that. I have discovered I recorded several movies I don’t want to watch. Surprise!
 
I decided, based on the 999 tiles limit, to play it “safe” and stick with small 2TB drives. I found a powered USB hub sitting in a drawer, so I connected that, and then connected a 2TB bus-powered drive. I let the hopper format it, then mounted it and an old 1TB 3.5” drive on a Linux PC and copied the files over. Browsing the files, it really does look like you could combine a couple of small/old drives onto a new one. The structure I see is simple and regular: 1 directory per show. I saw no special index/db file which makes me ASSUME you could get away with mixing and matching as you like.

It also looks like you could probably build a script to reorganize shows if you wished. Each directory has two files, that while not designed as human-readable text files, do have the textual descriptions of the show in that directory. So, with a little investigation into their structure, I suspect you could come up with a script that would only move shows matching a show name to a new drive.
 
The bm file seems to describe the current show title and seemingly the next show – like it’s a capture of what comes up when you press the browse (right arrow) on the remote?? The cat file seems to have the show description, but I don’t see the series name in that one.
 
Oh, I was wrong about the cat file not containing the show title. I just needed to get a bit smarter and look with strings instead of cat! (Because cat spews everything into the terminal window.) To prevent any confusion, “cat” is a Linux command, and each show directory coincidentally has a file named “cat”, so before I was doing “cat cat”, but “strings cat” is the better way.

Here is the output of "strings cat" for one video directory:

.ViP722k DVR
NOVA
The Elegant Universe: Einstein’s Dream
Series/Special. (2011) Author Brian Green (‘The Elegant Universe’) explains string theory; physicists attempt to unify general relativity and quantum mechanics. New (HD) (CC) (Stereo)
 
And Here’s how to get a list of video directories with NOVA in them (from the DishArc dir):

grep -l NOVA */cat

(Of course this is a bit crude since it will match NOVA in the description as well as the show title.)
 
  • Like
Reactions: charlesrshell
It looks like the fields in the cat file are surrounded by nulls, so grepping for a title surrounded by nulls should improve selectivity.

So I think this would be a way to copy all files with the title NOVA to some $outputdir (the destination depends on where you or your distro mounted the new disk)

outputdir=SOMEPLACE
cp -Rp `grep -lPa “\x00NOVA\x00” */cat | sed -n -e ‘s/\/cat//p’` $outputdir

It may need a sudo in front of it, if you’re not root.

I have not tried this!!! I just realized that the old disk that would most benefit from selective copying, was one that I had already copied 100%. Whoops!

I used to do stuff like this on a VAX 11/780 more than 40 years ago, so it just became amusing that this stuff is all pretty much unchanged after that time.
 
I’ve used many OSs from simple things like Altair DOS, CP/M, and Northstar DOS (which actually had a jump to address instruction in the CLI), to big iron stuff like MVS, CANDE/MCP, Kronos… but Unix just keeps going and going… And after reminding myself how easy it is to connect some commands to do what you want, I can see why.