Is a "dual install" possible?

[Forester1]

Hi everyone. My name is Forester1, and I am a trainzaholic.
I started with T:ANE, and within a few years I had taken advantage of the TMR plugin, the deluxe bundle add-on, and free and paid routes and assets to where I was soon consuming a Terabyte of hard drive space. I moved my install to an external drive but found things slowing down considerably.
When TRS19 came along, I thought this would be my chance to break the habit. I would perform a "Lean and mean" install, installing the platinum edition, with DLC I had already purchased, and not expand with more routes and assets until I had thoroughly explored everything I had. But over time, I found myself once again taking advantage of all of the wonderful things available: Philskene's wonderful routes, the megaroutes, assets by Dave Snow, MSGSapper, and a host of others, and now I have TRS19 taking up A TERABYTE AND A HALF on an external drive! Again, I am noticing big slow downs in trying to load routes. Yesterday I waited maybe an hour(?) after selecting "Create Session" for the small route "Chiyoda Branch Line". Large routes are becoming untenable.
I am finding that once I have built a huge install it is very difficult to try to reduce it. There are a lot of dependents and dependencies, and it becomes difficult to know what can be archived off to CDPs without eventually missing something. Especially when some of the largest files are just libraries and mesh files for really small assets!
So, what I would like to do now is start over again with a "Lean and mean" install but have the current install available as a sort of "Archive", to keep from having to search out and re-download a world's worth of assets that I already have (not that I want to do a lot of thar right away! ). My idea is to back up the current data folder, then uninstall TRS19, then reinstall TRS19, using the default data folder on the C: drive, but be able to point the launch window setting to the current install if I need to. It is very hard to extract something unless I can bring it up in CM, so I would want to be able to kind of point TRS19 back and forth at infrequent intervals.
Does that sound possible or even logical? Any ideas are welcome, except starting over with TRS22. I am not ready to buy the beta and hope to hold off until there is a platinum version of 22 offered. If I could do what I am proposing would installing the upcoming SP5 affect it? or does that only impact the executables folder?

[grazlash]

Being a Trainzaholic is a sad thing, I should know, I'm one too! Tragic isn't it.
All is not lost, however. What it sounds like you are attempting to achieve, would require a DBR each time you switch between databases. I have a couple of suggestions: one would be to archive the route cdp's and any associated modified files through CM until needed and can be stored online or on any other drive. The other would be to (if possible) install a large SSD drive on your computer, which will speed things up and allow you to run everything from a single drive without bogging down as it fills up. Loading times for Trainz and loading routes are much reduced, assets appear almost instantaneously and datbase rebuilds are no longer painful. Yesterday I updated 2 almost identical installs, one to TRS19 SP4 and on the SSD the latest TRS22 beta. The platter drive took 3 hours and 36 minutes to run the DBR, while the SSD DBR took 4 minutes, both DBs are around 900gb. An external SSD would be marginally slower than an internal due to transfer rate. Updating to SP5 will only affect the game, however you can't run 2 different builds on the same database.
Catch you at the next Trainz Tragics meeting!
cheers
Graeme

[JCitron]

I'm up there as well, Forester1 my database is about 1.6 or 1.7 TB and it's time I start cleaning house.

I think your issue is bottlenecked performance due to loading up from an external hard drive. The external hard disks tend to be slow devices to start with, meaning they're usually only 5400 RPM. This is a slow drive even when installed in a system. They're fine for writing backups to and retrieving backups from, but don't use them for reading a lot of data like you are doing.

There are two solutions for this.

1) Look into something like these Crucial External SSDs.

External SSDs | Portable Storage | Crucial.com

The 4TB device is suitable for data storage and is a lot faster than that external drive.

My concern though is it's an SSD and that needs to be backed up to something else because they do wear out a lot faster than a regular hard disk.

2) The alternative to that is to get an external drive bay and plug in a regular hard drive into that. With this solution, even though it is only USB 3.0, you still have the benefit of a 7200 RPM drive. Get a device that will support large hard disks to futureproof your setup so that you don't run out of capacity and need to purchase a new drive bay in the future if your drive capacity becomes too small.

Amazon.com: Sabrent USB 3.0 to SATA I/II/III Dual Bay External Hard Drive Docking Station for 2.5 or 3.5in HDD, SSD with Hard Drive Duplicator/Cloner Function [10TB Support] (EC-HD2B) : Electronics

This is only an example and there are many versions of these. I have an ancient 2TB maximum-support unit from Thermaltake which I have used in the past for Trainzing on the go. My device is going on 12 years old now that I think about it. When I was out storm chasing, I took my hard drive with me and plugged it into the external device similar to the one shown in the picture. I then had my complete Trainz data base with me. I did have a backup at home so there was nothing to worry about. When I got to the hotel each night, I plugged the device into a USB port on my laptop and Trainzed all night. When I got home, I put my drive back into my PC and continued where I left off.

Now with the external devices like these, you can have multiple databases. Once you setup your databases, meaning login and get it configured for your install, you can then switch between them without logging in each time and running a DBR as long as you keep the drive letters the same and the data-path the same. What you're doing here is substituting one database for the other. You'll have to shut down Trainz in between, but the drives can be ejected on the fly. The alternative is to setup the program on each drive with its own configuration associated with that particular database.

[G.M.]

Many installs are certainly possible as described


a few considerations:


Speed
I mean loading speed of a route:
-best place for the actual data = your local dir. is your fastest harddisk
preferably an m2.SSD, is an SSD directly on the motherboard
-smaller assets (files sizes) load faster, so think before you PBRize everything
-less assets loads faster, example a good route is a better route if less different ground textures etc.


Greed
not meant negative, but think about what routes and content you like best and actually use
if you think hard you may come to the conclusion its just 10%.


When I install a route, I only install the route, then slowly find and add to the route
only things that I deem essential and know that work (do repair on the fly)
I replace things I really don't want(or cant find) and then do delete missing assets


Purpose
are you a:
-collector, driver
-route builder
-content creator
-beta tester


for each category, quite different installs
FI a route builder, should have TANE, TRS19 build 100240 and a test version of the latest build
the version he/she works in should not have payware installed, why?
it then cannot end up in the route by accident.


Most of us are a bit of everything ofc.
here I have 3x TRS19 1x TRS22beta and many older version starting with TRS2004 fully working
but all versions as clean as possibly can.
Became Plus in december, but did not download any free DLC, go slow, give yourself a present everyday
enjoy that new item, like you collected model trains in the past.

[Forester1]

Thanks for the suggestions, guys. I believe you are right about getting an internal SSD, but large ones are still very expensive. I really think I would like to slim things down. M.G. under purpose, you missed one. I think I am a "pack rat"! Not so much greed as attracted by every shiny baubel that comes along, and with Trainz that is pretty much constant! I also download a lot from third parties, because I fear they may go away like TPR or the Backshop, but of course in that case I don't need to actually install them until needed, and I am learning that too slowly. Grazlash, I started trying to archive by setting CM to size, but it wasn't long before I ran into a lot of really large mesh and library files. Then you have to check dependents to see how many and which assets they are part of, and then check the dependents of those to see if they affect any routes or other assets. It gets to be quite a chore, but I guess I deserve it! Maybe I am better off just waiting and moving to TRS22 with lessons learned.

[Forester1]

Just to add a thought. Can you filter CM for assets where Dependents = False? I guess I'll find out, but that would really help to get things archived off.
EDIT: Looks like it only has "dependents include". Not sure how to phrase that. Dependents include none? null? blank? zero?

[1611mac]

Check this discussion - there's one tip given to find un-used assets. (see post #4) It was a while ago and may not apply now.... have not tried.
https://forums.auran.com/trainz/show...87#post1782687

[Forester1]

Thanks for pointing that out, 1611mac! I am going to have to read it a few times to understand it though. Even for a programmer like me, that is some logic: ​Open CM and select installed plus category -- sessions. Highlight all and select show dependencies recursively. In the new window add another asset kuid underneath the one shown. Copy the kuids in the top one to the bottom one. Change top selection to installed and add new plus selection below it and select AND NOT.
EDIT: Well, I did figure it out, but it didn't quite get me where I wanted to be. It may be listing assets (88,223) not used in routes and sessions I have installed, but it still returns assets that have dependents. So, If I am going to export assets to a .cdp, I still have to find the dependents and make sure they are packaged together. It is much closer than where I was, and may be doable, just more labor intensive. It would be great if they had included Dependents = FALSE in addition to all the other various options.

[1611mac]

Sorry that didn't work for you. Like I said, It had been a long time since I tried it. Maybe we're doing something wrong or there's an error in the instructions.

I try to stay positive.
- I just imagine that that any kuids I have installed that are "extraneous" (not being used) will be the key item needed in the next freeware route installed....
- I can alway obtain a larger disk/ssd
- Assume cost of SSD will continue to come down
- Assume CPU's GPU etc will get faster as they get cheaper
- Assume N3V will improve asset handling (caching, transfer, etc)

[JCitron]

I went through this cleanup process a year or so ago. I figured it was time to clean out the fridge and remove that old stale and dried up dishes of peas and carrots that end in the back behind the beer and soda.

I spent hours deleting old build 1.0 and build 2.0 assets I haven't seen since TRS2004 or TRS2006. The process was difficult at first because I thought about the people that made the assets who haven't been around for many reasons just like looking at old papers as we clean out draws. The process got easier and easier, and I removed as I went along. I felt great having accomplished this difficult task. I wish the process was so easy.

I used a simple sort by build, along with an installed and not pay ware type filter. I then highlighted a bunch of content and checked for dependents. If there were no dependents, I deleted the assets.

This worked okay, but there's a problem. Instead of telling us which asset is a dependency of something else, it just tells us there are dependencies when viewing the results. This then required sifting the data again, meaning highlighting fewer assets at a time to reduce the amount until I found the one or two assets that had dependents of something else.

For me the process became too painful, and I gave up because the amount of work needed to remove these few assets I could remove wasn't worth the effort. If we had a filter, however, along with another column in Content Manager as I've suggested in the past that could indicate that something is a dependency, we could easily glean through and remove old stuff without going through multiple hoops, summersaults, and ropewalks to clean up old assets.

[clam1952]

Safer to disable first when cleaning up, then look to see what's broken before deleting anything.

[Forester1]

Yes, I export to CDP prior to deleting anything. Then, if it is needed, I can just drag and drop it back onto the CM. I have been trying to sort by size, but that is not working well, since the 2019 stuff is much larger than anything before! Try sorting by size, and you find a few large routes at the top (which of course, are newer ones I want to keep!), but then see how many libraries and mesh files come up of 800MB or more! and the assets they go to may only be a few KB, but you have to keep the mesh files! Hence the dependents problem, as to delete say, Socalwb909's 735.8MB "53ft Container Trailers Mesh Library", I have to look up dependents, and there are two 45K container trailers. Or, I have a bunch of LBSCR Cab views by Camscott, all running 760+MB, but to delete them I have to find all of those locos and their dependencies. And, I have a 952.5MB file "CAJE-Libreria Maejestra Objetos" by guerrero1804. It has seven dependents, all but one under 100k, so I would have to gather them all up, but some of them belong to routes like "Vivay Vivarais 1960", so then I have to decide if I want to archive that route off and maybe others if I want to archive off that big Library file. Whew! So, it gets to be quite a decision tree!