This is an oc­ca­sion­al se­ries of posts where I will share ideas for pro­grams that don't ex­ist, but should. The goal is to in­spire de­vel­op­ers to cre­ate use­ful tools that can make our lives eas­i­er. Or, more like­ly, to re­mind me about these ideas so I can cre­ate them my­self. Or even more like­ly, to just get them out of my head so I can stop think­ing about them.

Idea 3: A program that does backups the way I want

I don't want to con­fig­ure things on many com­put­er­s. I want to con­fig­ure things once in one com­put­er and have back­ups for ev­ery­thing I care about, done prop­er­ly.

There is a great back­up pro­gram called restic that does back­ups RIGHT. It is fast, it is se­cure, it is easy to use, and it has a lot of fea­tures.

There is a frontend for it called backrest that does a lot of things right. It separates the concept of repo which is what you backup to and plan which is what and when you backup.

But I want more, I want to separate plan (when and how) from source (where the data is)

I want to work like this:

1. Create a repo like "this folder here on this computer is a repo called 'foo' and it has a password and whatnot"
2. Create a plan like "every day at 3 AM, backup to 'foo', keep 30 days of backups there, and every week at 2am on sundays, backup to 'bar' which is a remote repo and keep 10 weeks of backups there"
3. Create a source like "the /home/ralsina folder in my notebook"
4. Create a backup which is a combination of a source and a plan like "backup the /home/ralsina folder in my notebook to 'foo' every day at 3 AM"

Fur­ther, I want this to just work on all my com­put­ers as long as I have ssh cor­rect­ly con­fig­ured to al­low it.

My back­up con­troller should log in­to what­ev­er com­put­er the re­po is in and in­stall restic there. Then log in­to the source com­put­er anb in­stall restic there. Then cre­ate the re­po, and when the plan says it's time to back­up, it should log in­to the source com­put­er and run the back­up com­mand there, and then log in­to the re­po com­put­er and run some­thing there if it needs run­ning.

Then I want it to keep logs and no­ti­fy me via go­ti­fy or some­thing if some­thing goes wrong.

Is that too much to ask?

This is a new oc­ca­sion­al se­ries of posts where I will share ideas for pro­grams that don't ex­ist, but should. The goal is to in­spire de­vel­op­ers to cre­ate use­ful tools that can make our lives eas­i­er. Or, more like­ly, to re­mind me about these ideas so I can cre­ate them my­self. Or even more like­ly, to just get them out of my head so I can stop think­ing about them.

Idea 2: A nice web frontend for journald

I do some self host­ing. It's tempt­ing, when you self­-host, to run things as if it was a com­pa­ny's pro­duc­tion set­up. So, there are some who run mul­ti­ple large servers on ku­ber­netes and so on.

Not me, I run a sin­gle SBC with a bunch of dock­er­ized ser­vices.

So, how do I see logs if some­thing goes wrong?

Well, I log to the system's journal, so I can use journalctl to see the logs.

It's just this bit of YAML in your com­pos­er def­i­ni­tion:

  logging:
    driver: "journald"
    options:
      tag: "whatever"

That tags the logs from that con­tain­er with "what­ev­er". So, I can run:

journalctl -t whatever

This tool, journalctl is quite nice, and you can filter by date, grep for things, follow the live logs, and so on. But it's a command line tool, which I like.

But the ¨do it like a re­al prod thing"crowd us­es logstash or some­such, and have a web dash­board for this kind of things.

Well, I should have one of those too, but backed by journalctl

There is one that comes with sys­temd, but it's sort of crap­py, and there is no rea­son for it to be. It's run­ning in the same serv­er where the logs are, it's sim­ple, and it would be a nice lit­tle project to do.

This is a new oc­ca­sion­al se­ries of posts where I will share ideas for pro­grams that don't ex­ist, but should. The goal is to in­spire de­vel­op­ers to cre­ate use­ful tools that can make our lives eas­i­er. Or, more like­ly, to re­mind me about these ideas so I can cre­ate them my­self. Or even more like­ly, to just get them out of my head so I can stop think­ing about them.

Idea 1: A program that can automatically configure complex machine/monitor setups

My desk­top has mul­ti­ple ma­chines shar­ing mul­ti­ple mon­i­tors via one or more HD­MI switch­es. Those mon­i­tors are land­scape or por­trait, some are built in­to lap­tops and oth­ers are ex­ter­nal. Some are di­rect­ly con­nect­ed, some are not. A com­put­er may have be­tween ze­ro and three mon­i­tors con­nect­ed at any one time.

When they are con­nect­ed, in some cas­es, I want to run bar­ri­er to share the mouse and key­board, but with dif­fer­ent con­fig­u­ra­tions de­pend­ing on what mon­i­tors are con­nect­ed to what (say, is this com­put­er us­ing the mon­i­tor at the left of the one for the com­put­er with the key­board and mouse?).

So:

• I want to be able to con­fig­ure the mon­i­tors and their ori­en­ta­tion­s, and the key­board­/­mouse shar­ing, for each ma­chine.
• This should hap­pen au­to­mat­i­cal­ly when I con­nect a mon­i­tor or switch the HD­MI switch.
• It should start/stop/re­con­fig­ure bar­ri­er on each ma­chine.

Of course this does­n't ex­ist and is pretttt­ty hard to do.

For de­tails on how my set­up works now: mon­i­tors and kb­d/­mouse