mcusr / fb Goto Github PK

What I write as per today.

• DebugOutput.md
• Governor.md
• DelegateSkeleton.md
  Implement a major system change, change the name from DailySnapshot to HourlySnapshot, even if it happens every second hour. It is more to give and idea about the interval, than being precise.

both services part and and commandline tools
only log when in a service mode, or maybe we log to another log?
The backups we have are the evidence of the jobs done

It would be good to leverage upon it.

So, it is supposed to work out of the box from the pre-commit step.

to use it from the Makefile in the root directory would also be nice.

I want to see which folders there werent any files in, and
which folders that were successfully backed up in the
journal, in the notifications.

A bypass of password requiring sudo is tested, and done, from a service

I just wanted to add here, that though for now, the tar commands require sudo for running properly, and given that the scripts are run from the command line, the issue with being prompted with a password when you run it, is bearable.

Not so much for services, so, since I need root rights to execute tar, I'll probably give instructions as to how to set the setuid bit to root for the scripts in question, and configure the services to run with elevated rights, as a way to bypass, if you really really want to keep your sudo password.

For me, I see ChromeOs, as a rather large IP firewall, as my container is very safe from drive by attacks, and such, and I haven't set it up as an SSH server, with port forwarding, so I think having someone executing code, or even execute code with raised privileges to be less of an issue.

Now this doesn't cover me from downloading stupid, or malicious stuff or anything, but I don't see having the setuid bit on scripts that backs up to add to the problem, with the bit set, goes write protection, so, if someone wants to do a backup with elevated rights, from a malicious piece of code, then be my guest, this scheme, if I can make it work, will make the software somewhat safer, as it then can execute flawlessly, without requiring a passwordless sudo.

Dokumentere bruk av sudo, og hva maa gjoere hvis ikke sudo passord fri, og hva med services?

It saves a lot of notitifications.

• Investigate how I can best use shellcheck, and avoid error codes when things are okay.
• leave logs somewhere
• I'll run the runner from my makefile.

dieIf

services and everything. Stil only the copy script.

See to that it works as it should

• When there are no errors
• when there are errors, that we keep enough info in the log.

So, the number of backup sets, must be less or equal to, not
less than. and this is about HourlySnapshot.backup.sh

I need to read the install script more.
Start copying over the whole .bashprofile, or edit carefully, placing paths at the end.

• manager
• hasExcludeFile

• The sudo issue the fix for the sudo issue.
• Document canonical names list for vars.
• Update documents to reflect the new scheme names

• Use setbuf on the tar command when in CONSOLE mode.
• can I make setbuf work with sudo?

HourlySnapshot, and the services.

For completeness.

I'm not sure, maybe use some local log file? All the time, we don't need to do this by
journalctl, especially since we're watching the output in the console, and can just re-run the command, or pipe to a file.

Figure out which system areas afflicted for starters, and document!

Do both fbsnapshot and fbrestore work like they should?

I guess I should have the same machinery for parsing command line arguments as in OneShot.backup and OneShot.restore.

-[X] When there is nothing to back up.
-[ ] When there are stuff to back up.

I need to do this, and not let it take so long,
because this needs to be concluded before I have the full
overview over restore and backup rotation.

mayb it works now, and how can I see any logs??

Not concerning sudo, but that the general implementation works with exclude files, that it do rotate backups as it should and so forth.

The whole thing.

The prerequisites. concerning install.
I should leave comments while in the on google drive by test.

I'll either clone or download the zip file. Though, cloning lets me test the stuff before I create the tag and the release!

Then I'll run the install script.

I should have several services with the script installed.

Try it fully out as how it is to use by now, maybe make an accompanying script. to install 'jobs'

Adjust whatever documentation needs to be adjusted.

more portable, and less prone to break somewhere.

So, it is best to see that no backups are made because of
nothing to back up in the journal.
Therefore.. I'll implement that in services.

All cli error handlers write to stderr.

With trap, ,and timers, for echoing out more dots.