Thoughts about the "install/upgrade" operations

Core Development Feature Requests
1

Yeah, I know.....

In light of recent events, and the message:

Any errors will be logged to   /var/log/nodered-install.log

Alas that gets a bit BIG when looking for an error, or why it didn't update.

Suggestion:
(And I shall state I don't know the workings of the directories and why they are used.)
But if /var/log is a directory where logs are stored,

Looking in that directory, there are a lot of files AND directories.

Why not make a Node-Red directory and every time the update/upgrade is done, it makes a new file for that date?

Or, if these updates/upgrades (what ever) etc can happen at any time:
Simply break the log file into 24 hour frames. Each day, new log.

That way if someone's machine spits the dummy, but had been working for weeks, they just need to look at that day's log file.

Just a thought.

1 Like
2

If you care to look at the log you will see that it has lines like

Finished : Thu 6 Sep 15:56:19 UTC 2018

***************************************

Started : Wed 19 Sep 12:03:55 UTC 2018

So I think each run is fairly well separated / indicated.
The log file itself is rotated/cleaned up by the default linux logrotate daemon.

3

That is a very inefficient way of logging. The beauty of Linux is that everything has its place in the file system, all the logs go in /var/log/. If you need to look for something specific in a log file use grep:

cat /var/log/nodered-install.log | grep "what_I_am_looking_for"

Screenshot_2018-09-21_09-37-15
Screenshot_2018-09-21_09-37-15932Ă—116 15.8 KB

4

Ghayne,

I wasn't meaning to put the logs in a different log directory, but add a node-red directory in the log directory.

Looking at the log from my RPI today when it died: the file was BIG.
I only thought that the log file being broken to daily (or maybe even weekly) sizes would make error tracing for people easier and if they have to upload the relevant section, it would be less cutting and pasting.

But ok. I accept what you say.

5

DCJ,

Again, I stand corrected.

Just as I mentioned the log seemed very big. But if there are systems to clean/maintain the size: all well and good.

And I am sure if you looked at the log file I posted about the dead RPI after update:
I did search for that line and cut/pasted only the needed part.