New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Portable recovery #3206
Portable recovery #3206
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Did I see this change in your other PR (misc stuff) too?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Same comment here - same as in your other PR (#3205)
I miss some documentation here. Is it possible to add it? |
10f25d2
to
b261344
Compare
@gdha I added a short manual @jsmeix the help workflow had bad indentation which I of course fixed, besides that I reduced the whitespace changes. My goal at the moment is to get feedback on this and do some more tests. There will be a second phase of development with further optimisations (for example, I want to see if I can skip the |
Aaargh, now I get also #3168 (comment) because the About dracut not showing errors in the ReaR log, I could redirect the output to properly capture it. |
I did not have a look here (no time yet) Regarding Currently we have in sbin/rear
and we do not change TMPDIR in RECOVERY_MODE
so I think you need to enhance how RECOVERY_MODE is set |
@jsmeix about the dracut not showing errors problem I figured out what is going on: In normal (non-debug) ReaR mode stderr goes to How should we solve this? Maybe extract the "pull last lines from stderr" into a function that can then be used? Or add a I'm not solving this here but I think it is important to keep in mind and fix so that users can get an info about the problem for errors without running ReaR again in debug mode. |
I made the separated |
c0b7a8f
to
b4492e4
Compare
I found a bug in this code, it doesn't work when used from a dist install :-( I'll keep debugging then. |
It's the |
66b2640
to
3a5bf5f
Compare
3a5bf5f
to
05de9d7
Compare
@schlomo
You added the part about PORTABLE mode because of your above But when `chroot /mnt/local dracut' fails with
it fails not because there is no TMP_DIR in the rescue system It fails because there is no TMP_DIR So I think the part about PORTABLE mode in that comment
I think the actual root cause is that
@schlomo In sbin/rear we have currently
where "the rescue/recovery system" means here As far as I understand Or do I misunderstand something and PORTABLE mode is also |
@jsmeix ahhh, good point and good catch. ATM indeed the only use case for portable mode is recovery. I think that I felt that I shouldn't needlessly limit the portable mode to only that, but now that you mention it maybe that was stupid and indeed portable should be limited to recovery only. |
I will do a pull request... |
In sbin/rear set RECOVERY_MODE also in PORTABLE mode, see #3206 (comment) and #3206 (comment)
@schlomo For a quick "smoke test" it should be sufficient
DebugPrint message shows the intended TMPDIR value. |
@schlomo I think for that in
should be enhanced to a more specific test, BUT Perhaps the assumption that a rescue system runs in RAM My homeoffice workstation with a physical disk:
My ReaR "original system" test VM:
My ReaR recovery system VM
Perhaps when the 'findmnt' SOURCE is 'none' |
Add
OUTPUT=PORTABLE
and--portable
command line option to faciliate using ReaR in truly portable mode.The portable archive contains only ReaR, nothing else.
Tested with an OL9 restore via SystemRescueCD
I'll do some more testing both of portable usage and regular usage to ensure that this change doesn't hurt us.
Implements #3190 and should be merged after #3205, where I extracted the unrelated fixes. To review you can simply look at the last commit here.