[rear-users] ReaR rescue ISO as "live distribution"?

Schlomo Schapiro schlomo at schapiro.org
Thu Mar 14 11:17:04 CET 2019


like Johannes said, yes it can be done.

We would be happy to accept a pull request from you for a "rear rescue" or
"rear explore" workflow that would try to mount as much of the original
system under /mnt/local (the hard coded recovery mount point). Using this
workflow should disable further use of the "rear recover" workflow without
a reboot (to ensure a sane state of the system prior to rear recover).

I think that your idea is a nice extension to the original ReaR use cases
and I believe that other users also will find it useful.

For further and in-detail discussions I would appreciate if you would go
and create a GitHub issue at https://github.com/rear/rear/issues as that
helps us to track the discussion together with your code contribution.

What is important to remember is that the feature will only work reliably
for the same system that the rescue image was created on. That is part of
the design of ReaR.

Kind regards,

On Thu, 14 Mar 2019 at 09:51, Johannes Meixner <jsmeix at suse.de> wrote:

> Hello,
> On Mar 13 17:31 Philippe Andersson wrote (excerpt):
> >
> >> Because the information what filesystems are mounted at what mount
> >> points is available in the var/lib/rear/layout/disklayout.conf
> >> file I think it is not too hard to manually make a bash script
> >> that reads disklayout.conf and only mounts the filesystems.
> >
> > Well, it's certainly possible, but as you said yourself in one
> > of the posts you referenced above, probably not trivial either.
> in
> https://github.com/rear/rear/pull/1091#issuecomment-263819775
> what I meant therein with "that part will get really complicated"
> is that it will get really complicated to cleanly split that part
> out of the whole "recover" workflow, see my "Reason" therein.
> What I think should be not too hard is to make a separated
> script as I described above.
> See
> https://github.com/rear/rear/pull/1091#issuecomment-263638707
> how easy it can be in practice when the tree of filesystems
> is not overcomplicated and one knows the system (then a single
> mount command can be sufficient as in my testing example).
> In particular when you only need to repair one single filesystem
> it should be sufficient to only mount that specific filesystem.
> My example in
> https://github.com/rear/rear/pull/1091#issuecomment-263638707
> was a complete restore of all files and to do that the whole
> tree of filesystems needs to be mounted first (in my testing
> case that whole tree of filesystems was a single filesystem).
> Simply put:
> The simpler your system, the simpler your repair/recovery.
> Kind Regards
> Johannes Meixner
> --
> SUSE LINUX GmbH - GF: Felix Imendoerffer, Jane Smithard,
> Graham Norton - HRB 21284 (AG Nuernberg)
> _______________________________________________
> rear-users mailing list
> rear-users at lists.relax-and-recover.org
> http://lists.relax-and-recover.org/mailman/listinfo/rear-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.relax-and-recover.org/pipermail/rear-users/attachments/20190314/56f7cdaf/attachment.html>

More information about the rear-users mailing list