Bug 2164413

Some additional notes:

satellite-maintain restore procedure does not have "-t" option (which adds the --new-volume-script option when creating the archive, and which misses during extract).


Potential workaround: 
1) run the command:

tar --selinux --extract --file=/path/to/satellite-backup-2023-01-23-14-17-07/pulp_data.tar --absolute-names --overwrite --listed-incremental=/dev/null -M --directory=/ --new-volume-script=/usr/share/gems/gems/foreman_maintain-1.1.8/bin/foreman-maintain-rotate-tar

(add the latest argument there)

2) rename the pulp_data.tar
3) create a dummy pulp_data.tar, like:

touch /tmp/empty
tar cf /PATH/TO/BACKUP/pulp_data.tar /tmp/empty

4) Let the restore run again - it will not stuck on the same user prompt, but extract the dummy file instead; pulp data extracted manually in the right way

(procedure is pending a review, use with caution until KCS is created)

Perspective patch:

--- /usr/share/gems/gems/foreman_maintain-1.1.8/definitions/features/tar.rb.orig	2023-01-25 20:40:20.781292213 +0100
+++ /usr/share/gems/gems/foreman_maintain-1.1.8/definitions/features/tar.rb.new	2023-01-25 20:39:40.311747413 +0100
@@ -51,13 +51,16 @@ class Features::Tar < ForemanMaintain::F
     end
 
     if volume_size
-      split_tar_script = default_split_tar_script
       tar_command << "--tape-length=#{volume_size}"
-      tar_command << "--new-volume-script=#{split_tar_script}"
     end
 
     tar_command << '--overwrite' if options[:overwrite]
-    tar_command << '--gzip' if options[:gzip]
+
+    if options[:gzip]
+      tar_command << '--gzip'
+    else
+      tar_command << '-M --new-volume-script=#{default_split_tar_script}'
+    end
 
     exclude = options.fetch(:exclude, [])
     exclude.each do |ex|

(In reply to Pavel Moravec from comment #2)
> +      tar_command << '-M --new-volume-script=#{default_split_tar_script}'

It must be:

+      tar_command << "-M --new-volume-script=#{default_split_tar_script}"

to eval the variable properly.

One would need to confirm, but i'm not 100% sure the -M is needed when you pass the --new-volume-script flag.

In theory, extract from manpages of tar:
~~~
-F, --info-script=NAME, --new-volume-script=NAME
run script at end of each tape (implies -M)
~~~

I can confirm that the patch from #c2 + #c3 does work well on any backup, either with -t or without the option used. Also, it properly applies the -M --new-volume-script options only to pulp_data.tar - not to gzipped tarbals that cant be split that way.

And indeed, we can even omit the -M option there, since it is implicit from the other one. So the working patch is:

--- /usr/share/gems/gems/foreman_maintain-1.1.8/definitions/features/tar.rb.orig	2023-01-25 20:40:20.781292213 +0100
+++ /usr/share/gems/gems/foreman_maintain-1.1.8/definitions/features/tar.rb.new	2023-01-25 20:39:40.311747413 +0100
@@ -51,13 +51,16 @@ class Features::Tar < ForemanMaintain::F
     end
 
     if volume_size
-      split_tar_script = default_split_tar_script
       tar_command << "--tape-length=#{volume_size}"
-      tar_command << "--new-volume-script=#{split_tar_script}"
     end
 
     tar_command << '--overwrite' if options[:overwrite]
-    tar_command << '--gzip' if options[:gzip]
+
+    if options[:gzip]
+      tar_command << '--gzip'
+    else
+      tar_command << "--new-volume-script=#{default_split_tar_script}"
+    end
 
     exclude = options.fetch(:exclude, [])
     exclude.each do |ex|


It is not very clean patch as we should detect the tape size in a more clever way, but.. at least it works. As I successfully tested all combinations "backup|restore satellite with|without -t option".

Upstream bug assigned to pmoravec

Upstream bug assigned to pmoravec

*** Bug 2121082 has been marked as a duplicate of this bug. ***

Moving this bug to POST for triage into Satellite since the upstream issue https://projects.theforeman.org/issues/36012 has been resolved.

Verified

Verified with:
- Satellite 6.13.0 snap 15
- rubygem-foreman_maintain-1.2.7-1

Test steps:
- mkdir /backup
- satellite-maintain backup offline -t 50M /backup
- satellite-maintain restore -y /backup/satellite-backup-2023-03-20-09-59-46/

Observation:
- No error during restore.

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Important: Satellite 6.13 Release), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2023:2097