Restarting jool with a new NAT64 prefix fails

[telmich]

I modified jool.conf and launched jool restart on alpine. It gives the following error:

[11:02] router1.place6:/etc/jool# /etc/init.d/jool restart
 * Loading NAT64 instance default ...
Error: The kernel module returned error 22: Sorry; you can't change a NAT64 instance's pool6 for now.xlator_replace() failed. Errcode -22           [ !! ]
 * ERROR: jool failed to start
[11:03] router1.place6:/etc/jool# 

[11:03] router1.place6:/etc/jool# apk list -I | grep jool
jool-tools-static-4.0.6-r3 x86_64 {jool-tools} (GPL-2.0-only) [installed]
jool-tools-bash-completion-4.0.6-r1 x86_64 {jool-tools} (GPL-2.0-only) [installed]
joold-4.0.6-r1 x86_64 {jool-tools} (GPL-2.0-only) [installed]
jool-modules-lts-5.4.7-r0 x86_64 {jool-modules-lts} (GPL-2.0-or-later) [installed]
jool-tools-4.0.6-r1 x86_64 {jool-tools} (GPL-2.0-only) [installed]
[11:04] router1.place6:/etc/jool# 

[telmich]

It correctly (as expected) works when deleting the current instance first:

[11:04] router1.place6:/etc/jool# jool instance remove default
[11:04] router1.place6:/etc/jool# /etc/init.d/jool restart
 * Loading NAT64 instance default ...                                  [ ok ]
[11:04] router1.place6:/etc/jool# 

The pools are updated correctly:

[11:04] router1.place6:/etc/jool# jool pool4 display
+------------+-------+--------------------+-----------------+-------------+
|       Mark | Proto |     Max iterations |         Address |       Ports |
+------------+-------+--------------------+-----------------+-------------+
|          0 |   TCP |       1024 ( auto) |  147.78.195.251 | 61001-65535 |
+------------+-------+--------------------+-----------------+-------------+
[11:04] router1.place6:/etc/jool# jool pool4 display --help
Usage: display [OPTION...]

      --csv                  Print in CSV format
  -i, --icmp                 Print the ICMP table
      --no-headers           Do not print table headers
  -t, --tcp                  Print the TCP table (default)
  -u, --udp                  Print the UDP table
  -?, --help                 Give this help list
      --usage                Give a short usage message
  -V, --version              Print program version

Report bugs to jool@nic.mx.
[11:05] router1.place6:/etc/jool# jool pool4 display --udp
+------------+-------+--------------------+-----------------+-------------+
|       Mark | Proto |     Max iterations |         Address |       Ports |
+------------+-------+--------------------+-----------------+-------------+
|          0 |   UDP |       1024 ( auto) |  147.78.195.251 | 61001-65535 |
+------------+-------+--------------------+-----------------+-------------+
[11:05] router1.place6:/etc/jool# 

[ydahhrk]

Well, it's true that changing pool6 during a file update is currently defined to be illegal on NAT64 instances.

That's the reason why I defined restart as an instance remove followed by file handle (instead of just file handle) in my init script. This forces the pool6 change at the cost of losing the current instance's state (ie. the BIB).

Presumably, the init script of Alpine is only doing the file handle half. Although this seems to be their script and I don't see their definition of restart anywhere, so... I don't really know where the problem is.

Unless you want me to change the behavior of tweaked pool6 in NAT64 (which will require a discussion because it might yield technical issues), I can't really be of much help here. You might want to try contacting the Alpine maintainer.

[ydahhrk]

This seems abandoned, so closing.

Feel free to reopen if a problem persists.