[Fixed in 6.34] IPv6 address from Pool stays invalid

Am I doing something wrong?
(real address prefixes changed to documentation prefix for this post)

This problem began when I upgraded my 2011UAS-2HnD from v6.1 → 6.27

I am receiving a /60 from DHCPv6-PD and use it to create pool=ISP

[admin@routerboard] /ipv6 address> /ipv6 dhcp-client print detail
Flags: D - dynamic, X - disabled, I - invalid 
 0    interface=ether6 pool-name="ISP" pool-prefix-length=64 
      prefix-hint=::/60 status=bound prefix=2001:db8:c000:6310::/60 
      expires-after=3d23h37m1s duid="00030001d4ca6d31fd68" 
      add-default-route=yes use-peer-dns=no

I add an address from the pool, but it is flagged as invalid. (LAN is a bridge interface)

[admin@routerboard] /ipv6 address> add address=::1/64 interface=LAN from-pool=ISP advertise=yes eui-64=no

[admin@routerboard] /ipv6 address> print detail
Flags: X - disabled, I - invalid, D - dynamic, G - global, L - link-local 
 5 IG address=2001:db8:c000:6310::1/64 from-pool=ISP interface=LAN 
      actual-interface=LAN eui-64=no advertise=yes

[admin@routerboard] /ipv6 address> .. pool used print
POOL         PREFIX                                      OWNER        INFO       
ISP          2001:db8:c000:6310::/64                       Address      LAN        

[admin@routerboard] /ipv6 address> .. nd prefix print
(none)

Other facts:

  • If I remove this address and re-add it as a static address without specifying the pool, then it becomes active.
  • Even though the address is flagged Invalid, I am still able to ping/access it from the Internet.
  • Devices on LAN side do not get SLAAC.

I’m experiencing the same problem.
IPv6 addresses assigned from pool are invalid and ND doesnt work.

I’ve discovered that if you reboot the Mikrotik with the “invalid” address configured, it becomes valid and works.

I’m trying older version to figure out which rev. was the first to introduce this behavior.

EDIT:
This behavior first appears in 6.26 - In both .26 and .27, if you configure a “from-pool” address, it never removes the Invalid flag. Upon reboot it works properly and is not flagged invalid.

This problem seems to be present even in ROS 6.29.1. Mikrotik, could you please do something about this?

I am seeing this behavior as well.

Confirm problem.

Yes, I encounter this in 5.25. Disappointed that it’s still present in 6.29.

Bug seems to be that the prefix isn’t released to the Pool when the Address is de/reconfigured.

Disable/reenable DHCPv6 Client while Address is enabled also solves the issue (since this clears the Pool and makes the prefix available for the Address again).

6.33.5 here and it appears that the issues are still occurring. Confirm rebooting the router or release and renewing the DCHPv6 address on the router clears the invalid flag.


It also appears there cannot be more than one address from the pool active at any one time. If there are two address added to separate interfaces, one is always invalid. Rebooting or clearing the DCHP lease doesn’t help. Which address works and which one doesn’t can change with no apparent reason

Mikrotik can you please address this?

I cannot confirm that, I have separate IPv6 prefixes on two different networks and both of them work OK.

It is true that it sometimes fails when you first setup the whole thing and then starts working after reboot, as described
earlier in the thread. I remember I had that problem. Of course not good because when you are trying things and they
don’t work you first blame your own efforts.

Please contact Mikrotik support and send a supout file.

Alright. I’ll get you one shortly.

This has been fixed in 6.34

I’m not sure about the RC versions, but the official release works.
Thanks.

I have the same problem in version 6.37.1. My routerboard is a RB750UP.

See this video with the problem and how I solve them: https://dl.dropboxusercontent.com/u/28205339/apagar/IPv6_Wantel.mp4