There is a bug with IPv6 DHCP client over PPPoE (It's only tested with PPPoE). After PPPoE disconnect or router reboot it tries to rebind, but the rebinding is unsuccessful. The PPPoE server is HW router providing PPPoE with IPv6 using DHCPv6 and SLAAC simultaneously. If I manually release the pool, everything is fine and the pool is assigned correctly. If I wait the lease timer to expire, the renew is done successfully.

Same problem here, also with PPPoE. After a reboot, the IPv6 pool doesn't get created, and the DHCPv6 client shows this:
I have to manually hit the "Apply" button (without changing any settings), at which point it binds correctly:

Problem still, i updated to V6.42.3 and this bug still in Server.

The only way i can get it fized is rebooting the Server Router, then when al user reconects all get good for a short time then starts the problem all again.

Versions tested: bugfix - 6.40.8, current releases 6.41 6.41.1 6.41.2 6.42 6.42.1 6.42.2 6.42.3

All get same problems, and all tested server and client sides, i think the bug is in server side, because i not need to reboot the clients only the server roter for the fix.

My hardware: CCR-1036-8G-2S+

6.42.7 bug still present for me too. As @hebertonlp describes, IPv6 works well for a long time. My PPPoE was up for over 24 hours with IPv6 working. Then I rebooted the Mikrotik and left the office. When I came back about 6 hours later, the IPv6 prefix was bound, but clients had no route to any IPv6 addresses. I again rebooted the Mikrotik and immediately logged in to Webfig. Although the PPPoE connection came up immediately, it was stuck on "Rebinding" the IPv6 prefix, but as soon as I hit the "Apply" button (without changing anything) it immediately obtained a prefix, and clients were able to communicate over IPv6.

I'm affected by this bug too, sent a mail to mikrotik support yesterday, no response for now.
Tested with latest stable and latest testing releases

This workaround works for me. Create an on-up script under the PPP profile. This will force a release (and subsequent renewal). This is mine:

Thank you, it seems to work.
Regards

This issue is still there on 6.43.8

This workaround works for me. Create an on-up script under the PPP profile. This will force a release (and subsequent renewal). This is mine:

Code: Select all

:local interfaceName [/interface get $interface name];
:delay 10
:log info "profile-pppoe-isp client up: ipv6 dhcp-client release";
/ipv6 dhcp-client release [find interface=$interfaceName];

Excuse me, can you guide me with this script, I have no idea how to make it right for my config. Do I need anything else in the ppp config or just insert those line of script?

Excuse me, can you guide me with this script, I have no idea how to make it right for my config. Do I need anything else in the ppp config or just insert those line of script?

This script is supposed to go into the ppp-profile for your connection. You can just copy and paste it as is (if you use Webfig or Winbox). If you use the commandline, it would be the "on-up" field (/ppp profile edit <number> on-up).

Excuse me, can you guide me with this script, I have no idea how to make it right for my config. Do I need anything else in the ppp config or just insert those line of script?

This script is supposed to go into the ppp-profile for your connection. You can just copy and paste it as is (if you use Webfig or Winbox). If you use the commandline, it would be the "on-up" field (/ppp profile edit <number> on-up).

thank you for the commandline, it help me for the idea.

Well, two years since this topic was started, but it looks fixed in 6.46beta59:
*) dhcpv6-client - properly update bind time when unused prefix received from the server;
*) dhcpv6-client - properly update IPv6 address on rebind;
*) dhcvp6-client - fixed timeout when doing rebind;

I've tested without the script I had in ppp on-up and it seems to work right.