I can confirm that on my 750Gr3 now the conversion from master-slave to bridge works, and that on L2TP/IPSEC now dail-on-demand/timeout is now respected.
Updated: “*) bridge - fixed hw-offload disabling for Mediatek and Realtek switches when STP/RSTP configured;”
upload speed through router is still 500 Mbit/s slower when STP/RSTP is activated on the Bridge
UPDATE: Nstreme work fine… i add the “Management Protection Key” of the Sextant AP to the SXT station and connect OK with Nstreme… NV2 is not working!
You should never have queue with target 0.0.0.0/0. Such queue can limit traffic for specific IP only on one direction (upload and download is limited only by download limit) because every IP goes under 0.0.0.0/0 so queue does not know which is download and which is upload. For example, traffic between x.x.x.x and y.y.y.y will be always matched by this queue under download since x.x.x.x and y.y.y.y is under 0.0.0.0/0. If you set target as x.x.x.x then queue will know that if source address is x.x.x.x, then it is upload and if x.x.x.x is destination, then it is download.
That’s as expected since hw-offloaded STP/RSTP is not supported on Realtek, Mediatek and ICPlus175D switches: https://wiki.mikrotik.com/wiki/Manual:Switch_Chip_Features#Bridge_Hardware_Offloading. As I understand when you enable STP/RSTP on one of these switches the hw-offload should automatically be turned off, but that didn’t happen before, and now they fixed that.
While it is a worthy mention and any improvement in DHCPv6 support is welcome I feel it is important to remind you that your team is sorely lacking in this area from even a residential router perspective. Look at how OpenWRT / LEDE implements DNSMASQ in LUCI for a valuable interface. We need to be allow per server instance to set the network bits used, aka VLAN11 should always be provider::/56 ending in 11 like provider11::/64, this implies nesting or parent/child functionality that is currently not allowed with your pool system. Additionally, the DHCPv6 server needs to support client address allocation before it is useful. In parallel the SLAAC implementation should not parrot the upstream DNS servers, instead the administrator should be able to set their own. If I use your SLAAC implementation it results in all of my IPv6 DNS queries, which could be a single stack area of my network, not being cached.
I have a strange problem.
PPPoe connectivity to ISP modem.
Local devices can not connect to the web site IP address 185.58.74.135 which hosts for a larger number of sites. It can be checked which pages do not work with reverse ip (for example, https://www.ipneighbour.com)
The ROS 6.39.3 version works ok.
The Ros 6.41 and 6.42RC does not work.
Other sites work normally. I do not know if something else site does not work.
My site does not work either, I have a created a topic for that.
No feedback what’s wrong with pppoe.?
It’s not a problem for a RC release, but a current. Especially at this point, Downgrade is not simply because of the implementation bridge.
This is the main thing about roter, website review. it is not a special possibility. Some hosting does not open, they which have about 100 websites.
My production is at 6.39.3, so it’s not terrible. At work and at home in newer versions.
Tnx liviu2004 on the topic, I thought something was wrong with me.
Best regards.
Sorry for bad English
I tested and have the same problem. Test done with veritas.hr 185.58.74.135. IPv4 only.
RB750Gr3 6.42rc6: Website does not load
Mobile data on my smartphone: Website loads OK
Another shared host website, rtcs.com.my, loads fine.
6.41, sites from https://www.ipneighbour.com/#/lookup/185.58.74.135 list open normally. What error do you see?
The list opens normally, the web pages on that list do not open. For example, the web Page pupilla.hr, boletus.hr …
Just through the Mikrotik pppoe connection is a problem. "Status do not get off of “syn sent”
I have print logs, but the liviu2004 explained everything in his topic.
( I have a backup connection, pppoe on the router ISP, a Mikrotik connected via Lan to that router. . so the company that has to have access to one page has redirected it.)
New RC version released
What’s new in 6.42rc11 (2018-Jan-18 12:42):
*) bridge - fixed IGMP Snooping after disabling/enabling bridge;
*) bridge - fixed allowed MSTI priority values;
*) certificate - do not use utf8 for SCEP challange password;
*) chr - automaticly generate new systemID on first startup;
*) crs326 - fixed possible packet leaking from CPU to switch ports;
*) dhcpv4-server - added “dns-none” option to “/ip dhcp-server network dns”;
*) dhcpv6-client - implement confirm after reboot;
*) filesystem - fixed situations when “/flash” directory lost files after upgrade;
*) ipsec - fixed AES-CTR and AES-GCM support on RB1200;
*) kidcontrol - added initial support for “/ip kid-control” feature;
*) ppp - changed default value of “route-distance” to 1;
*) ppp - fixed change-mss funcionality in some specific traffic (introduced in v6.41);
*) webfig - fixed MAC address ordering;
*) webfig - fixed wireless snooper address, SSID and other column ordering;
*) winbox - added “dhcp-option-set” to DHCP server;
*) winbox - added “insert-queue-before” to DHCP server;
*) winbox - added DHCPv6 client info request type and updated statuses;
*) winbox - show Bridge Port PVID column by default;
*) wireless - fixed nv2 (introduced in v6.42rc);
It’s OK now with rc11. Tried veritas.hr, pupilla.hr and boletus.hr. All loading ok.
I guess this fixed it:
*) ppp - fixed change-mss funcionality in some specific traffic (introduced in v6.41);
Please give a bit more detail and/or workaround for this issue. Can it be fixed by removing change-mss in PPP and adding a change-mss firewall mangle rule?
When will 6.41.1 be released to fix it?
What could have been changed?!
RC9: *) ppp - changed default value of “route-distance” to 1;
RC11: *) ppp - changed default value of “route-distance” to 1;
Indeed it is really time to roll-out 6.41.1
Yes. You can add change mss mangle rules manually to fix the problem, you don’t need to turn off setting in ppp profile.
I have now some new behaviour (RC11) when using when using /IP-firewall-Mangle-Extra-Dst. Address Type.
I used here Address Type “!-local” to exclude the local addresses even it was not on the router itself but directly connected to one port. I have now to include the complete local address range to have it working as before. I use that to decide if a address is in the local network or the WAN and now I have to filter extra on 192.168.88.0/24 to make that decision.
It is now working as in the WiKi manual so I have to do more than before.
UPDATE:
I am back and while investigating why the above did not work any more my router disappeared, not literal but from my network, and I had to restore what was not possible because it wanted a password that was not there and was not needed. [Ticket#2018011822005351]
I went back to 6.40.5, never got a stable 6.41, and build-up from that back to 6.42.RC9 and there also “local” worked again.
The crash was my fault and I was looking at /IP addresses and noticed the line “192.168.88.1/24 192.168.88.0 bridge1” and I removed the “/24” because the Network was stated…bye bye connection and so router and up to a reset config. DON’T TRY THIS AT HOME OR WORK!!
So the workings of “local” changed in RC11 .
[Ticket#2018011822005351] has now been resolved and it was an error on my side that I missed to hit the checkbox “Don’t Encrypt” that time. Potential to avoid this confusing is still there in the graphical interface so that when you miss to click you know what is going to happen.
Yes, I saw that fix, everything is okay now.
One thing has not worked on Winbox for a long time.
If you set Schedule “Start Time = Startup”, Next Run shows ok. If you change your mind and put it for example “Start Time = 00:04:00” “Interval = 1d 00:00:00”. “Next Run” shows some strange time.
The only way to gain proper time is to create a new Schedule or set time in Terminal. The terminal works fine.
Did anyone else noticed a change (6.42.RC11) in behaviour of the Address Type: “local” as described in posting: https://forum.mikrotik.com/viewtopic.php?f=21&t=129034&start=50#p637725?