I think we have isuee with Letsencript the multi dns has problem:
progress: [error] could not resolve '01.abc.xyz,'02.abc.xyz,‘03.abc.xyz’
Can’t resolve the dns when is multiple with single the is no issue
WTF isn’t that a bit too much for the hap ac 3 hardware?
Everything worked fine until version 7.8. Since version 7.8 OpenVPN reconnects after 60-120 minutes.
7.8
*) ovpn - added AES-GCM and multicore encryption support;
*) ovpn - improved server stability;
*) ovpn - improved TLS-related error logging;
Are there any known issues regarding DHCP and mikrotik CPE (wireless bridge) in this version? I have 2 mikrotik devices in CPE mode and the clients connected to them on Ethernet do not receive an IP address anymore. And log on the AP is flooded with messages like “defconf offering lease 192.168.0.2 for 00:1D:EC:06:13:15 without success”
@nfabo
are yours dhcp subnet 192.168.0.x/nn?
Yes it is
Same here. Have to create an up and down script in the ovpn profile to set a route to the connection
:local ip [/ppp active get [find name=$user] address]
:local id [/ppp active get [find name=$user] name]
:local gt [/interface/ovpn-server get [find user=$id] name];
/ip/route/add disabled=no distance=1 dst-address=$ip gateway=$gt routing-table=main scope=10 suppress-hw-offload=no target-scope=10 comment="$gt";
/ip/route/remove [find comment="<ovpn-$user>"];
sry, figured it out. queue tree on bridge introduced the error. did not expect this traffic would leave the bridge…
Working fine for me
No traffic through OpenVPN since upgrading from 7.9.2 to 7.10.1 on CCR2004 
(connection is established but no traffic)
cAP AC
7.8 > 7.10.1 netinstall, keep configuration = no access to the device, lan/wlan are down.
7.8 > 7.10.1 netinstall, apply configuration, can’t save skin, default.json is always 0 kb. After rotaterboard upgrade device does not boot = only blue led is active. And why now all settings pages areas instead “all information in page”, containing just one small box in the middle of the page, where I must open/close sections and where I must scroll right, scroll down/up?
Why instead of fixing broken/adding new functionality developers did change the web UI, so that that is’t almost unusable?
Yes , and it sends escape sequence, not key.
Problem is that F1 is almost universally used as an HELP key for terminal application itself (and have to be disabled)!
Having previous behavior (from V6) was much better and worked from everything immediately.
Beside that, “?” is thing we are used to (Mikrotik, Cisco, juniper, etc… ) and really did not need change…
I think the reason it was changed is that a couple of other keys were VERY counter-intuitive, like Ctrl-V for “hotlock mode toggle”, and ? was interfering with cases where the user actually wanted to enter a question mark.
So someone decided to revamp the key mappings.
I checked four Windows terminal emulators before giving up on finding one that does that. Neither Microsoft’s own Windows Terminal, nor Cygwin’s MinTTY, nor SecureCRT, nor VSCode’s built-in terminal do that. The latter warns you the first time that other parts of the UI might want F1 for their own purposes, but it then offers to take you to the setting to make it send F1 to the terminal as an ANSI escape code instead, if that is your preference.
I consider these the top three free Windows terminal programs, plus the top commercial one.
What is this universal you speak of, then?
What is this universal you speak of, then?
For example mate terminal, previous gnome terminal as I remember…
but that is not a point, and no need to defend anything.
Decision has been made for whatever reason, so we have to get rid of old habits and get new ones…
I assumed “Windows” in my prior answer because that’s the only platform where I’d expect terminal behavior to be so broken. Every Unix, BSD, macOS, and Linux terminal should handle F1 as RouterOS 7 now expects it to for the very reason I gave above: ANSI X3.64 nailed the basics down in the 1970s, and we got multiple real implementations of ANSI-compatible terminal function key handling by the mid 1980s.
I don’t have a Mate VM laying around to test for this brokenness. The closest I have is an old Ubuntu 18.04 LTS VM, which runs Gnome Terminal 3.28.2. It doesn’t pull up the program’s help on F1, it sends it through to the running application.
To test this, I installed Midnight Commander, which does have an internal “help on F1” feature, and pressing F1 through Gnome Terminal did pull up the MC help, not the Gnome help, indicating that the terminal didn’t eat F1 as you claim Mate’s does.
Either fix your terminal program or get a better terminal.
So since i updated to the 7.10.1 software my RB5009UPr+S+IN router has decided to simply stop working on WAN and not link up at all regardless of any change of config to the ETH1 or SFP1, a full reset is the only thing that helps. After a full reset it works flawlessly for a while and then with no warning or error it simply drops WAN ETH link and stops working until I again do a full reset, again regardless if i use SFP or ETH1.
With bare minimum settings change from pure stock settings it still happens. This did not happen in previous software, ever.
Some kind of bug in 7.10.1?
OSPF MD5 problem, if I disable authentication it reaches full adjacency.
RouterOS 6.49.8 configuration that worked previously:
/routing ospf area
add area-id=0.0.0.10 name=site3
/routing ospf instance
set [ find default=yes ] router-id=100.127.255.10 use-dn=no
/routing ospf area range
add area=site3 range=192.168.10.0/24
/routing ospf interface
add authentication=md5 authentication-key=secret dead-interval=10s hello-interval=1s interface=vpn network-type=point-to-point
add interface=vlan1 passive=yes
add interface=vlan666 passive=yes
add interface=vlan667 passive=yes
/routing ospf network
add area=backbone network=100.127.252.36/30
add area=site3 network=100.127.255.10/32
add area=site3 network=192.168.10.0/24
Router OS 7.10.1 configuration which does not work:
/routing ospf instance
add disabled=no in-filter-chain=ospf-in name=ospf_v2 router-id=100.127.255.10 use-dn=no
/routing ospf area
add disabled=no instance=ospf_v2 name=backbone_v2
add area-id=0.0.0.10 disabled=no instance=ospf_v2 name=site3_v2
/routing ospf area range
add area=site3_v2 disabled=no prefix=192.168.10.0/24
/routing ospf interface-template
add area=backbone_v2 auth=md5 auth-id=1 auth-key=secret cost=10 dead-interval=10s disabled=no hello-interval=1s \
networks=100.127.252.36/30 type=ptp
add area=site3_v2 disabled=no networks=100.127.255.10/32,192.168.10.224/28,192.168.10.240/28 passive
It immediately starts working when I remove ‘auth’ from the point-to-point interface template (and disable it on the upstream).
Hex S after upgrade to ROS 7.10.1, Check Gateway with arp, bfd or ping does not work and it result mark-routing in mangle not working unless I set Check Gateway = none
You cannot have “check gateway” on a PPPoE link. And it makes no sense anyway, as the PPPoE protcol is performing that action by itself, and the interface will go down when the link fails, taking all routes configured like that with it.