/interface vxlan add group=224.0.0.188 interface=bridge-lan name=vxlan-iot port=8472 vni=123
/interface vxlan add mtu=1400 name=vxlan-iot port=8472 vni=123
/queue type
add cake-atm=atm cake-bandwidth=50.0Mbps cake-flowmode=dual-srchost cake-overhead=46 kind=cake name=cake-out
add cake-atm=atm cake-bandwidth=100.0Mbps cake-flowmode=dual-dsthost cake-overhead=46 cake-wash=yes kind=cake name=cake-in
/queue tree
add bucket-size=0.002 name=wan-in packet-mark=wan-in-pk parent=global queue=cake-in
add bucket-size=0.002 name=wan-out packet-mark=wan-out-pk parent=global queue=cake-out
The same issue with wireguard/Upgraded from 7.1.1 and my wireguard connections no l longer worked. No connection was being made and disabling and reenabling the interface had no effect. Rolled back to 7.1.1 and all was well again. Literally nothing in the config changed between both versions.
EDIT: realized I didn't state what platform I am using. I am running an RB5009.
*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled;
FWIW:Upgraded from 7.1.1 and my wireguard connections no l longer worked. No connection was being made and disabling and reenabling the interface had no effect. Rolled back to 7.1.1 and all was well again. Literally nothing in the config changed between both versions.
EDIT: realized I didn't state what platform I am using. I am running an RB5009.
*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled;
*) l3hw - added HW offloaded FastTrack support for inter-VLAN routing;
When you have a device with more than the 16MB flash found on low-end devices, always partiton the flash in 2 partitions so you can copy your existing install to the other partition before doing any upgrade.Short question, how do you roll back? Netinstall or just dumping a old npk on the device?
Different architecture. Someone else with a 5009 is reporting the same.
FWIW:
Hex on 7.1.1
mAP on 7.2rc2
wireguard active between both and running
Console has crashed; please log in again.
Thank you thank you thank you!!RouterOS version 7.2rc2 has been released "v7 testing" channel!
*) firewall - improved system stability when using address lists (introduced in v7.2rc1);
Awesome! I just fired up 7.2rc2 again and disabled my WireGuard queue and guess what? It works! So what about queues has broken things? I’m using a fq_codel queue.RB5009, my queues seem completely broken, I had to disable to get any packets through.
?*) winbox - added ZeroTier support;
Reading is a real art.......Is there going to be webfig or WinBox support for ZeroTier instead of just cli?
So brave, I am waiting for 7.2 stable before turning my CCR1009 into a rocking wireguard machine.On my CCR1009 --- Just loaded 7.2rc2 and that went very smoothly
so far everything is working as I expect ---- CPU is stable .... WireGuard is functioning and to my surprise its a bit faster now
Holy crap that was fast!Hotfix for queue processing should be available now.
What's new in 7.2rc3 (2022-Jan-28 16:33):
*) bridge - fixed filter and NAT "set-priority" action;
*) queue - fixed traffic processing (introduced in v7.2rc2);
Yes, at least it will decrease CPU utitilization a lot.So brave, I am waiting for 7.2 stable before turning my CCR1009 into a rocking wireguard machine.On my CCR1009 --- Just loaded 7.2rc2 and that went very smoothly
so far everything is working as I expect ---- CPU is stable .... WireGuard is functioning and to my surprise its a bit faster now
(1) This change I have questions about and to me it could be significant!
*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled;
This if not in place before should really speed up performance ??
As i still had connection through wifi on rb4011, i was able to connect to webfig and did the system/package/stable/download and install. after 7.1.1 i downloaded the rc1 from https://download.mikrotik.com/routeros/ ... c1-arm.npkShort question, how do you roll back? Netinstall or just dumping a old npk on the device?
[admin@MikroTik] > :put [ :terminal inkey timeout=60 ]
121
[admin@MikroTik] > :put [ :terminal inkey timeout=60 ]
60940
Clearly you need the OFF bridge config schema and dont forget to use safe mode ;-) ( viewtopic.php?t=181718 )7.2rc3 completely destroyed my RB5009 !
UPDATE2 : Managed to get my RB5009 back to 7.1.1 and I think I'm going to stay there for a while ;-) Enough excitement for a Friday evening ;-)
Was that new for you? For me that has happened on every reboot (and of course a reboot is part of an upgrade) since I have been using v7.Ipsec identities lost after upgrade from 7.2rc1 to 7.2rc2 and again after upgrade from 7.2rc2 to 7.2rc3 (SUP-60031).
Not new, but the problem has been acknowledged/reproduced on Jan 11th and "hopefully fixed in the next update" -- this is the next update :)Was that new for you?
IPv6 firewall continues to be broken with queues enabled in any 7.1+*) queue - fixed traffic processing (introduced in v7.2rc2);
I will go at it again and update my lab RB5009 to 7.2RC3 and see what happens this time ;-)jvanhambelgium, please generate a supout.rif file while you are at it. Without any usable information, it is impossible for us to figure out what went wrong.
Ok this time things went much more smooth. As precaution, I disabled IPv6 (because that cause some issues last time preventing PPPoE connection to the ISP)I will go at it again and update my lab RB5009 to 7.2RC3 and see what happens this time ;-)jvanhambelgium, please generate a supout.rif file while you are at it. Without any usable information, it is impossible for us to figure out what went wrong.
And generate supout.rif if needed.
Wireguard & ZeroTier work equally as good as they worked in 7.1.1 for me at least. Not using any other types of "vpn"Severe problem in openvpn
openvpn tcp and udp has problem.
im test android and windows client os
Severe problem in openvpn
openvpn tcp and udp has problem.
im test android and windows client os
--I have some difficulties to have this working.Code: Select all*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled; *) l3hw - added HW offloaded FastTrack support for inter-VLAN routing;
On which devices should this work ?
RB4011 ? RB5009 ?
Many thanks !
--Yes, at least it will decrease CPU utitilization a lot.(1) This change I have questions about and to me it could be significant!
*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled;
This if not in place before should really speed up performance ??
But as per my tests (7.2rc2 / 7.2rc3) on both RB4011 and RB5009, it does not work :
IPv6 on RB5009 is broken beyond being usable.2) Major issues with IPv6 in certain scenarios. It seems Linux based hosts (Synology for example) with everything standard, MTU 1500, etc are seeing 25-50% packet loss. On a 10G or 1G link download is around 250Mbps and upload around 25Mbps. The same on macOS 1G link I get gigabit both ways.
That's usual MT standard for new devices. RB4011 had many severe issues when it was new, taking almost 2 years of ROS updates to get everything working.I'm shocked at the quality of the RB5009 :(
"queue - fixed traffic processing" is MT speak for fixing queues completly broken in a way not moving any traffic at all.What's new in 7.2rc3 (2022-Jan-28 16:33):
*) queue - fixed traffic processing (introduced in v7.2rc2);
Works good for me, see viewtopic.php?p=909126#p909126I have some difficulties to have this working.Code: Select all*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled; *) l3hw - added HW offloaded FastTrack support for inter-VLAN routing;
On which devices should this work ?
RB4011 ? RB5009 ?
Many thanks !
[admin@rb5009] /queue> export
# jan/29/2022 11:39:40 by RouterOS 7.2rc3
# software id = 56R5-PRTF
#
# model = RB5009UG+S+
/queue simple
add max-limit=30M/200M name=limit-win3-upload target=192.168.2.65/32 time=7h-22h,sun,mon,tue,wed,thu,fri,sat
Thank you for your report, good news then I would say.Works good for me, see viewtopic.php?p=909126#p909126I have some difficulties to have this working.Code: Select all*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled; *) l3hw - added HW offloaded FastTrack support for inter-VLAN routing;
On which devices should this work ?
RB4011 ? RB5009 ?
Many thanks !
And no problems with my Wireguard-Peers
[admin@rb5009] /ip/firewall/filter> export
# jan/29/2022 11:47:34 by RouterOS 7.2rc3
# software id = 56R5-PRTF
#
# model = RB5009UG+S+
/ip firewall filter
add action=fasttrack-connection chain=forward comment="TEST FASTTRACK" connection-state=\
established,related hw-offload=yes
add action=accept chain=forward comment="TEST FASTTRACK" in-interface=vlan10 out-interface=vlan20
add action=accept chain=forward comment="TEST FASTTRACK" in-interface=vlan20 out-interface=vlan10
[admin@rb5009] /interface/bridge> export
# jan/29/2022 12:23:18 by RouterOS 7.2rc3
# software id = 56R5-PRTF
#
# model = RB5009UG+S+
/interface bridge
add frame-types=admit-only-vlan-tagged name=bridge1 protocol-mode=none vlan-filtering=yes
/interface bridge port
add bridge=bridge1 frame-types=admit-only-vlan-tagged interface=sfp-sfpplus1
/interface bridge vlan
add bridge=bridge1 tagged=bridge1,sfp-sfpplus1 vlan-ids=2
add bridge=bridge1 tagged=bridge1,sfp-sfpplus1 vlan-ids=10
add bridge=bridge1 tagged=bridge1,sfp-sfpplus1 vlan-ids=20
add bridge=bridge1 tagged=bridge1,sfp-sfpplus1 vlan-ids=30
[admin@RB5009] > interface/ethernet/switch print detail
Flags: I - invalid
0 name="switch1" type=Marvell-88E6393X mirror-source=none mirror-target=none mirror-egress-target=none l3-hw-offloading=yes
Is set to no on my rb5009
There is a switch property l3-hw-offloading, which I'm not sure only setting offloading for L3 VLAN routing only or also for L3 fasttrack.
I had to set to yes to get everything working. Mine was set to no, can't remember if this is the default or if it was me.
[admin@rb5009] /interface/ethernet/switch> print
Columns: NAME, TYPE, L3-HW-OFFLOADING
# NAME TYPE L3-HW-OFFLOADING
0 switch1 Marvell-88E6393X no
Totally agree. Just lack one PoE-out port, as we have on the RB4011, to power-up a LtAP gateway for example :)Once this all runs stable enough for production, this makes the RB5009 a perfect device combining router/coreswitch in one tiny box for branch sites.
So I understand you tested with tagged ports.However, with v7.2rc3 I have fasttrack forward rules with hw offload working on RB5009.
Did not find time for propper testing yet, but a quick ipperf run between routed VLANS on SFP+ shows >6GB/s with less than 50% total CPU load.
HAP ac^3 (RBD53iG-5HacD2HnD) turns off successfully. Thank you!*) arm - fixed "shutdown" command on hAP ac^2;
I too am awaiting a fix for this. my RB3011 idles at 9% CPU which is way higher than the RB2011IUAS I have at another site which only has a 600Mhz single core with similar traffic loads.. I would expect the dual 1.4Ghz core to be idling quite a lot lower than it is.Still CPU load not properly optimized for RB3011.
-"queue - fixed traffic processing" is MT speak for fixing queues completly broken in a way not moving any traffic at all.
It is really bothering how such a severe bug can make it into an rc2 release without being discovered.
How is it possible to miss a completly broken core feature in rc testing? With broken meaning not working at all?
(c) Lewis Carroll, "Through the Looking-Glass""Or else it doesn't, you know. The name of the song is called 'Haddocks' Eyes'."
"Oh, that's the name of the song, is it?" Alice said, trying to feel interested.
"No, you don't understand," the Knight said, looking a little vexed.
"That's what the name is called. The name really is 'The Aged Aged Man'."
"Then I ought to have said 'That's what the song is called'?" Alice corrected herself.
"No, you oughtn't: that's quite another thing! The song is called 'Ways And Means': but that's only what it's called, you know!"
"Well, what is the song, then?" said Alice, who was by this time completely bewildered.
"I was coming to that," the Knight said. "The song really is 'A-sitting On A Gate': and the tune's my own invention."
[admin@r1] > /ip/firewall/service-port/disable [find]
failure: module udplite is built-in and cannot be individually disabled
@ Mafioso Please read through this and tell us where Torch is notedTorch for IPv6 still not working. Can anyone confirm?
Look I think it's fair game to comment if the docs don't indicate something is broken. See https://help.mikrotik.com/docs/display/ROS/Torch could just say "IPv6 not support in V7" (and removed when it is). So the fact this isn't documented is a bug IMO.@ Mafioso Please read through this and tell us where Torch is notedTorch for IPv6 still not working. Can anyone confirm?
If you got this, far then clearly its not a case of : not having two clues and being lazy, - we can drop the lazy part. ;-P
Well, the problems I had with routing in 7.2rc1 are also not mentioned and they are fixed. Unless you mean that "router hangs when showing routes" is fixed by:There's no fix that mentiones torch, why would you expect it to work?
.*) route - improved routing table print speed;
Probably, yes. When you need to ask, then not by you.Here's another config question: Is it possible to make a mikrotik box into a "Bump on the wire"?
https://apenwarr.ca/log/20180808
Even devs wrote numerous times they make alot of silent changes even on big fixes and dont always write it in changelog..@ Mafioso Please read through this and tell us where Torch is notedTorch for IPv6 still not working. Can anyone confirm?
THEN copy and paste it into excel or any document and do FIND/SEARCH for the word Torch, just in case you missed it doing in manually.
What's new in 7.2rc2 (2022-Jan-28 11:00):
If you got this, far then clearly its not a case of : not having two clues and being lazy, - we can drop the lazy part. ;-P
Whenever I find someone defending Mikrotik by calling the user lazy, unaware tec it is always a forum guru.Look I think it's fair game to comment if the docs don't indicate something is broken. See https://help.mikrotik.com/docs/display/ROS/Torch could just say "IPv6 not support in V7" (and removed when it is). So the fact this isn't documented is a bug IMO.
@ Mafioso Please read through this and tell us where Torch is noted
If you got this, far then clearly its not a case of : not having two clues and being lazy, - we can drop the lazy part. ;-P
@ Mafioso Please read through this and tell us where Torch is notedTorch for IPv6 still not working. Can anyone confirm?
What's new in 7.2rc2 (2022-Jan-28 11:00):
This could be shorter.
If you got this, far then clearly its not a case of : not having two clues and being lazy, - we can drop the lazy part. ;-P
Well PPTP is safer than IPIP or GRE, which don't have any red error if used without IPSec... There are lot of way to configure a Mikrotik in unsafe ways, so picking on PPTP seems arbitrary since it's entirely possible the traffic inside may be encrypted thus PPTP wouldn't be particular unsafe.it is unusable this way
even if pptp is disabled, and only used when needed
the red color always suggest that there is error in interfaces ... but, no, no error
at least, change color to blue, like warnings in LOG
Ok, i triedNot sure but this may fixable (or should be at least) by using /system/device-mode. I can see perhaps in mode=home providing this kind of error, but mode="enterprise" the message ideally should be suppressed. The specific behavior of using device-mode set pptp=yes is undefined in the docs however, but seemingly if device-mode allowed PPTP, it's should not be an error.
I have also seen such red error message when configuring a tunnel with IPsec PSK (instead of using certificates). But it has gone away later.Well PPTP is safer than IPIP or GRE, which don't have any red error if used without IPSec... There are lot of way to configure a Mikrotik in unsafe ways, so picking on PPTP seems arbitrary since it's entirely possible the traffic inside may be encrypted thus PPTP wouldn't be particular unsafe.
to correct my self7.2rc3 broken IPv6 SNMP
SNMP['/usr/bin/snmpbulkwalk' '-M' '/opt/librenms/mibs:/opt/librenms/mibs/mikrotik' '-v2c' '-c' 'COMMUNITY' '-OQXUte' 'udp:HOSTNAME:161' 'IPV6-MIB::ipv6RouteTable']
IPV6-MIB::ipv6RouteTable = No Such Object available on this agent at this OID
7.1 working as expected
SNMP['/usr/bin/snmpbulkwalk' '-M' '/opt/librenms/mibs:/opt/librenms/mibs/mikrotik' '-v2c' '-c' 'COMMUNITY' '-OQXUte' 'udp:HOSTNAME:2001' 'IPV6-MIB::ipv6RouteTable']
IPV6-MIB::ipv6RouteDest[2000:0:0:0:0:0:0:0][3][2147483833] = 2000:0:0:0:0:0:0:0
Weird, checked it again and now l3-hw-offloading is no and I can't set it to yes neither
No idea how that worked before, I fiddled around a lot..
However, with v7.2rc3 I have fasttrack forward rules with hw offload working on RB5009.
Did not find time for propper testing yet, but a quick ipperf run between routed VLANS on SFP+ shows >6GB/s with less than 50% total CPU load.
Compared to <3GB/s with 80% total CPU load without L3 hw offload.
/ip/firewall/connection print
But it is working fine on 7.2rc3 on my RB4011to correct my self
no, it is NOT working on CHR, 7.1 & 7.2 RC both broken on CHR
Don't you think it would be a good idea to work towards a 7.2 release that is feature complete relative to 6.49.2 and has most of the visible bugs fixedThe reason why RB5009 received a huge inter-VLAN routing performance boost in 7.2rc3 is FastPath / FastTrack support by vlan-filtered bridges. Routing is still done by the software (CPU), but now it is going via Fast Path, and RB5009 has a fast CPU.
yes, when i tested, i tried only CHR on RCBut it is working fine on 7.2rc3 on my RB4011to correct my self
no, it is NOT working on CHR, 7.1 & 7.2 RC both broken on CHR
I tried it with 7.2rc3 and it works OK for me. I see that it returns the routing table data when there is any, and it displays that error message whenever there is no IPv6 route.yes, when i tested, i tried only CHR on RC
But it is working fine on 7.2rc3 on my RB4011
that was a reason for confusion
devices stayed on 7.1, CHR on RC
Don't you think it would be a good idea to work towards a 7.2 release that is feature complete relative to 6.49.2 and has most of the visible bugs fixed
that were introduced during the v6 to v7 transition?
Then the next step could be a 7.3 which adds all kinds of nice new features. But at least the users could "all" upgrade to v7 and/or have a fully functional
RouterOS again on their devices that only work with v7.
Right now it appears that more attention is going towards nice new features and speedups, but known unimplemented features are left behind.
So, faced same issue, FastTrack does not work for untagged ports in a VLAN-filtered bridge.So I understand you tested with tagged ports.However, with v7.2rc3 I have fasttrack forward rules with hw offload working on RB5009.
Did not find time for propper testing yet, but a quick ipperf run between routed VLANS on SFP+ shows >6GB/s with less than 50% total CPU load.
I'm scared untagged ports would prevent fasttrack to work for them...
Yeah, but several of those bugfixes are in features that were added in v7.The main focus is on bugfixes and v6 feature completion. According to the rc2 changelog, there are 82 changes (except Winbox): only 3 of them are new features, and 79 bugfixes or improvements.
How about getting “ PIM-SM” working! 🧐The main focus is on bugfixes and v6 feature completion. According to the rc2 changelog, there are 82 changes (except Winbox): only 3 of them are new features, and 79 bugfixes or improvements.Don't you think it would be a good idea to work towards a 7.2 release that is feature complete relative to 6.49.2 and has most of the visible bugs fixed
that were introduced during the v6 to v7 transition?
Then the next step could be a 7.3 which adds all kinds of nice new features. But at least the users could "all" upgrade to v7 and/or have a fully functional
RouterOS again on their devices that only work with v7.
Right now it appears that more attention is going towards nice new features and speedups, but known unimplemented features are left behind.
Yes, I hope the focus will be on that, at least for the moment. So we can have a single version again. And don't have to warn other users against blindly trying "upgrade".As soon as v7 is a no-brainer-update with a matching v6 stability, noone would like to stay on v6 for any reason. v6 could be shut-down and we would no longer talk a second about v6.
So, faced same issue, FastTrack does not work for untagged ports in a VLAN-filtered bridge.
So I understand you tested with tagged ports.
I'm scared untagged ports would prevent fasttrack to work for them...
Works as soon as ports are tagged.
SUP-73092 opened accordingly.
nope, devices working as it should rb760, 3011 ...So maybe you don't have IPv6 configured on that CHR.
I see. Thanks for the details.RB5009 does NOT support l3-hw-offloading (List of supported devices).
Setting hw-offload=yes for FastTrack firewall has a recommendatory meaning (i.e., "please offload if you can"). The actual HW offloading state of FastTrack connections appears in the connection list (H flag):
Completly agree regarding the interfaces available on RB5009, anything above 2.5GB intra VLAN routing with decent CPU load for NAS etc. at ether 1 is good enough.P.S. Marvell 88E6393 switch chip used in RB5009 is capable of offloading a very limited amount of FastTrack connections, but the feature is not implemented into RouterOS yet. Anyway, if you need hardware-accelerated near wire-speed routing, I suggest looking at CCR2116.
This would be of great value indeed. ZeroTier should not be missing on all Mikrotik routers. Mikrotik you are already doing a wonderful job do your best and implement ZeroTier on all routers starting with mipsbe.Still no ZeroTier for mips...
With ac, please.... or release new mAP with little ARM CPU. 😉😁
My understanding of the way Zerotier licensing works for hardware vendors is that they have to pay a license fee for every single device that may run ZeroTier. If this is indeed the case in MikroTik's situation, they would have to pay a fee to Zerotier for every single mipsbe architecture device that has ever been sold by MikroTik, which likely includes many old RB4xx and 7xx models that were discontinued years ago and probably do not even work anymore.This would be of great value indeed. ZeroTier should not be missing on all Mikrotik routers. Mikrotik you are already doing a wonderful job do your best and implement ZeroTier on all routers starting with mipsbe.
Are you sure? ZT makes money off the controller – more than one admin, an end customer pays. More devices ZT is on, more potential customers. Mikrotik's volume would seem to feeds their Freemium pricing model (similar plan as Tailscale). AFAIK, nothing stop anyone from deploying ZT on Linux, so their bet is "people are lazy" and they'll take the bet enough organizations are willing to $49+/month.My understanding of the way Zerotier licensing works for hardware vendors is that they have to pay a license fee for every single device that may run ZeroTier.This would be of great value indeed. ZeroTier should not be missing on all Mikrotik routers. Mikrotik you are already doing a wonderful job do your best and implement ZeroTier on all routers starting with mipsbe.
Zerotier has a manual on how to cross compile their product for different platforms. See the github https://github.com/zerotier/ZeroTierOneHopefully things like ZeroTier can be made to run as containers so there is no need for involvement by MikroTik and more than one such supplier can compete in this field.
It requires a little more refinement of the containers, maybe some way to edit configuration of a container from inside the RouterOS configuration system (winbox, webfig, terminal).
Then ZeroTier just has to setup a cross-compilation platform for all architectures they want to support.
Same experiences here including no response to my Ticket either :-(Does 7.2RC3 address SFP+ issues such as those called out in SUP-68278 which cause SFP+ links at 10G to flap every 5 minutes or so? This showed up in 7.1, and the only current solution is to force interfaces to 1G, which is obviously non-ideal. I can’t even get a response to the support ticket after over a month of waiting.
I'm not sure, but I recall reading about this before. I just had another look at the BSL that ZeroTier uses, MikroTik would not seem to qualify to include it in RouterOS free of charge as RouterOS itself is a commercial product not released under an open source license. So MikroTik would probably had to reach some kind of arrangement with them for a commercial license for ZeroTier in order to include it as an optional package. I remember reading that including it in a commercial hardware device was handled by a per device fee (this was a long time ago, when I first looked into ZeroTier for something else), but I don't recall where I saw that now. If it is the case, it might explain why they only released for arm and arm64.Are you sure? ZT makes money off the controller – more than one admin, an end customer pays.
Anyone having RADIUS/dot1x issues?
I have three User Manager instances (replicated) between my router and two switches. Most devices are now not responding and getting a timeout instead, I see responses to the EAP challenges but they don't authorize.
and, after all ...Ok so it is something different than "does not work on CHR" because for me it DOES work on CHR.
But I tried it with a static route, maybe that is different.
What I mean is that instead of working on stability and completeness, work is done on new features.I wouldn't call that "playing", those are features that many folks will like.
Probably a step forward ipv6 fasttrack we hope.
Thank you very much for this long waited feature.*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled;
I updated a CCR2004-16G-2S+ to 7.2rc3; after the 3rd reboot, I lost all 4 wireguard interfaces. Rebooted again and they were still missing. Loaded a backup and everything is ok again.
I would describe 7.2rc3 on my CCR2004 as 'weird'. Certain clients such as iPhone could not load images across many websites, such as twitter. Rebooted, nothing. Rolled back to 7.1.1.
Unsure how to provide more diagnostics, nothing I tried seemed to make a difference.
What about random reboots by watchdog on ccr2004 ?After upgrading to latest Beta on my CCR2204 I can reach 10 Gbit/s in inter-vlan filtering.Thank you very much for this long waited feature.*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled;
I no longer want to sell my CCR2004 ...
It works for me... (netwatch, route with ping check, monitoring router using ping). RB4011.ticket SUP-67221, still present here, about issue on icmp after 2/3 days of operation. It stops the output of icmp, so the icmp check route fails on all route, the router is not able to ping outside host, the router itself responf to ping from outside.
It seems a problem with mellanox connectx5 cards on x86 and CHR with passtrhough.It works for me... (netwatch, route with ping check, monitoring router using ping). RB4011.ticket SUP-67221, still present here, about issue on icmp after 2/3 days of operation. It stops the output of icmp, so the icmp check route fails on all route, the router is not able to ping outside host, the router itself responf to ping from outside.
The CCR2004 is no longer rebooting.What about random reboots by watchdog on ccr2004 ?
We are still getting at stable 7.1.1
Are you using it as ppp server ?The CCR2004 is no longer rebooting.What about random reboots by watchdog on ccr2004 ?
We are still getting at stable 7.1.1
If you are using vlans, speed at least doubled (I could not measure as it saturates my network).
CPU utilization is down, so it cannot harm on the reboot side.
Have fun!
In this release, v7.2rc3 they added winbox for ZeroTier:Is there going to be webfig or WinBox support for ZeroTier instead of just cli?
Despite not being the release notes*, good news is I see ZeroTier in web UI/"webfig" in v7.2rc3 too.*) winbox - added ZeroTier support;
Please don't reply to spammers that just repeat a random post from the thread to add their own advertisement once it gets accepted.Is there going to be webfig or WinBox support for ZeroTier instead of just cli?
Likely right. I'm actually annoyed by any footers in forums, so I ignore them.Please don't reply to spammers that just repeat a random post from the thread to add their own advertisement once it gets accepted.
Dropped a lot compared to what? 7.1.1?Internet speed through wi-fi dropped a lot
Yes, on 7.1.1 the speed was about 500 megabits per second, and on 7.2rc3 about 300.Dropped a lot compared to what? 7.1.1?Internet speed through wi-fi dropped a lot
I'd imagine it's the effect of the removal of the "route cache". In V6, a speedtest would hit the cache, while in V7 there is no cache to hit. See viewtopic.php?p=882867#p882867Yes, on 7.1.1 the speed was about 500 megabits per second, and on 7.2rc3 about 300.
Dropped a lot compared to what? 7.1.1?
It is the internet speed. The local network speed has not changed.
Tested With iOS.@sirbryan
The point is an existing configuration should work from one version to the next. It doesn't. (And this is supposed to be a release candidate.)Tested With iOS.@sirbryan
2022-02-09_06-23-47.png
Does that look like the same issue as is being discussed in this thread?Debugging the L2TP packets shows an initial CHAP challenge that is accepted and tunnel is created. Then a second tunnel starts with another challenge sent by the router, but the device doesn't reply. Eventually both tunnels are torn down and the router complains that the device didn't respond.
Hi,BGP vrf problem is acknowledged and will be fixed in one of the next versions.
Be patient, we already are waiting for "the new and improved BGP" for almost 10 years!could u make it fixed fast.
It's been 2 weeks i 've waiting.
No. That's voice over WiFi. This is an L2TP IPSEC tunnel from a phone or laptop coming back into the router. It works on 7.1.1/7.1.2, but not on 7.2rc3. That other issue has been going on since before rc3 was released.Does that look like the same issue as is being discussed in this thread?Debugging the L2TP packets shows an initial CHAP challenge that is accepted and tunnel is created. Then a second tunnel starts with another challenge sent by the router, but the device doesn't reply. Eventually both tunnels are torn down and the router complains that the device didn't respond.
viewtopic.php?t=181449
My CCR are getting rusty waiting.Be patient, we already are waiting for "the new and improved BGP" for almost 10 years!could u make it fixed fast.
It's been 2 weeks i 've waiting.
I believe a lot a people buy CCR for advance routing purposes and not for lte or wifi
My home router, a hAP ac2, is set to channel=stable, currently running 6.49.1. If I run check-for-updates, it offers me upgrade to 6.49.2.One hint is that v7, with the BGP implementation incomplete, buggy and unfinished, was promoted to stable version and offered as an upgrade.
Which channel do you have it set to?Disagree.
I have a hap ac2 running on 6.48.6 and it insist on upgrading to 7.1.1.
When you go to the MikroTik software download page https://mikrotik.com/download you'll see RouterOS v7 prominently displayed with Long-term, Stable and Testing channels, and v6 is only available as an option way down the page.My home router, a hAP ac2, is set to channel=stable, currently running 6.49.1. If I run check-for-updates, it offers me upgrade to 6.49.2.One hint is that v7, with the BGP implementation incomplete, buggy and unfinished, was promoted to stable version and offered as an upgrade.
It's not offering me v7. So sincerely: noone is pushed to install v7 (except on v7-only devices[*]) unless one reaches out to get it ...
When you go to the MikroTik software download page ...
Very odd.Which channel do you have it set to?Disagree.
I have a hap ac2 running on 6.48.6 and it insist on upgrading to 7.1.1.
It shows testing, which explains partially why it proposes that version.
But then it should have proposed 7.1.2 or 7.2rc3, no ?
Yup, these seem to work very well in this release. The GPS package even works with the Telit LM960s in this build (wrote up the steps here: viewtopic.php?p=912653#p912653 if anyone needs)*) lte - fixed support for Telit 960;
I've read it all, but it's unclear to me what you mean mikrotik should do on their wifi gear. Set the wireless interface to use fq_codel instead of sfq ? And if so, with what config exactly ? I haven't seen any results to back anything up yet in that article you've linked to.It's also not clear to me if mikrotik picked up the wifi work on any of their wifi gear: https://lwn.net/Articles/705884/
Yes, I can see there is a host-uniq parameter for PPPoE clients in this version. It is only available through the CLI, not through winbox/webfig. I am able to set it in my own device, but because my ISP does not require setting this, I cannot confirm that it works as expected.My ISP requires host-uniq as well as user/pass for PPPoE. Can someone confirm I can set this?
/interface/pppoe-client/set [ find ] host-uniq="<required string>"
this made my day 😂👌you stop using pptp
If you're still using PPTP in 2022 then you shouldn't consider yourself a professional!How is PPTP any less secure than running IPIP, EoIP or any other VPN protocol with no cipher enabled that will show no such warning? And what's next? All open WIFI interfaces will show red warnings too? ROS is supposed to be for professionals that know how to use it and what limitations each of these protocols may have. There are use cases for all of them, be it various vendor hardware or software compatibility, specific MTU requirements etc. There is no need for Mikrotik to be babysitting it's users.
This is true for several devices, but not all. In fact, just a small set of them support this. The link bellow is a good starting point:*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled;
*) l3hw - added HW offloaded FastTrack support for inter-VLAN routing;
*) l3hw - fixed HW offloaded NAT;
Wait, if this is true and it's applies to all devices this is going to be a major improvement. Im going to have to look into this more.
"Because of an error in data encryption, this session will end. Please try connecting to the remote computer again."
I'm afraid that's only related to the thing of perceived security. There is no username and password specified for IPIP or EoIP, so they are not considered a "VPN" in the sense an "average Joe" understands it (if you set the IPsec secret for these, they become much more secure than PPTP); PPTP does contain the magic username/password combination, so it must be secure, mustn't it?How is PPTP any less secure than running IPIP, EoIP or any other VPN protocol with no cipher enabled that will show no such warning?
...
There is no need for Mikrotik to be babysitting it's users.
Im using this one:Please show your script, there may be an error there.
Support confirmed this is an issue (posting for any others tracking this particular problem).Tried again with 7.2rc3 after tweaking the firewall and comparing the L2TP setup with stock ROS7 L2TP/IPSEC configurations, but no luck.
Installed 7.1.2/7.1.3 and devices connect right up.
Debugging the L2TP packets shows an initial CHAP challenge that is accepted and tunnel is created. Then a second tunnel starts with another challenge sent by the router, but the device doesn't reply. Eventually both tunnels are torn down and the router complains that the device didn't respond.
This is true for several devices, but not all. In fact, just a small set of them support this. The link bellow is a good starting point:*) bridge - added fast-path and inter-VLAN routing FastTrack support when vlan-filtering is enabled;
*) l3hw - added HW offloaded FastTrack support for inter-VLAN routing;
*) l3hw - fixed HW offloaded NAT;
Wait, if this is true and it's applies to all devices this is going to be a major improvement. Im going to have to look into this more.
https://help.mikrotik.com/docs/display/ ... Offloading
😳🤪For example, in the old days of Windows NT development there were about 5000 developers and about 5000 testers, that's why the resulting programs run so well, even to this day.
Comparing to 9x branch? It ran VERY well.😳🤪For example, in the old days of Windows NT development there were about 5000 developers and about 5000 testers, that's why the resulting programs run so well, even to this day.
Seriously? We must have a different understanding of "run so well".
Well, with BGP I understand how it can happen!I'm a regular home SOHO user, but looking at the quality of builds they are pushing out, even under 7.1 stable (yeah, call it "upgrade" branch...), with bootloops, BGP, VLAN, etc., issues, I am just mind boggled how it can happen.
I started using NT pretty early when it came out.Comparing to 9x branch? It ran VERY well.
😳🤪
Seriously? We must have a different understanding of "run so well".
Hehe: I would say if you run a network big enough so that you need BGP, you should have enough insight and in the first place not even try to put v7 in production. Keep your shiny new ROSv7 devices in your lab...So now that is where we are. A new kernel and exciting new features that have been promised for a long time, but unfortunately a BGP implementation that is unfinished and buggy.
It is a bit sad that there are no big leaps to be observed in BGP between the releases. I would have hoped that all hands were now put on its completion, rather than on fasttrack or l3hw, as it is a critical part of the system (for those that use BGP) that renders the whole version unusable for those that use BGP.
As soon as version 7 is feature-complete compared to v6, it would be possible to drop v6 development and focus on the new v7 features only.
Version 7 has so much interesting new features that we have all been waiting for that it is really a nuisance that there is this big showstopper for me and others.
I would even welcome an optional "BGP legacy" package that brings the v6 BGP into v7 to use it until the native one is finished.
I want to put v7 in use, but I cannot due to the sad state of BGP.Hehe: I would say if you run a network big enough so that you need BGP, you should have enough insight and in the first place not even try to put v7 in production. Keep your shiny new ROSv7 devices in your lab...