RouterOS version 7.2.1 has been released in "v7 stable" channel!

Before an upgrade:
1) Remember to make backup/export files before an upgrade and save them on another storage device;
2) Make sure the device will not lose power during upgrade process;
3) Device has enough free storage space for all RouterOS packages to be downloaded.

What's new in 7.2.1 (2022-Apr-06 11:52):

*) filesystem - improved long-term filesystem stability and data integrity;

To upgrade, click "Check for updates" at /system package in your RouterOS configuration interface, or head to our download page: http://www.mikrotik.com/download

If you experience version related issues, then please send supout file from your router to support@mikrotik.com. File must be generated while router is not working as suspected or after some problem has appeared on device

Please keep this forum topic strictly related to this particular RouterOS release.

To me, this is a strange start.

Why was that not in the 7.2 release.

Because this build is from last week, and not from last month.

Because an issue was discovered and fixed after release. Since it's not a big change, but an important one, we did not wait until 7.3

Paritioning on RB5009 still NOT WORKING.

Then 7.2 had a short life. Moving from release to testing in a short time.

Paritioning on RB5009 still NOT WORKING.

Was it mentioned in the single fix mentioned in the changelog?

Is this meant to solve the problem of partial config disappearance?

Paritioning on RB5009 still NOT WORKING.

Was it mentioned in the single fix mentioned in the changelog?

No, but is what (still) not work on 7.2.1, and is also about filesystem...
It has not been written what exactly was done, so one can only guess.

Then 7.2 had a short life. Moving from release to testing in a short time.

7.2 is just as stable as it was before
7.2.1 is in testing to test this one change only. You can keep using 7.2 it did not disappear.

RB5009 partitioning issue is a known problem and will be fixed in future releases. There is no need to mention it again and again in forum topics. We are aware of the issue.

[...] is also about filesystem...
It has not been written what exactly was done, so one can only guess.

@normis, @strods
Thanks.

I rather saw something as "7.2, as stable as 7.21".

Edit: then this testing version could be a mitagation of v7.2 bricking some devices. I am still on a RC brcause of that happening to others.

This is wonderful news. Thanks Normis, strods & whole Mikrotik team working on this issue, some of whom probably had a busy weekend hunting down bugs (more like finding a needle in a scrap yard with this one I'm sure) for the rest of us.

I just want to know if it would be possible to maybe share some more information about what was fixed/changed with this update? We obviously don't have the same insider information as RouterOS developers so we don't know how impactful this change could really be. Is it worth upgrading potentially dozens/hundreds of routers already on 7.2 to this new release or is it only recommended for new upgrades? Is this patch intended to fix the corruption on upgrade/reboot issue, or is it a broader & more general change which may or may not have this effect and we will just have to test? Any hints or pointers to address these types of questions will be greatly appreciated.. The only hint we have right now is the use of the word "improved" rather than "fixed" in the changelog and even that is quite vague as it doesn't really convey the extensiveness of the changes.
Maybe you can just help us weigh out our options and point us in a preferred direction depending what scenarios we are already in, given the potential significance this single issue (or its fix) can have.
Thank you.

Update: successfully upgraded my home RB5009 plus a few other local devices without any data loss thus far (Yipee!). Results are looking very promising thus far. Still too scared to try any remote devices for now, I'll think about it while waiting for some more more forum comments/feedback 😂

winbox is refusing to file transfer to my hap lite, will try again later.

edit: fixed, hap lite os and firmware upgraded.

You do not provide any relevant detail, why you post that?
Why you use 7.x on very-limited-all smips?

Well this is one of the scariest changelog entries. Is every prior RouterOS version affected by this "filesystem stability and data integrity" issue?

Nothing new, the 7.x topics have some "lost all configuration" issue
But for the few device I have with 7.x is never happened,
but for some of my device (neat 1 on 100) happen from upgrading 6.44.6 to 6.46.8,
and the problem is on the missing "ssh host key"...

Just I remember now something like netinstall for do faster installation do not generate ssh host key on first RouterOS boot on old version...

"improved something" means "fixed something" in MT-changelog-words. Nobody knows what was fixed actually. Why should anyone install this testing release? Without knowing what was a) improved or changed?? How can anyone judge a difference between 7.2 and 7.2.1? MT staff keeps commenting here but give no additional info to this very vague changelog entry.

You do not provide any relevant detail, why you post that?
Why you use 7.x on very-limited-all smips?

because it's the only Mikrotik with multiple ethernet ports that can be powered via micro usb
because it's used by thousands because it's the cheapest entry point for learning Mikrotik

I posted it because prior to 7.2, I could just winbox the smips .npk to files over the mac address, and reboot and it would upgrade.
I've enabled the dhcp client now, and upgraded via system packages.
There are Mikrotik devices that are only accessible on the Layer 2 network, and I'm just advising I had to enable Layer 3 networking to install this.

Why should anyone install this testing release?

If you're asking this, then perhaps you should be asking yourself: "why am I using Mikrotik products?"

You always have to ask yourself if you trust producer of certain item that that item is fit in all aspects to be used by you. And if you answer with kind of a "yes", then you should probably trust producer about future products as well. If you loose the trust, then probably the reason for loosing trust existed long before you lost it.

For example: if you think that MT introduced something evil in 7.2.1, then I guess you have no reason to believe 7.2 is any better ... or 6.48.5 for that matter.

Or, you can do what many people do: wait for some brave souls to install and test it and report the experience. And then decide if you want to install it (if you believe that reporters are true with their reports).

7.2.1 is in testing to test this one change only. You can keep using 7.2 it did not disappear.

I do not mind testing it, just the description is too vague to understand what was broken exactly. Could you describe a scenario that would break the router with 7.2?

@mkx
That is unfair and wrong. You take my argument and inverse it.

Anyone can update on "stable" channel without actually knowing the exact information to all the fixes. all you need to know: "i do my stable updates when they are available and known to be "safe" from reading this forum. Update to be up2date"

But testing release? "Hey folks, we have a testing release. We improved something. Some stability. Can you please give feedback about? on what? feedback on "improved stability of filesystem"? Is there not any further information than "improved xyz"? Is it that crazy to ask this question in a TESTING release thread?

If you experience version related issues, then please send supout file from your router to support@mikrotik.com

How can ANYONE give serious feedback to this question?

When there is ONLY A SINGLE CHANGE, then I install this testing-release. I can have a SPECIFIC look on the change and can tell: "yes, it indeed improved the stability for this before instable behaviour" or I can report "no, this change did not improve stability".

But it here is like: "I installed 7.1.2 and I do not see any difference/improvement/change to 7.2. It W0rKs!!11!11!!"

ontopic: Probably related to the config-corrupted-after-update that's been around the forums lately?
offtopic: Unfortunately MikroTik doesn't reveal the full changelog vs 7.2, again.
Comparing files you'll see differences in files like capsman.ko, wireless, ipsec, profiler, libcrypto.so, a lot of webfig(?) .info files that were still showing 7.2rc5, a lot of kernel modules but I'm sure there's just a version change there, maybe except sch_agr.ko and sch_sfq.ko which seem to have a little more diffs (probably still related to versioning).
So... which is it, MikroTik? only version change in those files or maybe something else was fixed/changed too? :)
Thank you!

For starters - this fix has nothing to do with the problem that some rare routers lose configuration after an upgrade. Still with all of your reports to support. we have not managed to reproduce the problem. As we thought before, most likely this problem is already fixed and affected are only routers which had some bad version installed on the in the past. Every by importing backup made before failed upgrade on another upgrade configuration is not lost.

Regarding this particular release - it fixes a very rare situation when a router could brick itself during the upgrade process by removing/corruption filesystem so the device could not read system files anymore. The router had to be get netinstalled. Now, this particular case is fixed. We decided to better release this fix as soon as possible rather than keeping it up to our sleeve. Since the fix is safe and could help some people to avoid unnecessary travels to their server rooms we released it in the testing channel. If no unexpected issues will be found, we will release it in the stable channel as an improvement, not an update for 7.2.1. This release has nothing to do with the 7.2 release. Simply we found this issue and fixed it just now.

Thanks @strods for detailed reply.

RB2011il upgraded, no issues found.

Thanks Strods, this gives clarity about the update and gives me hope to leave RC for, maybe testing. To me, it is still a fix to 7.2 and previous versions.

Regarding this particular release - it fixes a very rare situation when a router could brick itself during the upgrade process by removing/corruption filesystem so the device could not read system files anymore. The router had to be get netinstalled.

Does this affect all v6.x and v7.x releases? If not, when was the bug introduced?

Read twice the post #25
viewtopic.php?t=184992#p925750
and you find your answer

Doesn't help. I'm guessing this affects only v7 but it would be nice to know for sure, and whether this will be backported. I've lost at least one 6.x device after a failed upgrade.

Is this filesystem configuration change as opposed to completely changing filesystem from ext4 to btrfs for example? What filesystem is RouterOS using?

I for one am very happy with all the changes and continued improvements of Router OS 7. Just did the upgrade on one of my CCR2004 and from upgrade back to forwarding and management was less than 22 seconds. Try doing that with a Cisco ;)

I second that, I am a newcomer to Mikrotik, just a few months. Coming from 2 Ubiquiti edgerouters my Rb5009 and a HAPac3 a working perfectly and I am glad I am receiving updates instead of waiting months and months for a small update from Ubiquiti.
Both devices are upgrading in mere seconds (<30S)

I for one am very happy with all the changes and continued improvements of Router OS 7. Just did the upgrade on one of my CCR2004 and from upgrade back to forwarding and management was less than 22 seconds. Try doing that with a Cisco ;)

Will you two stop being reasonable? You're supposed to moan about every little thing here. Didn't you realize? 😛

Is like constructive posts are sold out...

Or, you can do what many people do: wait for some brave souls to install and test it and report the experience.

I like to wait about six months to a year before touching MikroTik firmware ... with a feather.

Or, you can do what many people do: wait for some brave souls to install and test it and report the experience.

I like to wait about six months to a year before touching MikroTik firmware ... with a feather.

Oh, you're such a faint of heart.

Better is wait what some WISP, with more than 4000 devices, suggest... ;)

I want to change my RB2011 router for CCR2004-16G-2S+ could someone tell me if this works with RouterOs Version 6?

As it says on that product's specs page, it's v7 only.

test chr updated, no errors so far, but still no ipv6 traffic through eoip (ipv4) tunnel

Since 7.2 my HexS is booting every now a then ... Log shows:

> system,error,critical router was rebooted without proper shutdown, probably kernel failure

Will 7.2.1 fix this?

Try netinstall or open separate topic for that, I have an hex s at my home and from 7.2 update I never get that problem

Since 7.2 my HexS is booting every now a then ... Log shows:

> system,error,critical router was rebooted without proper shutdown, probably kernel failure

Will 7.2.1 fix this?

I saw that on my hEX S as well with version 7.2. Before that I also have an RB5009 and with version 7.1.5 also saw the same error.

Where is the fix for L2TP/IPSec (Error on Phase 2)??? and OVPN??

Right now 7.x is kind of useless without these features stable and rock solid like ver 6.x has.

I'm sure you got a reply by now to your properly submitted bug report regarding OpenVPN and L2TP/IPSec.

tuxedo0801 please create a supout.rif file and contact support@mikrotik.com ,check if there are any autosupout.rif files, if there are please share them as well. Without examining them we can't speculate on what caused the specific issue.

I am not sure if that's something new or if it existed before.
When the router is restarted and OSPF starts sync with other routers, I always get those warnings:



After a while, those warnings disappear.
All routers run v7.2 and have almost identical OSPF config.

Regarding this particular release - it fixes a very rare situation when a router could brick itself during the upgrade process by removing/corruption filesystem so the device could not read system files anymore.

So, basically the way I am reading this, if we are already on 7.2, there is no need to update to 7.2.1 (because we are on 7.2 and therefore have already gone through the "during the upgrade" process and didn't hit the bug).

In other words, don't upgrade to any version less than 7.2.1, make sure to upgrade to at least 7.2.1 when you upgrade. However, don't upgrade to 7.2.1 just for the sole purpose of of getting the fix because future versions will have the same fix and the bug only happens during upgrades anyway.

Is all of the above correct?

If that is correct then it should be a release and not a testing version.

If that is correct then it should be a release and not a testing version.

Testing version exists for testing before rolling out to stable...

Mikrotik has three branches, long term, stable and testing.

Everything gets tested in testing first before reaching stable, certain stable versions get chosen as long term every three months.

This is how mikrotik test releases, and this is the release process.

Regarding this particular release - it fixes a very rare situation when a router could brick itself during the upgrade process by removing/corruption filesystem so the device could not read system files anymore.

So, basically the way I am reading this, if we are already on 7.2, there is no need to update to 7.2.1 (because we are on 7.2 and therefore have already gone through the "during the upgrade" process and didn't hit the bug).

In other words, don't upgrade to any version less than 7.2.1, make sure to upgrade to at least 7.2.1 when you upgrade. However, don't upgrade to 7.2.1 just for the sole purpose of of getting the fix because future versions will have the same fix and the bug only happens during upgrades anyway.

Is all of the above correct?

Stay on stable release unless you are interested in bug reporting and fixing.

Updated a CCR2004-16G-2S+ to 7.2.1 Testing
Routing aroung 2.2 Gbps
~3900 ipv4 routes OSPF
~1700 ipv6 routes OSPFv3

No queues.
15% of traffic masquarade.
Fasttrack enabled.
Around 58% cpu usage, no memory leaks with current config
OSPFv3 interoperability with v6 now looks to be working well. (Had issues in 7.1.3)!!
So far stable.
CPU usage seems a bit high?

I'm sure you got a reply by now to your properly submitted bug report regarding OpenVPN and L2TP/IPSec.

Reported the issues since v7.1.5 and 7.2 and I am not the only one sharing the same issues.

Would this fix be related to running scripts such as the ATT gateway bypass? My ATT gateway bypass was broken in 7.2 but working in 7.1.5

EDIT: Just tested on an RB4011, ATT gateway script still not working. Reverted back to 7.1.5.

upgrade 7.2 to 7.2.1
on hAP941 fix memory and cpu usage

on hEXr3?
i lost, all bridges, all filter and nat rules, all wireguard interface (peers is ok) :-(

Probably if you just restart the device, you would have lost it anyway...
Before updating, always reboot first, to remove any doubts...

upgrade 7.2 to 7.2.1
on hAP941 fix memory and cpu usage

on hEXr3?
i lost, all bridges, all filter and nat rules, all wireguard interface (peers is ok) :-(

not fix, one hAP have still problem

19:43:31 system,error,critical router was rebooted without proper shutdown
19:43:31 system,error,critical kernel failure in previous boot
19:43:31 system,error,critical out of memory condition was detected

Probably if you just restart the device, you would have lost it anyway...
Before updating, always reboot first, to remove any doubts...

last restart was 10 minutes before upgrade

[...] to remove any doubts...

Doubts removed...

Would this fix be related to running scripts such as the ATT gateway bypass? My ATT gateway bypass was broken in 7.2 but working in 7.1.5

EDIT: Just tested on an RB4011, ATT gateway script still not working. Reverted back to 7.1.5.

This fix is ONLY to stop routers getting bricked during an upgrade making it require a netinstall. If you can access your router at all, then this fix is not intended for your specific issue.

upgrade 7.2 to 7.2.1
on hAP941 fix memory and cpu usage

on hEXr3?
i lost, all bridges, all filter and nat rules, all wireguard interface (peers is ok) :-(

not fix, one hAP have still problem

19:43:31 system,error,critical router was rebooted without proper shutdown
19:43:31 system,error,critical kernel failure in previous boot
19:43:31 system,error,critical out of memory condition was detected

This fix is ONLY to stop routers getting bricked during an upgrade making it require a netinstall. If you can access your router at all, then this fix is not intended for your specific issue.

Report your missing configuration here viewtopic.php?t=184711 they are requesting these examples

upgrade 7.2 to 7.2.1
on hAP941 fix memory and cpu usage

Yes 👍 same here, for SMIPS device like lite and mini it's fixed lots of issues.

Would this fix be related to running scripts such as the ATT gateway bypass? My ATT gateway bypass was broken in 7.2 but working in 7.1.5 EDIT: Just tested on an RB4011, ATT gateway script still not working. Reverted back to 7.1.5.

What error are you getting?

Would this fix be related to running scripts such as the ATT gateway bypass? My ATT gateway bypass was broken in 7.2 but working in 7.1.5 EDIT: Just tested on an RB4011, ATT gateway script still not working. Reverted back to 7.1.5.

What error are you getting?

Im not getting anything. After booting up and script running gateway bypass (bridge method) and waiting the 3 mins the DHCP client never pulls an IP address or GW. It just sits there and does nothing. When I revert back to 7.1.5, all works as normal.

tuxedo0801 please create a supout.rif file and contact support@mikrotik.com ,check if there are any autosupout.rif files, if there are please share them as well. Without examining them we can't speculate on what caused the specific issue.

I created a ticket: SUP-79498

Ooo, more updates. Neat. :)

Not updated anything yet since the 7.1.1, have been hovering over an imaginary "update" button for my zero conf switch for a few days.
Dare I push it? CRS326-24G-2S+RM.
Not touching my router yet.

7.2 is a much better and more stable version than 7.1.1, no matter how you look at it.

Regarding this particular release - it fixes a very rare situation when a router could brick itself during the upgrade process by removing/corruption filesystem so the device could not read system files anymore. The router had to be get netinstalled. Now, this particular case is fixed. We decided to better release this fix as soon as possible rather than keeping it up to our sleeve. Since the fix is safe and could help some people to avoid unnecessary travels to their server rooms we released it in the testing channel. If no unexpected issues will be found, we will release it in the stable channel as an improvement, not an update for 7.2.1. This release has nothing to do with the 7.2 release. Simply we found this issue and fixed it just now.

That's quite an important fix indeed! Thank you for clarification. This makes this 7.2.1 a really important one. And people having encountered this described filsystem-corruption situation in the past, can try to upgrade from their backups to this 7.2.1 release. Given, it is reproducible from a backup.

See how important clear and meaningful communication of changes is?

changelog: "we improved filesystem stability"
strods: "we fixed a rare situation where upgrade could turn your device into a brick. netinstall only to recover it."

I don´t know if this behavoiur has something to do with 7.2.1, but since the update my Router RB5009 looses after different times half of the packets in direction internet.
Exactly every second PING is lost. After a reboot the connection is ok, but only for a while.
I´ve downgraded back to 7.2 for comparison......I´ll report about the behaviour with 7.2..........
I´ve nothing changed in the config between 7.2 and 7.2.1...no idea what´s going on :-/

@Ullinator I've had that problem when I was bridging software vlans on my Rb5009. Do you have such configuration by any chance? Also make sure you aren't showing any "bridge port" columns in winbox/cli/webfig (eg arp list, DHCP leases window) which is known to cause similar instability on the Rb5009 (SUP-71476)

@Ullinator I've had that problem when I was bridging software vlans on my Rb5009. Do you have any vlans? Also make sure you aren't showing any "bridge port" columns in winbox/cli/webfig (eg arp list, DHCP leases window) which can cause such instability on the Rb5009. Wondering when Mikrotik are going to finally fix this issue as it seems to affect all Rb5009s and I've seen people RMA their units thinking it's a hardware problem when it isn't (unless all our devices do have such a hardware problem after all)

@fragtion: indeed I have one vLAN in the bridge, but exactly as it is adviced by MT: Could this really be the problem?
Thank´s for advice :-)

For starters - this fix has nothing to do with the problem that some rare routers lose configuration after an upgrade. Still with all of your reports to support. we have not managed to reproduce the problem. As we thought before, most likely this problem is already fixed and affected are only routers which had some bad version installed on the in the past. Every by importing backup made before failed upgrade on another upgrade configuration is not lost.

Do you really have no idea what is the exact cause of this issue that has affected so many of us? I mean, even when you are not able to reproduce the problem, with the inside knowledge of RouterOS you should be able to state some cases that might ultimately lead to loss of configuration, usually an entire section. You should know how configuration is stored and how it can be lost.
It is often reported as "after an upgrade" but I think in fact it is "after a reboot" (which of course is always done during an upgrade, but can also be done independently).
Do you have any indication that the "/export the config, netinstall the device, /import the config" route is fixing the problem? At least, it is what I noticed on my device and I think others noticed it as well.
Is it triggered by upgrading from v6, something that you may not be doing so often anymore in the development department?
When it can be fixed by export/import and does not need a format, would you consider a "repair config database" command that can do this automatically (export, remove database, import) without netinstall and without session connected to MAC address?

It may seem like a minor inconvenience for you, but for users in the field it is not only a major annoyance, but it also affects our trust in the stability of the platform and our confidence in updating from remote.
Please treat it with the urgency and attention that it requires, and implement workarounds when it is not clear how it can really be fixed.

"improved something" means "fixed something" in MT-changelog-words.

I always smile when I read "improve stability when doing ..." in the changelog, which just means "fixed a problem that caused a crash" :-)

I don´t know if this behavoiur has something to do with 7.2.1, but since the update my Router RB5009 looses after different times half of the packets in direction internet.
Exactly every second PING is lost. After a reboot the connection is ok, but only for a while.

That sort of hints towards a second default route that somehow gets installed and that leads to an ECMP situation with one working and one not working route.
It heavily depends on your config how that situation could come about. E.g. running scripts, PPPoE and DHCP clients, autorouting protocols, etc.

@fragtion: indeed I have one vLAN in the bridge, but exactly as it is adviced by MT:
Could this really be the problem?
Thank´s for advice :-)

Test and let us know if it helps !!🙂 Then if you confirm it's a problem despite being implemented in the way Mikrotik recommends, then I suggest opening a support ticket with Mikrotik about this specific issue with the bridged software vlans as I have not done so yet (I only have ticket open for the DHCP leases lockup issue)

I don´t know if this behavoiur has something to do with 7.2.1, but since the update my Router RB5009 looses after different times half of the packets in direction internet.
Exactly every second PING is lost. After a reboot the connection is ok, but only for a while.

That sort of hints towards a second default route that somehow gets installed and that leads to an ECMP situation with one working and one not working route.
It heavily depends on your config how that situation could come about. E.g. running scripts, PPPoE and DHCP clients, autorouting protocols, etc.

@pe1chl: Thank´s for the tipp, but there are no other def. routes, not static, not dynamic and not via script. And the exact same configuration runs with all other ROS-releases without probs.:-/
The RB5009 is my router to the internet and the IP comes from my ISP via DHCPv4. It´s a cable connection behind an cable modem without PPPoE or something else. Only native DHCP via Ethernet.
My default route targets to the ISP (as it should). Other routes are only dynamic onces, like WireGuard, but no additional def. gateways. Of course another for IPv6 ;-)
There is only one static route is for another department (eg. 192.168.50.0/24), that´s all.
Until now it runs without probs with ROS 7.2 after my downgrade....hmmm.

Start a new thread or post your config if you want assistance........ (preferably both)........... as for the others, eat some lucky charms, tricks or (guessing) is for kids!!

Start a new thread or post your config if you want assistance........ (preferably both)........... as for the others, eat some lucky charms, tricks or (guessing) is for kids!!

@anav: the only reason why I post this in this treat was, that the behaviour first startet shortly after the update to 7.2.1.
And yes, I will open another thread :-)

Updated a CCR2004-16G-2S+ to 7.2.1 Testing
Routing aroung 2.2 Gbps
~3900 ipv4 routes OSPF
~1700 ipv6 routes OSPFv3

No queues.
15% of traffic masquarade.
Fasttrack enabled.
Around 58% cpu usage, no memory leaks with current config
OSPFv3 interoperability with v6 now looks to be working well. (Had issues in 7.1.3)!!
So far stable.
CPU usage seems a bit high?

Looks to be par for the course. Disappointing, but not surprising since they're similar in horsepower to an RB4011 or RB5009.

Odd that you had issues with 7.1.3. That's the only one where OSPF behaves on my network. Everything newer causes some kind of loop on my network, likely related to filtering changes in newer releases.

I look at these devices as less about achieving full-rate 10Gbps and more about meeting needs of 1-2.5Gbps across multiple links. That's also what others have mentioned on the forums and in online reviews.

I did a quick series of tests between a CCR2116 running 7.2 and a CCR2004-16G-2S+ running 7.1.3, bridged over the office CCR2004. The bridging 2004's CPU hovers around 34% for all tests. The CPU mentioned is the 2004 (the 2116 twiddles its thumbs at 5%).

TCP 2004 -> 2116 = 5.7Gbps @ 84% CPU
TCP 2116 -> 2004 = 5.6Gbps @ 90% CPU
TCP 2004 <-> 2116 = 1.2Gbps each way, 90% CPU
UDP 2116 -> 2004 = 9.4Gbps @ 80-90% CPU
UDP 2004 -> 2116 = 6Gbps @ 93% CPU

Since 7.2 my HexS is booting every now a then ... Log shows:

> system,error,critical router was rebooted without proper shutdown, probably kernel failure

Will 7.2.1 fix this?

are you using jumbo frame?
in my hap ac2 fixed disabled jumbo frame >1500 mtu, using default mtu 1500

Updated a CCR2004-16G-2S+ to 7.2.1 Testing
Routing aroung 2.2 Gbps
~3900 ipv4 routes OSPF
~1700 ipv6 routes OSPFv3

No queues.
15% of traffic masquarade.
Fasttrack enabled.
Around 58% cpu usage, no memory leaks with current config
OSPFv3 interoperability with v6 now looks to be working well. (Had issues in 7.1.3)!!
So far stable.
CPU usage seems a bit high?

Looks to be par for the course. Disappointing, but not surprising since they're similar in horsepower to an RB4011 or RB5009.
[.....]

I spent some time working on this last night. After upgrading a CCR2004 and two RB5009's to 7.2, I'm able to get 6Gbps between an 2116 and 1036 routing through the CCR2004 and RB5009. The 2004 is topping out around 60% because the RB5009 is maxing out at 90% or so at the 6Gbps.

In other words, significant improvement for the CCR2004 in 7.2(.1) compared to 7.1.x. My OSPF issues (seen in 7.1.4/5) are gone in 7.2 as well.

Found a bug on routers with several public IPs where ipsec tunnel negociation is handled with the proper defined local-address address, but then data traffic is sent using the lowest public IP, making, as a result, receiver to reject it.
SUP-79720 created accordingly.

Edit : was facing this : https://help.mikrotik.com/docs/display/ ... Paddresses
Thank U MikroTik support !

RouterOS can't generate multiple Let's Encrypt certificates for different domains (using the /certificate enable-ssl-certificate dns-name="different.domain.com" command), it simply keeps re-generating the first one already created.
SUP-79724 created accordingly.

Hi, I have l2tp with ipsec between two mikrotiks and with ros 7.2.1 it is hanging up every couple minutes. I dont know if that sessions are closed by server or client but it is happening.

EDIT :
Never mind, probably my fault. Disabled some rules on firewall.

EDIT 2 :
problem still persist, I will do downgrade to ros 7.2 and I will let known....

EDIT 3 :
downgrade of l2tp server side (rb5009) to ros 7.2 has bring solution of this problem

Upgraded 1009-7G-1C from v6 stable to v7, half hour passed not coming back online, I hope I'm not victim of upgrade like here..
viewtopic.php?t=181525

Need to go to location tomorow anyway but comon mikrotik how you not solve this issues by now?

L2TP/IPSEC vpn on ccr1009 not possible from windows and linux, now it does not connect and puts some processors at 100%.
What is going on !!!!

Hi MT Devs
maybe i am wrong, but ...

[IP-FORWARD-MIB::ipCidrRouteIfIndex] => 0
why i have a feeling that this should to point to intermediate gateway ???

BTW, still need to use -Cc on both IPv6 and IPv4 routing table :(

RouterOS V7 Don't Exit The user from PPP Active after Disconnected the users

It's Exit from the Mikrotik Radius but it's stil showing it's connected in PPP active

RouterBoard : 1009 , 1036

Version(1009) : 7.2.1
Version (1036) : 7.1.5

the two RouterBoard has same proplem
this video for the proplem
https://youtu.be/KRQPvA4Y2Dg

RouterOS version 7.2.1 has been released in "v7 stable" channel!

Before an upgrade:
1) Remember to make backup/export files before an upgrade and save them on another storage device;
2) Make sure the device will not lose power during upgrade process;
3) Device has enough free storage space for all RouterOS packages to be downloaded.

What's new in 7.2.1 (2022-Apr-06 11:52):

*) filesystem - improved long-term filesystem stability and data integrity;

To upgrade, click "Check for updates" at /system package in your RouterOS configuration interface, or head to our download page: http://www.mikrotik.com/download

If you experience version related issues, then please send supout file from your router to support@mikrotik.com. File must be generated while router is not working as suspected or after some problem has appeared on device

Please keep this forum topic strictly related to this particular RouterOS release.

RouterOS V7 Don't Exit The user from PPP Active after Disconnected the users

It's Exit from the Mikrotik Radius but it's stil showing it's connected in PPP active

RouterBoard : 1009 , 1036

Version(1009) : 7.2.1
Version (1036) : 7.1.5

the two RouterBoard has same proplem
this video for the proplem
https://youtu.be/KRQPvA4Y2Dg

do i need to worry about? i didn't change anything related to the frequency.
the previous 7.2 was fine

It will disappear when you click Settings and in that screen select AUTO for the frequency.

thanks

@MT: are there any changes between 7.2.1 (Testing) and 7.2.1 (Stable)?
Do I have to reflash the 7.2.1 (stable) when I had already flashed 7.2.1 (Testing)??
Thanks!!

Is the same

L2TP/IPSEC vpn on ccr1009 not possible from windows and linux, now it does not connect and puts some processors at 100%.
What is going on !!!!

I have the same issue too on ccr1009 and 7.2.0 version.

I am not alone :( on this.....
I was forced to reconfigure from scratch a RB4011 ( 16 hours or so of work) at the end.... another surprise the scripts wont work for no-ip!!!??? linux machines save me..
It seems that each router model has its diferent problems!? So bizarre... or lack of quallity control or testing, usb does not work too......Linux save me NUT ups control....
Funny fact the RB750Gr3 (the small runs fine).!!!
Easter hollidays for all i will hope this will improve asap....
I know mikrotik should be in big trouble but quality control should go up everytime.
Mikrotik wake up..

L2TP/IPSEC vpn on ccr1009 not possible from windows and linux, now it does not connect and puts some processors at 100%.
What is going on !!!!

I have the same issue too on ccr1009 and 7.2.0 version.

same issue too...

OSPF on CCR2116-12G-4S+ with FW 7.2.1 not work!

I have the same issue too on ccr1009 and 7.2.0 version.

same issue too...

L2TP/IPSec semi broken in 7.1.5 and now completely broken in 7.2 and 7.2.1

Is the extra package zerotier on x86 7.2.1 not available yet?

Terminal worked just after the upgrade but after 1st reboot stopped. Cant copy paste

BGP is missing in SMIPS v.7.2.1 (Stable) package.
Is it a “new normal” or mistake?

P.S. I upgraded my RB941 to v.7.2.1 directly from v.6.49.2, so I can’t comment about SMIPS BGP status in previous v.7.x releases.

Hello.
Hardware:
hEX S + cAP ac

Software:
3 main functions are configurated:
- IPSEC/IKEV2 VPN with certificates for mobile clients and for site-to-site with the same hardware config on remote location. (no problem here)
- CAPSMAN with VLANs and multiple wireless networks. (no problem here)
- def network and added some Vlan to separate networks for servers, relatives, IoT, etc. (problem with def network)

Well ... in the network that is NOT in VLAN (usually use it to access routers)... I have constantly pkg loss from the device in this network (ping 8.8.8.8) - from the Mikrotik terminal, I don't have this problem. On all others Vlans network also - works normally.
So... maybe is some bad configuration.
For now, I switch to long-term to "downgrade" to 7.1.5... where everything works normally.
When I have time (And no working people from home) will export the config and investigate...

ARP list on LTE interface is filling up with all-Internet-wide IP's.
This ARP list is going to be endless.

Equipment:
wAPR-2nD with R11e-LTE (LTE firmware: MikroTik_CP_2.160.000_v020)
mipsbe v.7.2 (stable)

ARP list on LTE interface is filling up with all-Internet-wide IP's.
This ARP list is going to be endless.

That is likely a configuration mistake. You enabled proxy-ARP?

I have a strange behaviour with L3HW Offloading an IPv6. After enableing everything was smooth, but after accidently changing the bridge protocol from MSTP to none with L3HW active the problems begun.
The IPv6 connections were lost and even after deactivating L3HW offload and reboot it didn't came back.
My last try was factory reset all switches with L3HW support and built them up with terminal export without activating L3HW offload and from the beginning without STP.
But nothing changed. All switch chips without L3HW offload like RB5009 or RB4011 are working fine with IPv6 but all CRS switches where I had activated L3HW offload in the switch menu lost IPv6. IPv4 is working fine all the time.
It started with ROS 7.2 and didn't changed with 7.2.1. I have no idea what to do next.
Any suggestions?
Thank you!!

RouterOS version 7.2.1 has been released in "v7 stable" channel!


To upgrade, click "Check for updates" at /system package in your RouterOS configuration interface, or head to our download page: http://www.mikrotik.com/download

I know there are other ways to upgrade, via File manager and what not, but trying to follow written instructions is only giving me a DNS timeout. Tried both WebFig and WinBox
Solution?

Terminal worked just after the upgrade but after 1st reboot stopped. Cant copy paste

I've been experiencing this too (and was told it's a feature, but I think that's confused with another behavior. This is an actual issue which shouldn't be happening).
Turns out the problem was occurring for me when using the new windows 11 notepad (I'm on an insider Dev build too). If I copied from Microsoft word then it pasted fine in terminal. So it's something with the new notepad or winbox is not properly filtering out special formatting from clipboard. Try using word or a different source to copy and see if it is helps

Still having issues with CAPsMAN working with 7.2.1 on RB3011, CAP's fail to connect with timeout, also OpenVPN breaks, will not bring sessions up, roll back to 7.1.5 and everything comes back up fine.

I tried 7.3b33 and same result.

Something has changed between 7.1 and 7.2+ and its breaking my setup.. for now I'll hang back on 7.1.5.

ARP list on LTE interface is filling up with all-Internet-wide IP's.
This ARP list is going to be endless.

That is likely a configuration mistake. You enabled proxy-ARP?

For LTE interfaces there is no ARP config at all.
But looking to APR list there is a feeling exactly like in case of PROXY-ARP.
At this moment, according to ARP list, my LTE up-link is closely related with 1786 IP’s from all over Internet, but all on 1 common MAC address..

Maybe you configured it in BRIDGE mode instead of (NAT) ROUTING?
In that case there is probably little else it can do because the clients do not know about a gateway?

Terminal worked just after the upgrade but after 1st reboot stopped. Cant copy paste

I've been experiencing this too (and was told it's a feature, but I think that's confused with another behavior. This is an actual issue which shouldn't be happening).
Turns out the problem was occurring for me when using the new windows 11 notepad (I'm on an insider Dev build too). If I copied from Microsoft word then it pasted fine in terminal. So it's something with the new notepad or winbox is not properly filtering out special formatting from clipboard. Try using word or a different source to copy and see if it is helps

Thnx and im gonna try this workaround. Without terminal we cant work.

I upgraded from 6.49RC2 and only lost one (from three in total) static route, a simple one, but o so essentail.

It was one not to a gateway name but one to an IP. Till now, all works fine.

Without terminal we cant work.

Of course an alternative for terminal is to use an external terminal program like PuTTY. I use a terminal window on Linux and "ssh".
Ok, it cannot connect in MAC mode. But once you have your router configured that usually is not required.

4G USB internet via Huawei E3372 modem is not working on 7.2.1 (didn't try older 7.x). The router (RB 951G) was freshly installed via Netinstall. The ppp-out1 interface state became "running" just one time then never did that again (however it shown "Connected" but there was no IP address nor route added). After downgrading to 6.49.6 (via Netinstall) it works just fine. ppp-out1 is set identically (unchecked "Dial on demand" and entered the APN name). Strangely, 6.49.6 added the interface automatically in a disabled state as expected, but 7.2.1 did not, I had to add it myself. So for now I have to stay at v6 for now, which is sad because I hoped to use WireGuard.

Did it not create an LTE interface instead? (this depends on the firmware running on the E3372 I think, did you hack that or did it always use PPP?)

Again, the winbox connections keep dropping 2 seconds after having logged in. This is for most devices here, after upgrades to 7.2.1, and with the latest winbox.
This issue never existed with older winbox versions and older ROS releases (pre 7).
Super annoying. Constantly having to Clear cache and jump all sorts of hoops to get in. No way to find out what's actually wrong here.

Did it not create an LTE interface instead? (this depends on the firmware running on the E3372 I think, did you hack that or did it always use PPP?)

On the "General" tab it shows the "Port: usb1" and it was properly recognized as a USB device.
No I didn't hack it, just had a freshly installed router with default configuration, plugged in the USB modem, entered APN name and wanted to use it as usual :)
It always used PPP. On 6.x it always added the interface automatically, in disabled state, so I just unclicked "Dial on demand", entered APN name and clicked Enable, and the connection was always made automatically.
On 7.2.1 it just did nothing after plugging in, so I added "PPP Client" myself, and it also recognized USB, but was never able to connect.
It seems broken...

Sorry for my ignorance.

But if I read it correctly, my well-trusted RB493G can not be upgraded to version 7.
Is this true?

Thanks

False, but do it only if you really needs new features, or remain on 6.48.6 long-term

Without terminal we cant work.

Of course an alternative for terminal is to use an external terminal program like PuTTY. I use a terminal window on Linux and "ssh".
Ok, it cannot connect in MAC mode. But once you have your router configured that usually is not required.

So...... nope. Pasting in terminal from wordpad or word under windows 11, does not solve the problem. At pasting i have characters and sympols.

Is there any case that mikrotik can take a look at that plz?

Of course an alternative for terminal is to use an external terminal program like PuTTY. I use a terminal window on Linux and "ssh".
Ok, it cannot connect in MAC mode. But once you have your router configured that usually is not required.

So...... nope. Pasting in terminal from wordpad or word under windows 11, does not solve the problem. At pasting i have characters and sympols.

Is there any case that mikrotik can take a look at that plz?

Ah ok. It worked for me once that I tried, but I haven't done enough testing, clearly it's not consistent & there's still a problem. Let's wait hopefully MT fixes this in an upcoming update.

It seems that Mikrotik will have to work around a problem that is not theirs actually. While terminal inside winbox is something they've got to fix (use something like "paste special - unformatted text"), other terminal software (e.g. putty) is not their doing. And trying to tidy up some formating pseudocodes from console input is ... something in grey area. I guess they should do it never the less to avoid possible future vulnerabilities (similar to cleansing input from web page to prevent SQL injection attacks).

On device wAP R R11e LTE modem does not work (On 6.48.6 (long-term) works perfect):
Default configuration. factory settings on ROS6 and 7.

It seems that Mikrotik will have to work around a problem that is not theirs actually. While terminal inside winbox is something they've got to fix (use something like "paste special - unformatted text"), other terminal software (e.g. putty) is not their doing.

Guessing what I sometimes experienced when working with Windows (I try to limit that horror) is that it has pasted UTF-16 text into a program that expects 8-bit or at most UTF-8 characters... that won't work, obviously.

Minor cosmetic bug with 7.2.1, really can't disable some things with 'Design Skin'

step 1: trying to disable some types of interfaces on Interfaces/Add New
step 2: disabled on "Design Skin", hit "Save" on the Skin name, then exit "Design Skin". "Save" do *NOT* change the .json file size (or timestamp) on the disk, which doesn't seem OK
step 3: everything seems fine, but a logout/login will bring back all those options again, just like on step 1. I even tried editing the .json file, adding the options inside "Interface" section and uploading it to the MK, but got no success neither.
EDIT: some interface types, while working to disable via Design Skin, are being listed on an 'undefined' section on the .json file, while I believe they should be all under 'Interfaces'. Maybe this 'undefined' might point to something related to the problem

The LHG XL 52 ac reboots every five or two minutes after updating from 7.1.5 to 7.2.1

sdcard on modem e3372 can't detected by mikrotik

False, but do it only if you really needs new features, or remain on 6.48.6 long-term

Thanks for the reply.

I am trying to see and learn about new features of V7.

I am confused b/w what you meant as upgradable, and V7 upgrade error that i got. See attcahed image.



Would you be kinfd enough to give me some hints.
Many thanks.

I am confused b/w what you meant as upgradable, and V7 upgrade error that i got. See attcahed image.
Would you be kinfd enough to give me some hints.

You should first upgrade to the latest stable v6 (v6.49.6) so you get that ignore-missing option.
And REMEMBER TO PARTITION YOUR ROUTER (and copy the running version to the second partition) before you upgrade!

Updating an LTAP LTE6 from version 6.49.5 to 7.2.1, ROS upgrade was successful
After i upgraded the firmware, from 6.49.5 as well, the device upon boot had lost the LTE interface... A second reboot solved it...
Am not sure if that was supposed to happen or not ...

i can confirm that copy-paste in terminal from word is working, but not from wordpad. Same issues as pasting from notepad .rsc or .txt. Characters or whatever lines.........

i can confirm that copy-paste in terminal from word is working, but not from wordpad. Same issues as pasting from notepad .rsc or .txt. Characters or whatever lines.........

Yeah works for me too that way. Notepad++ also works, something to try. Are you on Windows 11? Maybe it's a problem with Windows 11 notepad... When pasting to winbox terminal some strange corruption occurs , terminal reveals strange lines with \b binary and contents which have nothing to do with the script being pasted, as if it's pulling the data from somewhere else, and complains that rules already exist even if it's fresh config. Glad you got it working with word too now

Although it would seem like a bug in winbox (using 3.35 64-bit) , it only happens with 7.2.1 router connections;
- showing a list of Active PPP Connections, if a connection disconnects and reconnects, it then shows 2 connections (old and new)
- if continued disconnect and reconnects happen, more sessions will show
- you cannot remove those 'phantom' connections
- under PPP/Interface, it shows correctly the number of connections
- disconnecting winbox and connecting back shows the correct active connections so in actuality they are not really even there
- not a problem with 6.x

i can confirm that copy-paste in terminal from word is working, but not from wordpad. Same issues as pasting from notepad .rsc or .txt. Characters or whatever lines.........

Yeah works for me too that way. Notepad++ also works, something to try. Are you on Windows 11? Maybe it's a problem with Windows 11 notepad... When pasting to winbox terminal some strange corruption occurs , terminal reveals strange lines with \b binary and contents which have nothing to do with the script being pasted, as if it's pulling the data from somewhere else, and complains that rules already exist even if it's fresh config. Glad you got it working with word too now

Yeap im on windows 11....... not only notepad, but also wordpad doesnt work for me

Seeing double again. Updated a hEX-S from 6.x to 7.21 recently and noticed that I have....again....double entries in columns:
Interface --> Ehternet

I expected this was resolved months ago but it is back. I cleared cache and the wiv file, and it still occurred.

update: saw this happening on more places in Winbox.

Seeing double again. Updated a hEX-S from 6.x to 7.21 recently and noticed that I have....again....double entries in columns:
I expected this was resolved months ago but it is back. I cleared cache and the wiv file, and it still occurred.

This problem has been present for years, but as now the columns are sorted by name instead of shown in the order they would appear in the listing, it has become more apparent.
Before, it was easier to see which one is "the setting" and which one is "the actively used value (status)" based on the grouping with other settings and statuses.
But now, you just have to try. Enable one, when it does not work enable the other.

You should first upgrade to the latest stable v6 (v6.49.6) so you get that ignore-missing option.
And REMEMBER TO PARTITION YOUR ROUTER (and copy the running version to the second partition) before you upgrade!

After updating to 6.49.6 and copied the active partition to the second one, I reran the upgrade but got no further than same error msg of:

So what I should do next, please?

It should have shown a "Ignore Missing" option, no idea why it does not do that for you.
However, you can do what it tells you there: disable those packages, then probably reboot first, then try the upgrade again.

The error message mentions you could disable the listed packages. Or you can uninstall them. Did you try doing any of it?

OpenVPN issue on hEX S

hEX S as OpenVPN client is not able to connect to OpenVPN server on CCR1009 if hEX S is on 7.2 or newer.
Issue is with authentication - it works only when Auth/Cipher on OpenVPN server is set to null.
Any other combination leads to error: <ovpn-OpenVPN-client>: terminating... - peer disconnected
When downgraded to 7.1.5 it works again without any OpenVPN modification - if upgraded to 7.2 or 7.2.1 it is broken again

And now for a positive report:

I spent a couple nights last week migrating a border router from a CCR2004 with 6.47.10 on it to a CCR2116 with 7.2.1.

Wow. What. A. Task.

After connecting the 2116 to the switch (a 317), I bonded two 10G ports and configured the VLANs in the 317. The 2116 had been up as a blank slate. I opened another terminal to the 2004, then spent a couple hours copying all the settings over in order, connecting up VLANs and VLAN interfaces but disabling them until later. Everything but BGP and OSPF settings & filters copied over nicely. The bulk of the time was spent translating the BGP and OSPF configs from v6 to v7. At trigger time I disabled the VLAN interfaces on the 2004 and turned them up on the 2116, then re-enabled the BGP peers.

I will say, I love how the CCR2116 is essentially a switch (like the CRS3XX series) with a 40Gbps pipeline to the CPU. It makes inserting these things into the mix in preparation for a migration very convenient. And what little experimentation I did with Layer 3 switching was phenomenal.

There's only one problem I have with L3 offloading. Currently the docs show how to do it on a per physical port basis, and how to make certain routes exempt for offloading. The NAT/FW examples shown in the documentation assume that you'll assign the world-facing IP(s) (that need NAT/FW capabilities) to a physical port.

My situation is such that I typically bring customer or ISP/BGP peer links into a central switch, then LAG from that switch to the router(s) using 2 or more connections (usually 10Gbps), allowing me to route the traffic to any router in the switch, which makes upgrades and scaling much more flexible. But there's no way I can see to tell L3-HW-Offloading how to exclude a VLAN from offloading, especially on a LAG. Since you have to enable L3 switching on all ports in order for it to work as expected, all NAT/FW filters are bypassed (they aren't assigned to a physical port that can be excluded).

For now, I can live with hw-switched Fasttrack connections, and the load went from the CCR2004's 30% on 6.47.10 to the CCR2116's 6% on 7.2.1 with the same amount of traffic (~600Mbps) + firewall and NAT for some locally-connected subnets.

If the chipsets allow it, it would be nice to filter offloading on a per-VLAN basis, at least so that initial packets can be punted to the CPU for FW/NAT processing.

Why this warning?

Why you do not read all before posting?

EDIT: real -> read

Why you do not real all before posting?

The problem is that "correct" setting is not apparent. And the correct setting since just recently is "auto".

In theory it's a good thing because it allows CPU clock scaling. When idle, it'll save a few miliwatts of electricity. When under load, it'll (short term) overclock to boost performance. In practice we'll see if auto setting can cause some weird side effects.

Suggestion to MT: if ROS can figure out that CPU is not running at default clock setting, why not display the "correct" setting as part of warning message?

Why you do not real all before posting?

The problem is that "correct" setting is not apparent. And the correct setting since just recently is "auto".

The correct setting never has been apparent. There should be some marker in the selection list which setting is the default.

The problem is that "correct" setting is not apparent. And the correct setting since just recently is "auto".

The correct setting never has been apparent.

Sure. But if one went into deep dive and found the specifications page (like @honzam did) and set the clock setting to that value, ROS was happy. But not any more. AFAIK auto setting is not universally available, it's there only for select (ARM-powered?) device models.

For @honzam

MARK II: Why you do not read all POSTS before posting?
viewtopic.php?p=927524#p926526

For @honzam

MARK II: Why you do not read all POSTS before posting?

Seems only few do, so why should @honzam be one of those rare people?

Plus I think that MT really have some room for improvement the CPU frequency scaling. Like @pe1chl noted, the default value could be clearly marked ... both in CLU and GUI. And both in setting part (in GUIS in drop.down liast of possible values, likewise in CLI) and in this warning message.

So perhaps we should actually bitch about this a bit more to draw attention of MT devs.

So perhaps we should actually bitch about this a bit more to draw attention of MT devs.

I think it is too trivial to pay much attention to it...

Suggestion to MT: if ROS can figure out that CPU is not running at default clock setting, why not display the "correct" setting as part of warning message?

Maybe the wording of the error message is not really correct with the introduction of the "auto" option.

What that error message says is that set cpu frequency value is not the same as default value, it doesn't read the actual CPU frequency to display the error message.

What that error message says is that set cpu frequency value is not the same as default value, it doesn't read the actual CPU frequency to display the error message.

Exactly. So ROS actually knows default value and it could include that value to the warning message ... to spare a few users from pulling their hair.

Also, red implies that it's a serious problem for it to be set to a value other than the default. Save red for emergency conditions.

Ideally it should rather only give a warning if neither the "specification" frequency OR auto is set (ie, user manually overclocked the router). No use showing a red warning if the user is running at factory frequency, even if that frequency is no longer the default in current software. Doesn't bother me, but should spare us all the noise from concerned users

+ 1

Cant upgrade x86. System go off. When trying starting it says :
Starting.....
soscket: Address family not supported by protocol
startind devices....
CHR can run only on x86_64

heh i try donwload iso and install .... same thing.

7.1.1 > 7.2.1

After upgrading RB760iGS (hEX S) from 6.49.6 to 7.2.1 sfp information is no longer visible with Mikrotik or OEM modules, although sfp seems to work.
The other problem is that if Hardware Offload is active on any port in the bridge with RSTP turned on the bridge stops forwarding packets...

Create a new routing table if you want to do policy routing. Or use a VRF if you need additional security for separating l3 networks.
https://help.mikrotik.com/docs/display/ ... g+Examples
https://help.mikrotik.com/docs/pages/vi ... eId=328206

Cant upgrade x86. System go off. When trying starting it says :
Starting.....
soscket: Address family not supported by protocol
startind devices....
CHR can run only on x86_64

heh i try donwload iso and install .... same thing.

7.1.1 > 7.2.1

Same problem with 7.2
Dell R420 without virtualization. Trying to install 7.2, 7.2.1 from iso x86. Instalation pass but cant run system. Cant upgrade 7.1.1 to 7.2 or 7.2.1. System not running after reboot. 7.1.1 working but sometimes reboot without reason....
Starting.....
soscket: Address family not supported by protocol
startind devices....
CHR can run only on x86_64

With the RB951Ui-2HnD it does not let me see the graphs of the CPU and memory usage

My wifi stopped working after upgrading to 7.2.1. it is always on obtaining IP address for both 2Ghz and 5Ghz. the Ac HP2.

It also reset my wifi settings for the 5Ghz
it also change the DHCP Server to Wifi1 instead Eth1.

How do I restore my WIFI please help!! I tried to reverse back to 7.1.5, it didn't help, it changed something in the settings and I don't know what.
It says IP configuration error when trying to connect, it always shows connection on Registration devices but they disconnect almost after 5-10 sec and it's never connected on my devices.

Just upgraded a map lite from 6.48.6 directly to 7.2.1. YOLO. Works like a charm. Consumes way more RAM, but nvm.

it also change the DHCP Server to Wifi1 instead Eth1.

Both devices are ports of a bridge? Set the DHCP server for that bridge!

detect internet stops working when you restart the router.

IPV6 Routing Table not work

Upgraded both my devices, seems ok.

Router RB1100AHx4 (RoS 7.2.1)
Switch CRS326-24G-2S+RM (RoS 7.2.1)

1. Detect Internet doesn't always work.

2. Router is straight up dropping or doing something with the ICMP traffic. I can't ping anything! In packets captured on the router, I only see ICMP Echo Request message and no responses whatsoever.

With the RB951Ui-2HnD it does not let me see the graphs of the CPU and memory usage


cpu.jpg

Grapphing.jpg

After restarting it several times and seeing that it was still the same, I turned it off and then turned it on and it was fixed.

1. Detect Internet doesn't always work.

+1.

I reported this back here a few times for ROS 7.x now, but it didn't get fixed . IIRC this is a feature that among other things the (default) firewall rules are based upon, so maybe worth looking into.

I have opened a support ticket ...

I reported this back here a few times for ROS 7.x now, but it didn't get fixed . IIRC this is a feature that among other things the (default) firewall rules are based upon, so maybe worth looking into.

No, the default firewall rules do not really require "detect internet". It is a toy that was added at some time maybe with the intention to extend it into a fully automatic configuration of the router (just plug it in in a random way and it will find out by itself what is your LAN and what is your WAN) but it wasn't really completed.
I think it is best if it is removed. It just causes annoying side-effects and is "calling home" which we do not want.

Lost all bridges and ipip interfaces after upgrading from 7.2 to 7.2.1 on hAP ac^2

Lost all bridges and ipip interfaces after upgrading from 7.2 to 7.2.1 on hAP ac^2

No need to report that anymore, it is just waived away with "very rare cases" and "we cannot reproduce that so we cannot do anything about it".
Sad, indeed.

Lost all bridges and ipip interfaces after upgrading from 7.2 to 7.2.1 on hAP ac^2

No need to report that anymore, it is just waived away with "very rare cases" and "we cannot reproduce that so we cannot do anything about it".
Sad, indeed.

It should be reported, but probably best to channel the incidents to this dedicated thread which support have created

Still having issues with CAPsMAN working with 7.2.1 on RB3011, CAP's fail to connect with timeout, also OpenVPN breaks, will not bring sessions up, roll back to 7.1.5 and everything comes back up fine.

I tried 7.3b33 and same result.

Something has changed between 7.1 and 7.2+ and its breaking my setup.. for now I'll hang back on 7.1.5.

Ok, I resolved the CAPsMAN issue, up to and including 7.1.5 my CAP's connected to the RB3011 fine on the Untagged/Native VLAN, starting with 7.2+ I needed to add the Tagged VLAN for the subnet that their primary IP was part of, no to the configuration on the PoE switch (Mikrotik CSS106-1G-4P-1S) that powers the CAPs, without this the CAPs could *see* the RB3011 as a controller but wouldn't successfully Provision showing timeout messages in the logs.

I've since added the tagged VLAN to my 7.1.5 configuration (in addition to the native vlan on the primary bridge) CAPsMAN Interface configuration so it doesn't trip me up in the future.

However, OpenVPN is *VERY* broken and unstable in 7.2.1 and the last two 7.3 beta's the session comes up but massive amounts of packet loss and sessions just simly falling apart randomly and in other cases pausing for minutes at a time then resuming for no reason whatsoever, no indication that the session is flapping though.

Dropping back to my 7.1.5 partition and OpenVPN performs reliably and predicably in TCP mode. .

I have been using TCP which should be stable, I know OpenVPN with UDP packets in 7.x has been a dumpster fire as far as usability goes so i've avoided that and gone for the known good / slower TCP option.

For now i'll stick to 7.1.5 on all units, that is stable enough on my hAPac2's (CAPs) and the RB3011.

Kind Regards,
Jim.

I was too excited about Wireguard on V7 so I upgraded all my site to V7 and done well with site to site VPN. And then suddenly I realized my hotspot page is not working and many people found the same problem here in the forum. So I have to downgrade all 5 sites back into V6.49 with OpenVPN and yes it was great experience in V7. Hope this will resolved soon.

try 7.3 beta34 hotspot is working with simple queue though i hate to admit i haven't thoroughly test it, earlier build whenever you put some limits it borks

Found a bug (on CCR2116, did not test on other devices) where, with "too many" VLAN interfaces / IP addresses configured, some IP do not respond anymore.
Disabling some of the IPs make the other ones available again.
With a rather random behavior...
SUP-80578 created accordingly.

Mikrotik, please have a look at Safe Mode. It has some serious flaws! Today I used Safe Mode to tinker around with some L2tp+ipsec tunnel config. Within that process I first disabled all firewall files rules and then even deleted them for testing purpose. I made several changes to l2tp and ipsec too. Then I disconnected from Winbox - "are you sure? your changes will be lost". Afterwards I reconnected to the device to see if all is good. But hell no! Firewall rules were back again, but ALL DISABLED. Holy moly!!!! This is SAFE MODE?

I experienced safe-mode issues in the past already. 5 changes and one random change was not reverted. I always thought that I may be wrong. But this time, holy crap, this thing fucked up my while firewall filter rules. I can now restore from export, as I can't tell if all rules are correct again. Holy moly.

Please please harden and battle-test safe-mode. This is crucial! Nobody wants to fuck up his configuration just because safe-mode has some serious bugs. This has to be top most of the top mostest priority of priorities.

edit:
I was not able to reproduce the bug. But I can't exactly remember in which order I made changes in safe-mode session.

What I know for sure is:
I have a disabled l2tp client interface. This was the first thing I enabled in safe-mode. while troubleshooting I adjusted firewall filter rules. Then I finally gave up (damn ipsec) and disconnected from winbox. Connected again. The l2tp interface was disabled again. Ipsec confs back. Good. But then it blew the sh** out of me, when I saw all firewall filter rules disabled. That frightened me to death.

I will repeat what my colleague wrote:

"This fix has nothing to do with the problem that some routers lose configuration"

We are working on that problem and it will be fixed in next release.

@infabo why so scared? I'm sure you did those tests on a non critical device that's used only for tests, as any sane person would do.

Znevna, why are you trolling?

With your argument you can drop safe mode. no need for that feature anymore.

because znevna claims:

* you should do your tests only on non-critical lab device. when you fail, you just do a netinstall. easy cheezy

And even such a statement is nonsense. Should I mistrust safe-mode and do an /export before and after using safemode? This would be ridicolous! That would need no safe-mode at all. I would just do my changes on the device and if it is not worky worky, I just restore a backup or compare and cherry-pick from the /export I made before I started to change configs. And then who says, that restoring a backup puts my device into the state it was before the backup?

I will repeat what my colleague wrote:

"This fix has nothing to do with the problem that some routers lose configuration"

We are working on that problem and it will be fixed in next release.

You are aware of safe-mode issues and working on a fix already?

That was what normis just did write.

You are aware of safe-mode issues and working on a fix already?

File a bug report.

Trying to add a VXLAN interface on rb5009 fails when trying to using a multicast group and interface, in my case bridged, with the error, "unable to set MTU".
Not using a multicast group works fine.

It works on older devices (MMIPs, MMIPBEs, etc.) that have been upgraded from 6.49.6 to 7.2.1, but not on the rb5009.

You are aware of safe-mode issues and working on a fix already?

File a bug report.

I certainly would if I could reproduce. Sadly I have no spare Chateau to tinker around until I can file a step by step bug report. Instead I am diffing my messed up config with my reference export to get things back on track. Safe mode, more like screwed-mode.

PIM-SM still not working on CCR2004-1G-12S+2XS

I reported this back here a few times for ROS 7.x now, but it didn't get fixed . IIRC this is a feature that among other things the (default) firewall rules are based upon, so maybe worth looking into.

I have opened a support ticket ...

The MikroTik support was able to reproduce the problem based upon my information, so a fix will be included in upcoming ROS versions.

Thanks!

Exciting times for v7. Devs have fixed some haunting/lingering bugs which means focus will gradually shift to more exciting and long-requested features going forward.

File a bug report.

I certainly would if I could reproduce. Sadly I have no spare Chateau to tinker around until I can file a step by step bug report. Instead I am diffing my messed up config with my reference export to get things back on track. Safe mode, more like screwed-mode.

Apparently fixed in 7.3beta37.

The safe mode issue (assuming such a thing even exists) is not the same thing as the config loss on package upgrade issue. I don't see anything about safe mode improvements in the v7.3beta37 release notes. Not sure why everyone (normis included, lol) is confusing these matters. The former is yet to be dealt with, the latter is allegedly resolved. Actually I would venture to say that v7.3beta37 should provide a more "stable" experience than 7.2.1 "stable", at least in theory in spite of the beta label, but I'm sure we'll see a new stable release follow up soon anyways as well. Don't take my word for it though, I'm not speaking from a position of knowledge/authority, just basing this on improvements in the changelogs

OSPF on CCR2116-12G-4S+ with FW 7.2.1 not work!

I have the same problem with 7.2.1 on a CCR1072 en a CCR2004.
Both work perfectly on 7.1.5, but OSPF crashes on 7.2.1

If OSPF instances enabled on boot, it is not working at all.
If i disable OSPF instances, reboot, enable instances it is working for a short time, but then it is crashing.

The tab Interfaces and Neighbors are empty, while Instances, Interface Template and Areas are correctly filled.

When i downgrade to 7.1.5 everything is working again.

Exciting times for v7. Devs have fixed some haunting/lingering bugs which means focus will gradually shift to more exciting and long-requested features going forward.

I'm still seeing port flapping on NAND activity with a CRS328. My support ticket has gone unanswered since Dec: SUP-68278. Hopefully they work on responding to support tickets with serious issues before they start working on new features.

OSPF on CCR2116-12G-4S+ with FW 7.2.1 not work!

I have it working on two CCR2116's, along with BGP and handling 600+Mbps of traffic each. Both routers are talking to each other as well as routers running 6.47.10.

What does your config look like?

Version 7.2.1 could be installed to XCP-ng server using x86 CD image iso but after VM stops on boot. The last output is "No errors found. Starting..." and then VM stops.

1100AHx2
netinstall 7.2.1
clear configuration with "no default configuration" checkbox
rename interface ethernet13 to ethernet13_
/export all interfaces renamed!
if to try import this fresh configuration got error:
failure: already have interface with such name

if to try import this fresh configuration got error:
failure: already have interface with such name

The order of rename commands in export is wrong. When the first rename command is executed, ether3 (with its default name) still exists.

While this does indicate problem with export file generation, it's a very hard problem because the export procedure would have to be very smart to get the order right in all cases. Alas, renaming interfaces just to shuffle the order is plain stupid IMO, this way it's very hard to notice that interface names don't correspond to markings on the chasis.

Well, likely (when his description is correct) he did not do those renames, the system came up with them by itself.
In that case it would be a bug. However, in case the story is not complete: it can also be the result of loading a backup file made on another device.
E.g. when you previously had a RB2011, made a backup on that, and load that on the RB1100, you get a mess like that.
(with the interface names not matching the markings on the chassis)
So don't do that. Ever.

Then, simply run this after restore :

OSPF on CCR2116-12G-4S+ with FW 7.2.1 not work!

I have it working on two CCR2116's, along with BGP and handling 600+Mbps of traffic each. Both routers are talking to each other as well as routers running 6.47.10.

What does your config look like?

On this moment we have a VLAN where 4 CCR routers are connected
router1, router2, router3 and router4.

It is a /24 subnet and they have a OSPF network type broadcast, for example 10.1.1.0/24
router1 and router2 are also doing VRRP for servers/clients in the same /24 subnet.
router3/4 doing PPPoE Server.

We have another 2 routers on remote side with a PTP OSPF, that is not giving a problem.

I have the idea (but not sure) the problem starts when i enable the VRRP interface.
Is it recommended to separate OSPF in a separate VLAN and another subnet?

There are other places where MAC addresses are being used (wlan, bridge, ...)

For the first, there must be a similar command for resetting all wireless MACs to their defaults.

For the second, the bridge MAC is equal to one of the interfaces in the bridge, by default under control of the auto-mac setting. Resetting the interface MACs will make the bridge select a new MAC.