In my situation. Two RB5009, same ISP, on both ends symetrical 800Mbps links. When I’m uisng IPSEC to transfer large files i see on both routers that one core is going to 100% and speed over IPSEC is no more than 280Mbps, when before i was easly reaching 600Mbps. I didn’t change any settings. It looks like HW acceleration is not working at all.
Any suggestions?
Try latest stable build?
I did today. I installed 7.22.2 and boom - issue fix.
Correct me if i’m wrong but long term should be the most tested and most stable release, something like gold star version in Cisco. I don’t know how Mikrotik could promote such bugged version to be long-term.
Maybe you need to consult the dictionary:
Mikrotikish<->English conversion table/dictionary
(now you can also listen to the Mikrotik song while reading it) 
God damm. Now all is clear!!! Thank you!
Every software is buggy, No software is bug-free.
IMHO, “long-term” new version should be released each time there are max 3-5 bugs fixed ( but wtih 0% regression). I
Noone is “forced” to upgrade at each release, but we should be able to buid confidence and be certain there are no REGRESSIONS.
But with - 0 new features, 0,0000 regressions and max 5 bugs fixed.
people could then decide it they apply the update or wait ( eg: because they are not impacted by fixed bugs…)
Exactly. How it’s possible to breake IPSec acceleration which was working flawless for so many versions? It’s look that Mikrotik don’t have even basic lab to test software before relaese and lack of participation on this forum more look like they hired too many interns with bad attitude.
I’m using MIkrotik routers (i don’t like switches) for years and still i have in cabinet RB2011 on which i was learning. On ROS v6 it was much better, sometimes i found bugs but in general it was working without issues. When ROS V7 pop up, first releases was absolutly unusable due to random and idiotic bugs but it was marked as stable.
IMHO, Mikrotik should stop playing with idiotic containers and fix main router features first! This is core functionality of those devices. Next if for examle current stable 7.22.2 will proof to be reliable after long testing and observing forum than such version should be promoted to long-term.
Some people don’t understand that it’s almost impossible to make lab to imitate whole organization. I have more than 100 routers and the oldest is RB3011 and RB750 and newest RB5009. If i installing new software i really undestand that some bugs can be present for NEW features - sure, but old features like OVPN (many of us remeber what they did on HexS and i need to send router on second end of world because due to bug it was constantly rebooting), IPSec acceleration should be rock solid!
I know that Ubiquiti routers are beyond MIkrotik, especailly regarding possibilities and price, but damm - i have few of them and update after update works fine and getting new functions. Once or twice they f… something like recently protect but they are fixing it fast. Mikrotik looks like they don’t care and using userers like test enviroment.
If it was this easy for MikroTik to release "0 regression", MikroTik would release stable version without regressions as well. That is has nothing to do with long term or such. It is a skill you have to build/develop in your software release cycle. And MikroTik seems to skip that part. Most probably the do some kind of smoke testing, as you rarely never see broken devices after upgrade. But I can't see proper testing of newly added features, as you always see a lot of "additional fixes" and "introduced in" fixes. That has one common source: insufficient testing.
the more you put in it, the higher the regression risk is… and the more complicated it is to assess if the risk of upgrading is worth taking
So now this "stable" seems to be more "long term" build then current LongTerm it seems.. After two days, i dont read serious bugs with this build or im wrong?
One would hope that they have a test setup where new releases are installed and a number of tests are automatically run before they are released. Ideally there would be a test not only for basic functionality, but also for each and every problem (bug) that has been fixed in the past.
Probably there is something like that, but the coverage of tests is not optimal. I can understand that something like “IPsec hardware acceleration no longer works” (probaby only under certain configurations and on some subset of the hardware) would not be part of those tests, but after this problem has been fixed ideally it should be added, so this cannot happen again.
It likely is a lot of work to establish and maintain such a test setup, and understandable that (like documentation) it gets less priority than it should. Then maybe that is a difference between getting a lot of functionality and free upgrades for a low price, and getting less functionality for 5 times the price + mandatory support contract to get updates after purchase?
I might say something that sounds “unpopular” regarding to the actual topic, but it can be that business model of the company can limit the possibilities for software development and tesing.
Vendors who have NGFW products do have software licensing and cloud service subscriptions in their revenue stream while Mikrotik does not. That can set its’ own limitations to what you get…
HEX S original. Running stable for several years. Latest ROS 7.20.8 long term. Decided to upgrade to latest long term.
Check for updates
Download / Install
Lost connection and router never came back.
All lights working (power / LAN). Turn off when unplug, turn on when plug in.
Other than that, nothing. Not showing up in neighbors with static IP set.
Cannot connect via DHCP. Connection fails.
Really regret doing that upgrade.
Reset button - device restarted, but still inaccessible.
Unplug wall wart. Wait. Plug in. No change.Still inaccessible.
Ended up having to netinstall, which did work.
Will make me think 100x before doing an upgrade unless absolutely need it.
Hard to see what problems others had as most of the discussion seems to be unrelated to this specific version update.
On the other hand if you already have established the business model that does not provide for a vast developer resources you probably shouldn't be playing with every single feature available on the market (NAS, containers, proxy etc...) and trying to push them in to routers but instead focus on core network functionalities that you actually can support with what you have and that most of your users really depend on...
That as well - but it looks that if business model is mostly based on hardware sales then only way to increase the revenue is to try to expand the market share. One of the ways is expanding to new market segments, but that starts to add to development costs.
As for me, making Eierlegende Wollmilchsau from MikroTik devices is not a right way.
Especially when it turns a Wolpertinger....
Mikrotik does though? They sell Cloud Hosted Router (CHRs) licences.