# v7.16rc \[testing\] is released!

**URL:** <https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857>\
**Category:** Announcements\
**Created:** [August 6, 2024, 7:57am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857 "2024-08-06T07:57:22Z")\
**Posts on this page:** 20\
**Page:** 16

<div class="post-metadata">

**Author:** ![pe1chl](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/pe1chl/32/55588_2.png) [@pe1chl](https://forum.mikrotik.com/u/pe1chl)\
**Post date:** [September 19, 2024, 1:43pm UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/301 "2024-09-19T13:43:21Z")

</div>

> [@](#):
>
> This “software based offload” looke like multithreading to me in this case…

Multithreading is often not useful in these cases because at least a single connection will have to be processed in sequence.  
I.e. when you have a single PPPoE client connected to your ISP via a single network interface, there is not much that can be multithreaded.  
Of course when you have two ISP each with a PPPoE client, they can be distributed over different cores, at least when the hardware allows it.  
(unfortunately in todays routers there often is a switch chip connected to the CPU with only a single connection, and all incoming traffic is handled the same, no matter what interface it arrived on)

---

<div class="post-metadata">

**Author:** ![SystemErrorMessage](https://forum.mikrotik.com/letter_avatar/systemerrormessage/32/6_449c533f586939b6e5f246632a46421f.png) [@SystemErrorMessage](https://forum.mikrotik.com/u/SystemErrorMessage)\
**Post date:** [September 19, 2024, 1:55pm UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/302 "2024-09-19T13:55:08Z")

</div>

> [@](#):
>
> > [@](#):
> >
> > This “software based offload” looke like multithreading to me in this case…
> 
> Multithreading is often not useful in these cases because at least a single connection will have to be processed in sequence.  
> I.e. when you have a single PPPoE client connected to your ISP via a single network interface, there is not much that can be multithreaded.  
> Of course when you have two ISP each with a PPPoE client, they can be distributed over different cores, at least when the hardware allows it.  
> (unfortunately in todays routers there often is a switch chip connected to the CPU with only a single connection, and all incoming traffic is handled the same, no matter what interface it arrived on)

in programming thats not the case. you can have multiple queues bound to a single io. that io can be its own thread, the other queues can be their own threads.  
In python it is recommended to thread your IOs/request and multiprocess anything CPU intensive. multi threading is bound to a single process but can better use multiple cores to handle io tasks even if its to a single destination.

---

<div class="post-metadata">

**Author:** ![pe1chl](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/pe1chl/32/55588_2.png) [@pe1chl](https://forum.mikrotik.com/u/pe1chl)\
**Post date:** [September 19, 2024, 2:25pm UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/303 "2024-09-19T14:25:36Z")

</div>

The point is that when you use constructs that allow you to process a single stream of information in multiple threads, and also you need to guarantee that the processing is done in sequence, you are only wasting effort on thread synchronization that is not made up for in the increased performance of multiple cores.  
Remember the PPPoE processing of data packets is trivial. It is not like processing a web request, or similar.

---

<div class="post-metadata">

**Author:** ![CoMMyz](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/commyz/32/1578_2.png) [@CoMMyz](https://forum.mikrotik.com/u/CoMMyz)\
**Post date:** [September 19, 2024, 3:01pm UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/304 "2024-09-19T15:01:01Z")

</div>

Do not confuse PPPoE and multi-threading. Its not the PPPoE that does not get multi-threaded.

Each distinct flow/connection that passes through the PPPoE can get multi-threaded but only if there is no active Simple Queue on it.  
Once you enable any kind of queuing the entire interface will pass through a single core as the CPU must measure all flows in order to be able to queue them.  
If they multi-thread this scenario - each process would have to sync each connection with the other cores to avoid out-of-sync flows effectively creating more harm than good as this is costly for a CPU to do.

These scenarios affect all protocols not only PPPoE.

---

<div class="post-metadata">

**Author:** ![fischerdouglas](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/fischerdouglas/32/2357_2.png) [@fischerdouglas](https://forum.mikrotik.com/u/fischerdouglas)\
**Post date:** [September 19, 2024, 5:59pm UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/305 "2024-09-19T17:59:08Z")

</div>

> [@](#):
>
> Do not confuse PPPoE and multi-threading. Its not the PPPoE that does not get multi-threaded.
> 
> Each distinct flow/connection that passes through the PPPoE can get multi-threaded but only if there is no active Simple Queue on it.  
> Once you enable any kind of queuing the entire interface will pass through a single core as the CPU must measure all flows in order to be able to queue them.  
> If they multi-thread this scenario - each process would have to sync each connection with the other cores to avoid out-of-sync flows effectively creating more harm than good as this is costly for a CPU to do.
> 
> These scenarios affect all protocols not only PPPoE.

This is a good point, and a question that needs to be properly validated…

I’m sure I’ve faced several situations of B-RAS in PPPoE without the use of QUEUEs in which subscribers didn’t go beyond 800-900 Mbps.  
And during diagnostics we found that when the subscriber tested bandwidth, one of the Cores was at 100% and the rest of the cores were practically idle.  
I reiterate that this scenario was WITHOUT THE USE OF QUEUEs.

In one of these cases I made a point of taking the subscriber’s traffic to B-RAS in L2 and doing a proof of concept by changing from PPPoE to Static IP directly assigned to the interface. And during these changes the bandwidth tests reached 1Gbps without any problems. In addition, there was no CPU imbalance stuck on just one of the cores.

---

<div class="post-metadata">

**Author:** ![nz\_monkey](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/nz_monkey/32/430_2.png) [@nz\_monkey](https://forum.mikrotik.com/u/nz_monkey)\
**Post date:** [September 20, 2024, 12:56am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/306 "2024-09-20T00:56:55Z")

</div>

> [@](#):
>
> > [@](#):
> >
> > I think we may be talking about different use cases.
> > 
> > Which other manufacturers and products are you comparing Mikrotik to ?
> 
> All those manufacturers that make the routers that ISPs supply with the 1Gbps+ internet connections they deliver today.  
> AVM, Arcadyan, Linksys, etc etc etc.

Those are all CPE type routers.

My main concern and use case is for Service Provider BNG’s, terminating hundreds, thousands or tens of thousands of connections.

---

<div class="post-metadata">

**Author:** ![fischerdouglas](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/fischerdouglas/32/2357_2.png) [@fischerdouglas](https://forum.mikrotik.com/u/fischerdouglas)\
**Post date:** [September 20, 2024, 7:07am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/307 "2024-09-20T07:07:50Z")

</div>

> [@](#):
>
> Which other manufacturers and products are you comparing Mikrotik to ?

Search by Marvel on those links…

[https://isp-tech.ru/en/switch-asic/#](https://isp-tech.ru/en/switch-asic/#)  
[https://sonic-net.github.io/SONiC/Supported-Devices-and-Platforms.html](https://sonic-net.github.io/SONiC/Supported-Devices-and-Platforms.html)  
[https://dent.dev/ecosystem/hardware-compatibility/](https://dent.dev/ecosystem/hardware-compatibility/)  
  
  
Marvell Chips are used by several big vendors, including Juniper and Cisco.  
Just recently (after 2018, after buying Cavium) Marvell started to appear more on the market with their own name.  
Usually, each switch vendor that the chip would come with a name defined by this buyer.  
Before OCP, it was considered ill-favored not using “your own” chips.

I remember (long time ago) taking in my hands two line-cards of Cisco chassi device, same part-number of Cisco, and in one of those the chip had the Cisco logo and name, and the other was identical except that in the chip in the other line card had Marvell wrote on it.  
I, in my ignorance, even considered that was a falsification… Not! Just production bath different.

Marvell is not the Godzilla that Broadcom is, but is a very respectful solution.  
MikroTik that do not use all the potencial of Marvell chips.

We all should push then to use better the resources of the solution they chose themselves.

---

<div class="post-metadata">

**Author:** ![pe1chl](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/pe1chl/32/55588_2.png) [@pe1chl](https://forum.mikrotik.com/u/pe1chl)\
**Post date:** [September 20, 2024, 7:40am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/308 "2024-09-20T07:40:49Z")

</div>

> [@](#):
>
> > [@](#):
> >
> > All those manufacturers that make the routers that ISPs supply with the 1Gbps+ internet connections they deliver today.  
> > AVM, Arcadyan, Linksys, etc etc etc.
> 
> Those are all CPE type routers.
> 
> My main concern and use case is for Service Provider BNG’s, terminating hundreds, thousands or tens of thousands of connections.

Ok but my concern is MikroTik selling CPE type routers that underperform relative to other manufacturer’s routers, despite using the same hardware. People are struggling with their hAP type routers getting 1Gbps performance, while the ISP-branded white box CPE router does the same thing without problem.  
That is something that MikroTik should be able to fix and make more customers happy than working for the rare ISP that uses MikroTik for a PPPoE service.

---

<div class="post-metadata">

**Author:** ![Guntis](https://forum.mikrotik.com/letter_avatar/guntis/32/6_449c533f586939b6e5f246632a46421f.png) [@Guntis](https://forum.mikrotik.com/u/Guntis)\
**Post date:** [September 20, 2024, 7:49am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/309 "2024-09-20T07:49:43Z")

</div>

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

---

<div class="post-metadata">

**Author:** ![Rox169](https://forum.mikrotik.com/letter_avatar/rox169/32/6_449c533f586939b6e5f246632a46421f.png) [@Rox169](https://forum.mikrotik.com/u/Rox169)\
**Post date:** [September 20, 2024, 8:14am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/310 "2024-09-20T08:14:23Z")

</div>

2 Guntis: Whats wrong with this RC, when we can expect stable version?

---

<div class="post-metadata">

**Author:** ![merkkg](https://forum.mikrotik.com/letter_avatar/merkkg/32/6_449c533f586939b6e5f246632a46421f.png) [@merkkg](https://forum.mikrotik.com/u/merkkg)\
**Post date:** [September 20, 2024, 8:30am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/311 "2024-09-20T08:30:51Z")

</div>

@Guntis,

People post stuff everywhere and anywhere to just be seen or heard on their issues they have and that are affecting their business

- SUP Tickets are created no response
- Forum threads are created and no comments to these specific issues from any Mikrotik person.

All we ask is for communication and “possible timelines” and i’m sure things will be kept to correct topics.

---

<div class="post-metadata">

**Author:** ![SystemErrorMessage](https://forum.mikrotik.com/letter_avatar/systemerrormessage/32/6_449c533f586939b6e5f246632a46421f.png) [@SystemErrorMessage](https://forum.mikrotik.com/u/SystemErrorMessage)\
**Post date:** [September 20, 2024, 8:39am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/312 "2024-09-20T08:39:58Z")

</div>

on the Rb5009 the 7.16rc4 keeps rebooting itself every few hours. at this point i’ve pretty much had it with the instabilities. i have exams coming up and i need a reliable internet connection. The only thing i’ve used are vlans, ip segmentation, HDD (samba), a container (adguard but not in use anymore). It was so bad in 7.15 but its still not stable yet.  
I’ve been powering it with a bench PSU so voltage/power isnt the problem.

I tried downgrading but that didnt even work, i followed the documentation and the router still did not downgrade itself.

---

<div class="post-metadata">

**Author:** ![Kanzler](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/kanzler/32/3006_2.png) [@Kanzler](https://forum.mikrotik.com/u/Kanzler)\
**Post date:** [September 20, 2024, 8:45am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/313 "2024-09-20T08:45:51Z")

</div>

@SystemErrorMessage,  
Have you sent a supout to support?

---

<div class="post-metadata">

**Author:** ![infabo](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/infabo/32/55692_2.png) [@infabo](https://forum.mikrotik.com/u/infabo)\
**Post date:** [September 20, 2024, 10:39am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/314 "2024-09-20T10:39:31Z")

</div>

> [@](#):
>
> 2 Guntis: Whats wrong with this RC, when we can expect stable version?

There must be some showstopper(s) apparently. If there are upcoming changes/fixes I hope we se another RC instead of going straight stable. Too high the risk of regressions in last-minute fixes.

---

<div class="post-metadata">

**Author:** ![eworm](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/eworm/32/1441_2.png) [@eworm](https://forum.mikrotik.com/u/eworm)\
**Post date:** [September 20, 2024, 11:00am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/315 "2024-09-20T11:00:32Z")

</div>

Wondering if it is what ever kind of regression or something in relation to Winbox 4… Anybody can shed some light on this?

---

<div class="post-metadata">

**Author:** ![cowgirl](https://forum.mikrotik.com/letter_avatar/cowgirl/32/6_449c533f586939b6e5f246632a46421f.png) [@cowgirl](https://forum.mikrotik.com/u/cowgirl)\
**Post date:** [September 20, 2024, 11:07am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/316 "2024-09-20T11:07:55Z")

</div>

3 weeks passed. Its time for RC5 or a Release…

---

<div class="post-metadata">

**Author:** ![kJpermol](https://forum.mikrotik.com/letter_avatar/kjpermol/32/6_449c533f586939b6e5f246632a46421f.png) [@kJpermol](https://forum.mikrotik.com/u/kJpermol)\
**Post date:** [September 20, 2024, 11:16am UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/317 "2024-09-20T11:16:04Z")

</div>

On CRS112-8P-4S, webfig menu switch menu vlan does not show up. It also does not show up on winbox 4beta. Reverting to 7.15.3 corrects the problem.

---

<div class="post-metadata">

**Author:** ![Paternot](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/paternot/32/1646_2.png) [@Paternot](https://forum.mikrotik.com/u/Paternot)\
**Post date:** [September 20, 2024, 1:44pm UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/318 "2024-09-20T13:44:13Z")

</div>

> [@](#):
>
> 3 weeks passed. Its time for RC5 or a Release…

Pretty please, no. Just no. A new version (RC or stable, doesn’t matter) should be released ONLY when finished. I’d rather wait 2 months for a new RC than keep battling bugs later.

---

<div class="post-metadata">

**Author:** ![fischerdouglas](https://forum.mikrotik.com/user_avatar/forum.mikrotik.com/fischerdouglas/32/2357_2.png) [@fischerdouglas](https://forum.mikrotik.com/u/fischerdouglas)\
**Post date:** [September 20, 2024, 2:15pm UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/319 "2024-09-20T14:15:41Z")

</div>

> [@](#):
>
> > [@](#):
> >
> > 3 weeks passed. Its time for RC5 or a Release…
> 
> Pretty please, no. Just no. A new version (RC or stable, doesn’t matter) should be released ONLY when finished. I’d rather wait 2 months for a new RC than keep battling bugs later.

i++;

---

<div class="post-metadata">

**Author:** ![riv](https://forum.mikrotik.com/letter_avatar/riv/32/6_449c533f586939b6e5f246632a46421f.png) [@riv](https://forum.mikrotik.com/u/riv)\
**Post date:** [September 20, 2024, 2:23pm UTC](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857/320 "2024-09-20T14:23:39Z")

</div>

BGP missing ibgp-rr-client local.role

[Previous page](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857.md?page=15)

[Next page](https://forum.mikrotik.com/t/v7-16rc-testing-is-released/177857.md?page=17)
