RouterOS v6.28 released

Wed Apr 29, 2015 11:59 am

The simple problem is TCP over NV2 in a PTMP environment is increasingly painful...
PTMP environment... that was the significant information that was missing....
Please, make sure that you point that out.. cause PtP it works as it should.

No. If your traffic passes multiple ptp connections your tcp-speed decreases with every nv2 ptp hop. So the problem ist with ptp, too.

Also there is no need to refer to this problem in every unrelated forum topic, it doesn't work - you need to write to support@mikrotik.com

They know the problem but it has no priority due to not enough users urge them to work on that. Writing to a forum is "urge them". Your statement is: Everything is fine just ignore this.

In any case PTMP NV2 is not the most common config, especially with 802.11ac around. So on global scale significance of the fix vs. time to fix it simply doesn't cost the effort. So you might need to work around it.

There is no solution to work around. You can try to use nstreme but nstreme is very sensitive to interference so you get faster tcp-speeds but with disconnects.

Wed Apr 29, 2015 12:14 pm

No. If your traffic passes multiple ptp connections your tcp-speed decreases with every nv2 ptp hop. So the problem ist with ptp, too.

Over saturated links - no problem. I have 5 ptp hops
If link is almost empty, then i observe the same decrease, but it is only logical.

They know the problem but it has no priority due to not enough users urge them to work on that. Writing to a forum is "urge them". Your statement is: Everything is fine just ignore this.

Sorry, but Hijacking topics is simply impolite! Create a topic and spam in it as much times as you like!

Lakis · Wed Apr 29, 2015 2:52 pm

macgaiver: 'cause PtP it works as it should' - using nv2 hm... 99% I am not sure about that, well you should try nstreme and see how it should work - except disconnect problem (yep ste:nstreme is very sensitive to interference) but noting is done past 2 years.
And again Mikrotik please give me a simple answer when AC txpower will be fixed?

Wed Apr 29, 2015 3:09 pm

macgaiver: 'cause PtP it works as it should' - using nv2 hm... 99% I am not sure about that, well you should try nstreme and see how it should work - except disconnect problem (yep ste:nstreme is very sensitive to interference) but noting is done past 2 years.
And again Mikrotik please give me a simple answer when AC txpower will be fixed?

Just stop spamming

bclewl1ns · Thu Apr 30, 2015 5:52 am

Well it looks like i touched a nerve. Look I believe that Mikrotik is a fantastic system. But I am noticing an ongoing problem with NV2 and TCP connections. And as everyone here knows TCP works in multiple streams However. When your delivering a 25 Meg Connection to a user and they get on a speediest server and continually complain about speed. It becomes and issue. Especially when UBNT does great at those tests. I have also submited trouble tickets on tcp and nv2 speed problems and you get no response the only place you have to URGE a response is in the forums. There are also MANY posts on this in the forums. To draw some attention we post in the New Release forums. It is not an attempt to complain to complain or to spam for spamming sake it is an attempt to let the company who I have spend Thousands of Dollars on know that it is a REAL PROBLEM. We invest in this infrastructure we rely on it. We make our livings off of it. Please explain to me why in a forum on RouterOS release and beta testing it is considered spamming a topic to discuss a PROBLEM with ROUTEROS.

Lakis · Thu Apr 30, 2015 11:50 am

Well it looks like i touched a nerve. Look I believe that Mikrotik is a fantastic system. But I am noticing an ongoing problem with NV2 and TCP connections. And as everyone here knows TCP works in multiple streams However. When your delivering a 25 Meg Connection to a user and they get on a speediest server and continually complain about speed. It becomes and issue. Especially when UBNT does great at those tests. I have also submited trouble tickets on tcp and nv2 speed problems and you get no response the only place you have to URGE a response is in the forums. There are also MANY posts on this in the forums. To draw some attention we post in the New Release forums. It is not an attempt to complain to complain or to spam for spamming sake it is an attempt to let the company who I have spend Thousands of Dollars on know that it is a REAL PROBLEM. We invest in this infrastructure we rely on it. We make our livings off of it. Please explain to me why in a forum on RouterOS release and beta testing it is considered spamming a topic to discuss a PROBLEM with ROUTEROS.

ste quote "Just stop spamming

" it was just a joke

look bclewl1ns NV2 does not work like that, not like on Mikrotik web "Performance test results", in real world u can not expect in PTMP situation AP to deliver 150mbit to station. And in some parts u are right nv2 needs lot of improvements
especially for the P2P situation

turnip · Thu Apr 30, 2015 4:02 pm

I can confirm this - updated 3 routers to 6.28 and enabled RoMON. Tested it then switched it off. CPU usage sits at 100% for single processor (CRS125) and 50% for dual core (x86) router. Switched it back on and disabled all RoMON ports, and it's back to normal.

ROS 6.28, RB2011UAS-2HnD

try this

1) Tools -> RoMON, check ENABLE, click OK

2) Tools -> RoMON, uncheck ENABLE, click OK

CPU usage will be 100%, you must enable it again.

mnemo · Sat May 02, 2015 8:46 pm

I have a RB850Gx2 and when I update from v6.26 to v6.28 with the upgrade package I had a problem becouse the multicast package disappeared.

The solution is to load the muticast package from "All packages" package.

Please, solve it in the next update.

rpingar · Sat May 02, 2015 10:10 pm

The simple problem is TCP over NV2 in a PTMP environment is increasingly painful...
PTMP environment... that was the significant information that was missing....
Please, make sure that you point that out.. cause PtP it works as it should.
No. If your traffic passes multiple ptp connections your tcp-speed decreases with every nv2 ptp hop. So the problem ist with ptp, too.

Also there is no need to refer to this problem in every unrelated forum topic, it doesn't work - you need to write to support@mikrotik.com
They know the problem but it has no priority due to not enough users urge them to work on that. Writing to a forum is "urge them". Your statement is: Everything is fine just ignore this.

In any case PTMP NV2 is not the most common config, especially with 802.11ac around. So on global scale significance of the fix vs. time to fix it simply doesn't cost the effort. So you might need to work around it.
There is no solution to work around. You can try to use nstreme but nstreme is very sensitive to interference so you get faster tcp-speeds but with disconnects.

I have the same exact experience,
tcp single stream is a very big issue in umltiple nv2 hop and in a ptmp.

nstream is very sensitive and with disconntions, only latest release and ac hardware seems little bit better.

Hope MT can get better.
But lately (2 years) they seems more focused on other part of the business.

regards
Ros

bclewl1ns · Sun May 03, 2015 7:49 am

Well it looks like i touched a nerve. Look I believe that Mikrotik is a fantastic system. But I am noticing an ongoing problem with NV2 and TCP connections. And as everyone here knows TCP works in multiple streams However. When your delivering a 25 Meg Connection to a user and they get on a speediest server and continually complain about speed. It becomes and issue. Especially when UBNT does great at those tests. I have also submited trouble tickets on tcp and nv2 speed problems and you get no response the only place you have to URGE a response is in the forums. There are also MANY posts on this in the forums. To draw some attention we post in the New Release forums. It is not an attempt to complain to complain or to spam for spamming sake it is an attempt to let the company who I have spend Thousands of Dollars on know that it is a REAL PROBLEM. We invest in this infrastructure we rely on it. We make our livings off of it. Please explain to me why in a forum on RouterOS release and beta testing it is considered spamming a topic to discuss a PROBLEM with ROUTEROS.
ste quote "Just stop spamming " it was just a joke

look bclewl1ns NV2 does not work like that, not like on Mikrotik web "Performance test results", in real world u can not expect in PTMP situation AP to deliver 150mbit to station. And in some parts u are right nv2 needs lot of improvements
especially for the P2P situation

Not sure were you got the 150M in my posts. And I am aware that NV2 Does not work like that and I can never expect 150 Meg on a PTMP link I am looking back and can't find were in this thread i have stated that. However, their are items that need to be addressed.

1. Latency - Even using auto on the NV2 Setting latency is way to high.
2. Jitter - It is increasely getting worse on NV2 Links
3. TCP Throughput Should be more stable and better then 10% of UDP

I am aware of NV2's benefits and its shortcomings. My only POINT is that instead of doing all these new features LETS Stabilize a issue that is effecting current users.

ibm · Sun May 03, 2015 11:50 am

1. Latency - Even using auto on the NV2 Setting latency is way to high.
2. Jitter - It is increasely getting worse on NV2 Links
3. TCP Throughput Should be more stable and better then 10% of UDP

I am aware of NV2's benefits and its shortcomings. My only POINT is that instead of doing all these new features LETS Stabilize a issue that is effecting current users.

+1

spacemind · Mon May 04, 2015 2:00 am

Hello,

Upload rate in queues is shown in bits instead of kb alson on this version.
The queues are set in radius.

Chupaka · Mon May 04, 2015 3:11 pm

Hello,

Upload rate in queues is shown in bits instead of kb alson on this version.
The queues are set in radius.

1228800 bps is 1228,8 kbps, and WinBox does not round limits in output. for example, 1228000 should be shown as 1228k

ddejager · Mon May 04, 2015 4:32 pm

Issue with sector writes rising will be fixed in version 6.29. If you want to test it, then fix for it will be included in RouterOS starting from 6.29rc13 version when it comes out.

This item does not seem to be listed in the changelog for 6.29. Is it, in fact, fixed? If it will take a long time for 6.29 to be released (due to the addition and testing of fast-track) can this fix be provided independently sooner?

[Update: it is showing up in the changelog now, perhaps I just missed it earlier]

bosuagwu · Mon May 04, 2015 4:48 pm

is quite cool. Nice work.

usmanahmed · Tue May 05, 2015 12:50 pm

Dear Mikrotik Support,

I have CR1036-12G-4S on version 6.28 and we are facing issue in guest page (hotspot) page that its not opening. we have to restart the router to resolve the issue. please look into it and if any logs is required from my side, please let me know, i will provide the logs on this forum

Usman Ahmed

PastuhMedvedey · Tue May 05, 2015 2:26 pm

jan/02/1970 00:41:33 system,error,critical System rebooted because of kernel failure
jan/02/1970 00:41:33 system,error,critical Out of memory condition was detected
jan/02/1970 00:41:33 system,error,critical router was rebooted without proper shutdown
jan/02/1970 00:41:37 interface,info sfp-sfpplus1 link up (speed 10G, full duplex)
jan/02/1970 00:41:38 interface,info sfp-sfpplus2 link up (speed 10G, full duplex)

During the test (traffic generator), the router reboots spontaneously. The test involved two 10G interface.
Router - CCR-1036 8g-2s+ . Supout file sent to support.

rextended · Tue May 05, 2015 9:55 pm

Now are 15 days after I have installed the 6.28 on my border BGP routers, my gateways, firewalls and user-managers.

No one single problem, except I must reboot one time each devices for update the software from 6.27 to 6.28

MikroTik: Is possible to update the software without reboot the machines?

Chupaka · Wed May 06, 2015 12:54 am

Now are 15 days after I have installed the 6.28 on my border BGP routers, my gateways, firewalls and user-managers.

No one single problem

on 17th day one of my CCRs became loosing packets on _some_ neighbour routers until reboot... I hope it was just a Moon in wrong phase

wfebbo · Wed May 06, 2015 2:09 am

After update Router OS 6.28 RB1100AHx2 not send e-mail backup!

Error: timeout occured

paoloaga · Wed May 06, 2015 3:14 am

I am experiencing troubles with the MTU on PPPoE session through a VLAN on a RB2011. The automatic mangle rule that changes MSS to 1440 seems not to work correctly. If I manually add a mangle rule that sets MSS to 1400 everything works again. The router was working correctly with v6.22. I also tried 6.29rc14 and the behaviour is the same as v6.28.

I hope to find the time tomorrow to open a ticket, generate the supout and try to find the exact steps to reproduce this bug.

sinaexer · Wed May 06, 2015 8:28 am

Hi
Unfortunately, the new version (6.28) web proxy does not work
Please check

kivimart · Wed May 06, 2015 9:23 am

Hi
Unfortunately, the new version (6.28) web proxy does not work
Please check

i have 2 diffrrent RB that runs web-proxy on 6.28, no problem

bclewl1ns · Wed May 06, 2015 10:10 am

1. Latency - Even using auto on the NV2 Setting latency is way to high.
2. Jitter - It is increasely getting worse on NV2 Links
3. TCP Throughput Should be more stable and better then 10% of UDP

I am aware of NV2's benefits and its shortcomings. My only POINT is that instead of doing all these new features LETS Stabilize a issue that is effecting current users.
+1

So Interesting info to share. We replaced some of our backhaul links with Mimosa. Previously we had PTP RB921 Net Metals with NV2 because well it is the only one stable. We were getting 275 Meg up and Down. (Not to shabby) on the NetMeals. The Mimosa is doing 720 up and Down. Also Latency is only 2 MS verses 8 to 15 on the NV2 Link. All that is great However the most interesting item to note is The bandwidth on all of our NV2 PTMP sectors Went back to normal. Customers called us to inform us that the bandwidth was better then it was before. Testing we were getting 90 Meg UDP and 85 TCP on 911ac CPE's connected to a 921 Netmetal Ac sector. I was amazed out how much the CPE performance was increased when we moved out the NV2 PTP's in the system.

We also tested this with a older RB911 N link to a secondary site with ROS 5.26 using NSTREAM - Same Result as with the mimosa. Increase in overall performance and CPE speed.

In our network we are going to move all NV2 PTP's our and replace them with other systems. Mikrotik Please make it a point in the future to fix NV2 in PTP environments

nmaton · Wed May 06, 2015 1:23 pm

For me this is one of the buggiest releases yet.

My router hangs and stops forwarding packets.
I can still connect via serial but when you issue the system reboot command .. it hangs and spews out the following (but does not reboot)

] > system reboot
Reboot, yes? [y/N]:
y
system will reboot shortly

Rebooting...
Stopping services...
(0,3) WARNING: waiting for reply message: timed out
(0,3) Now cycle 562417970394666, timeout 40000000000 cycles
(0,3) Caller 0xfffffffc00042c58 with ICS 0
(0,3) Message type 0x10f, dest 0x40
(0,3) chan 1: 271
(0,3) chan 2: 271
(0,3) chan 4: 271
(0,3) chan 5: 271
(0,3) chan 17: 271
(0,3) chan 18: 271
(0,3) chan 19: 271
(0,3) chan 20: 271
(0,3) chan 21: 271
(0,3) chan 33: 271
(0,3) chan 34: 271
(0,3) chan 35: 271
(0,3) chan 36: 271
(0,3) chan 37: 271
(0,3) chan 49: 271

This on my 2 CCR1036 routers which are configured solely for PPTP termanation purposses.

Wed May 06, 2015 2:47 pm

nmaton,

please contact MikroTik support (support@mikrotik.com) with attached support output files from one of your routers.
We will try to find out what is the reason for your issues. We are sorry for any inconvenience.

rextended · Wed May 06, 2015 3:59 pm

Hi
Unfortunately, the new version (6.28) web proxy does not work
Please check

1) Clear Cache
2) Change the webproxy folder
3) clear again cache

retry...

formatc · Wed May 06, 2015 4:50 pm

L2TP/ipsec firewall config:

/ip firewall mangle

chain=input action=mark-packet new-packet-mark=ipsec passthrough=yes
protocol=udp dst-port=500,4500 log=no log-prefix=""

/ip firewall filter
3 chain=input action=accept protocol=udp dst-port=1701 packet-mark=ipsec
log=yes log-prefix=""
4 chain=input action=reject reject-with=icmp-network-unreachable
protocol=udp dst-port=1701 packet-mark=no-mark log=no log-prefix=""

L2TP/IPsec packet does not match rule 3, but match rule 4.
It seems there is no marker after the decryption IPSEC packet.

downgrade to 6.27 solve this problem.

KillerOPS · Wed May 06, 2015 11:35 pm

bclewl1ns, i think you should also consider the increased self-interference in your setup.
I've experienced this myself. You keep adding links and customers on the towers, and you can't expect the overall performance not to degrade. Also, your competitors don't stand still. They also keep adding links and customers.

4 years ago 5Ghz was clean. Now, I think in some places 2.4Ghz is cleaner than 5.
Also, it's a pretty known fact that nstreme on n p2p links is much better compared to NV2. Nv2 has that TDMA thingie which is very useful in P2MP, but almost useless in p2p.

bclewl1ns · Thu May 07, 2015 12:13 am

I am well aware of the increased self-interference . We plan sites very carefully to avoid this as much as possible. I also agree Nstream is allot better for PTP however it has proven to be almost useless when it comes to the new AC radios. nstream disconnects frequently and throughput is about half that of NV2. Understand this is not a criticism but a info share that is showing that there needs to be some work on the PTP side of NV2 or better fixing nstream to work with ac radios. More info on self-interference, The Mimosa radios we have replaced our backhaul with are 5 Ghz radio 4X4 so it uses 160 Mhz of spectrum. If anything we polluted our environment more by adding them in. However Speeds increased significantly. They are also TDMA based radios. But they have GPS Sync.

paolopoz · Thu May 07, 2015 12:27 pm

Does anybody experienced problems with wireless disconnections after upgrading to 6.28?
I have a PtP link and after upgrading one of the two sides I keep getting disconnections with error:

lost connection received deauth class 2 frame received (6)

After downgrading to 6.27 the problem disappeared.
The board is RB711-5Hn-MMCX
The wireless card is AR5008 Wireless Network Adapter (rev: 1)

PastuhMedvedey · Thu May 07, 2015 3:14 pm

Router - CCR-1036 8g-2s+
RouterOS v6.28

Slow download speeds BGP prefixes, and uneven loading the CPU.
Wrong time BGP session timer.

paolopoz · Thu May 07, 2015 3:38 pm

Slow download speeds BGP prefixes, and uneven loading the CPU.

This is a known limitation: BGP routing table loading is single threaded thus using only one CPU.

PastuhMedvedey · Thu May 07, 2015 4:18 pm

This is a known limitation: BGP routing table loading is single threaded thus using only one CPU.

skibi82 · Sun May 10, 2015 11:39 am

When the gentlemen of mikrotik deign to reply:

Ticket#2015042166000534 Data: 2015-04-21 16:08
Post
2015-04-21 15:57
Critical bug in 6.28
After updating to version 6.28 is the problem of detection ipsec esp.

Before that could properly be described in Mangle encrypted traffic Packet

Markt: IPSEC and then on filter rule i culd filter IP inside the ipsec
tunnel

Since version 6.28 does not have the capability of this.
Because the above error stanol traffic on one of the VPN concentrator.
Please fix it as fast as possible

1 Rule on mangle

/ip firewall mangle
add action=mark-packet chain=input comment=IPSEC new-packet-mark=\
ipsec-encrypted passthrough=no protocol=ipsec-esp

2 Rule on forward
/ip firewall filter
add action=passthrough chain=forward comment=IPSEC
packet-mark=ipsec-encrypted

L2TP/ipsec firewall config:

/ip firewall mangle

chain=input action=mark-packet new-packet-mark=ipsec passthrough=yes
protocol=udp dst-port=500,4500 log=no log-prefix=""

/ip firewall filter
3 chain=input action=accept protocol=udp dst-port=1701 packet-mark=ipsec
log=yes log-prefix=""
4 chain=input action=reject reject-with=icmp-network-unreachable
protocol=udp dst-port=1701 packet-mark=no-mark log=no log-prefix=""

L2TP/IPsec packet does not match rule 3, but match rule 4.
It seems there is no marker after the decryption IPSEC packet.

downgrade to 6.27 solve this problem.

rextended · Sun May 10, 2015 5:34 pm

When the gentlemen of mikrotik deign to reply:

Ticket#2015042166000534 Data: 2015-04-21 16:08
Post
2015-04-21 15:57
Critical bug in 6.28
After updating to version 6.28 is the problem of detection ipsec esp.

Before that could properly be described in Mangle encrypted traffic Packet

Markt: IPSEC and then on filter rule i culd filter IP inside the ipsec
tunnel

Since version 6.28 does not have the capability of this.
Because the above error stanol traffic on one of the VPN concentrator.
Please fix it as fast as possible

1 Rule on mangle

/ip firewall mangle
add action=mark-packet chain=input comment=IPSEC new-packet-mark=\
ipsec-encrypted passthrough=no protocol=ipsec-esp

2 Rule on forward
/ip firewall filter
add action=passthrough chain=forward comment=IPSEC
packet-mark=ipsec-encrypted

You mark the packet on input chain and apply the filter on forward chain.

Obviously you can not see any input marked packet on forward chain.
Use on both rules the input chain or on both rules the forwad chain.

skibi82 · Mon May 11, 2015 4:55 pm

1. According to the scheme packet floow, described ipsec traffic on input should be visible on a forward with mark.
Everything works fine until 6.27
After the update above 6.28 there is now the possibility to describe traffic that it's ipsec with the ipsec
A change in the documentation does not make this any indication of change Packet floow

Please you to firstly understand what is written:
http://wiki.mikrotik.com/wiki/Manual:Pa ... decryption

700px-Packet_Flow_Example_5c.png

rextended - And by the way, you can run as dekryptor ESP if you want to input you lead srcip

When the gentlemen of mikrotik deign to reply:

Ticket#2015042166000534 Data: 2015-04-21 16:08
Post
2015-04-21 15:57
Critical bug in 6.28
After updating to version 6.28 is the problem of detection ipsec esp.

Before that could properly be described in Mangle encrypted traffic Packet

Markt: IPSEC and then on filter rule i culd filter IP inside the ipsec
tunnel

Since version 6.28 does not have the capability of this.
Because the above error stanol traffic on one of the VPN concentrator.
Please fix it as fast as possible

1 Rule on mangle

/ip firewall mangle
add action=mark-packet chain=input comment=IPSEC new-packet-mark=\
ipsec-encrypted passthrough=no protocol=ipsec-esp

2 Rule on forward
/ip firewall filter
add action=passthrough chain=forward comment=IPSEC
packet-mark=ipsec-encrypted

You mark the packet on input chain and apply the filter on forward chain.

Obviously you can not see any input marked packet on forward chain.
Use on both rules the input chain or on both rules the forwad chain.

Mon May 11, 2015 6:02 pm

@skibi82: ipsec is in transport mode or tunnel mode?

skibi82 · Mon May 11, 2015 8:37 pm

In Mail from 2015-04-21 15:57:23.04.2015 - Ticket#2015042166000534 was picture incluted:

wtf.png

YES is in tunnel mode

@skibi82: ipsec is in transport mode or tunnel mode?

Sykotic · Tue May 12, 2015 12:34 pm

Hi Mikrotik support.

6.28 broke BGP signaled VLPS.
I don't have any logs as I don't have the spare equipment to do it outside of my production environment yet, however it's easy enough to reproduce. Looks like there is at least one other that can confirm the problem.

http://forum.mikrotik.com/viewtopic.php?f=14&t=96705

pingus · Tue May 12, 2015 9:46 pm

I have the same problem with rb600a:

jan/02/1970 00:41:33 system,error,critical System rebooted because of kernel failure
jan/02/1970 00:41:33 system,error,critical Out of memory condition was detected
jan/02/1970 00:41:33 system,error,critical router was rebooted without proper shutdown

And I am not able to downgrade, after the reboot I have still 6.28 installed...

pingus · Tue May 12, 2015 10:08 pm

Ok, got it. The npk file was in a subfolder ...

rgunderson · Wed May 13, 2015 6:16 am

6.28 disabled the SFP on my CCR1009-8G-1S. Downgrading back to 6.27 let it work again.

dmi3 · Wed May 13, 2015 2:30 pm

Hi!
strods told me to write my problem with 6.29RC here, and i must update it from 6.28 cuz "sector writes bug" so:
1. i have a question. Where can i find additional packets like ntp, ups, capsman2 for RC platform?
2. now i use wireless-fp to use capsman, and i got 2 routers in configuration, and i got 2 bridges in master router
i got datapath1 configurated to add all master-wifi's in first bridge_work
and cap on master router did it, but no packets pass through router to, even when cap is in bridge_work, inet working but no local connect
on my second datapath2 configuration the slave-wifi's added to bridge_free
i make simple rules to allow bridge_free clients access to 2 of bridge_work adresses and its worked, like it should

when i use capsman2 on 6.28 i didnt have such trouble, all my master-wifi's client with datapath1 can connect each others and every bridge_work (lan) clients

 caps-man datapath print 
 0 name="datapath1" client-to-client-forwarding=yes bridge=bridge_work 

 1 name="datapath2" client-to-client-forwarding=yes bridge=bridge_free

 interface bridge port print
Flags: X - disabled, I - inactive, D - dynamic 
 #    INTERFACE              BRIDGE              PRIORITY  PATH-COST    HORIZON
 0    ether8                 bridge_work             0x80         10       none
 1    ;;; master1
      ether1                 bridge_work             0x80         10       none
 2    ;;; master2
        ether7                 bridge_work             0x80         10       none
 3 I  wlan_work              bridge_work             0x80         10       none
 4 I  ether6                 bridge_work             0x80         10       none
 5    ether9                 bridge_work             0x80         10       none
 6    ether10                bridge_work             0x80         10       none
 7 I  wlan_guest             bridge_free             0x80         10       none
 8 ID cap7                   bridge_work             0x80         10       none
 9  D cap8                   bridge_free             0x80         10       none
10 ID cap9                   bridge_work             0x80         10       none
11  D cap10                  bridge_free             0x80         10       none

[/size]

karlisi · Wed May 13, 2015 3:22 pm

2. now i use wireless-fp to use capsman, and i got 2 routers in configuration, and i got 2 bridges in master router
i got datapath1 configurated to add all master-wifi's in first bridge_work
and cap on master router did it, but no packets pass through router to, even when cap is in bridge_work, inet working but no local connect
on my second datapath2 configuration the slave-wifi's added to bridge_free
i make simple rules to allow bridge_free clients access to 2 of bridge_work adresses and its worked, like it should

when i use capsman2 on 6.28 i didnt have such trouble, all my master-wifi's client with datapath1 can connect each others and every bridge_work (lan) clients

CAPsMAN v1 works fine if you have separate device for CAPsMAN and separate wireless devices, but if you are using for CAPsMAN router with wireless, it won't allow packets between wireless interface and other interfaces on this router. This bug apparently is corrected in v2. There is no exact mention about it in changelog, only this: ' improved CAP<->CAPsMAN data connection protocol'

dmi3 · Wed May 13, 2015 5:00 pm

CAPsMAN v1 works fine if you have separate device for CAPsMAN and separate wireless devices, but if you are using for CAPsMAN router with wireless, it won't allow packets between wireless interface and other interfaces on this router. This bug apparently is corrected in v2. There is no exact mention about it in changelog, only this: ' improved CAP<->CAPsMAN data connection protocol'

Thx fo reply!
I would gladly use capsmanv2, on my running 29rc18, but i didnt know where can i pick a npk file for RC

and one more time in 6.28 with v2 all works great

i got vicious circle
6.27 cant fasttrack
6.28 have "sector writes"
6.29rc doesnt have capsmanv2

what should i do? back to 6.27?

grusu · Wed May 13, 2015 5:21 pm

CAPsMAN v1 works fine if you have separate device for CAPsMAN and separate wireless devices, but if you are using for CAPsMAN router with wireless, it won't allow packets between wireless interface and other interfaces on this router. This bug apparently is corrected in v2. There is no exact mention about it in changelog, only this: ' improved CAP<->CAPsMAN data connection protocol'
Thx fo reply!
I would gladly use capsmanv2, on my running 29rc18, but i didnt know where can i pick a npk file for RC
and one more time in 6.28 with v2 all works great

i got vicious circle
6.27 cant fasttrack
6.28 have "sector writes"
6.29rc doesnt have capsmanv2

what should i do? back to 6.27?

Link for capsmanv2:

http://www.mikrotik.com/download/share/ ... mipsbe.npk

Change "mispbe" with you router architecture.

dmi3 · Wed May 13, 2015 5:53 pm

Link for capsmanv2:

http://www.mikrotik.com/download/share/ ... mipsbe.npk

Change "mispbe" with you router architecture.

THX alot!
there are you get this? any archive or beta page with list of stuffs, or you simply write it with own logic?

pukkita · Wed May 13, 2015 6:14 pm

I am well aware of the increased self-interference . We plan sites very carefully to avoid this as much as possible. I also agree Nstream is allot better for PTP however it has proven to be almost useless when it comes to the new AC radios.

Have yet to test, does this apply also if set in N mode?

grusu · Wed May 13, 2015 9:13 pm

Link for capsmanv2:

http://www.mikrotik.com/download/share/ ... mipsbe.npk

Change "mispbe" with you router architecture.
THX alot!
there are you get this? any archive or beta page with list of stuffs, or you simply write it with own logic?

The information I got from another forum post.

honzam · Thu May 14, 2015 4:08 pm

Issue with sector writes rising will be fixed in version 6.29. If you want to test it, then fix for it will be included in RouterOS starting from 6.29rc13 version when it comes out.

Which boards have this problem? Only mipsbe?
It is real problem or only "cosmetics" ?
Thanks

Thu May 14, 2015 4:19 pm

It was problem for all devices. Simply all the time something was written in memory. It did not cause any problems. Only additional process was running while Firewall menu was opened.

bclewl1ns · Fri May 15, 2015 3:25 am

I am well aware of the increased self-interference . We plan sites very carefully to avoid this as much as possible. I also agree Nstream is allot better for PTP however it has proven to be almost useless when it comes to the new AC radios.
Have yet to test, does this apply also if set in N mode?

I have set to only N mode however if the radio is and AC radio the problem persists.

dmi3 · Fri May 15, 2015 10:00 am

Hi guys! Updating to 6.29rc20 an have some questions:
1. i found capsmanv2 on direct link http://www.mikrotik.com/download/share/ ... mipsbe.npk , but where can i find NTP packet for RC releases? Maybe someone have user-friendly web page where we can download all rc packets we need? it would be nice
2. whether fasttrask added to the list of IPV6-FIREWALL?
3. i added and configured 6to4 interface through 192.88.99.1 but cant get good score on tests because MT blocks ICMPv6 to local bridge, i added some rules based on https://gist.github.com/tiernano/4344701
computer pinged good from MT and MT from Internet

4. MT cant ping ipv6 adresses with "invalid value" error, all clients can (look terminal @ screenshot above)
5. When i try SYSTEM-AUTO APGRADE while updating from RC18 to RC20 i got "there is now newer alpha/beta/rc version" but i know they exist

thx for help and any information!

bajodel · Fri May 15, 2015 12:33 pm

Problems on bridges interfaces, input traffic is not counted !!

Tested on platforms: CRS125-24G-1S and RB2011
RouterOS affected versions: 6.29rc14 6.29rc18 and 6.29rc20 (last one screenshot attached ..scroll down to view all)

Sob · Fri May 15, 2015 10:00 pm

3. i added and configured 6to4 interface through 192.88.99.1 but cant get good score on tests because MT blocks ICMPv6 to local bridge, ...

If you have 6to4 with 2002::/16 address derived from IPv4, unspecified remote address for 6to4 interface and 192.88.99.1 as default gateway, you're running into "my" 6to4 bug (broken since 6.20, in current state since 6.24). The problem is that default gateway specified as ::192.88.99.1%6to4 does not work at all. So you can only access 2002::/16 and that's it. And if someone tries to connect to you from production (non-2002::/16) range, packets get to your router, but replies don't get out.

Current workaround recommended by MikroTik support is to use two 6to4 interfaces:

/interface 6to4
add local-address=<your IPv4 address> mtu=1480 name=6to4
add local-address=<your IPv4 address> mtu=1480 name=6to4gw remote-address=192.88.99.1
/ipv6 address
add address=2002:<your IPv4 address>::1/16 advertise=no interface=6to4
/ipv6 route
add distance=1 gateway=6to4gw

dmi3 · Mon May 18, 2015 9:04 am

Current workaround recommended by MikroTik support is to use two 6to4 interfaces:

I did as suggested, and now http://www.ipv6-test.com/ tells me there is no ipv6

some other services tell there is.. in fact clients got addresses, can go ipv6, can ping, download etc
but ping still didn't reach clients after MT
suppose its wright?:

Sob · Mon May 18, 2015 3:15 pm

I did as suggested, and now http://www.ipv6-test.com/ tells me there is no ipv6

Ok, that does not sound like improvement. I'd suggest to create new topic and post more details, config export, etc..

jondavy · Mon May 18, 2015 3:32 pm

the package all_packages-ppc-6.28.zip this damaged after installation and check
and also another bug:
he reappears the WebBox interface (which does not work) in RB1100AH

TomosRider · Tue May 19, 2015 11:31 am

One thing that happened yesterday, but i saw earlier posts about this....after power loss, Mtik lost all static DHCP leases....running ROS 6.27, will do update to .28.....any ideas why is this happening? No bad blocks....

TomosRider · Tue May 19, 2015 5:22 pm

After upgrade, same thing...also, cant create backup anymore. When i try to create backup file, i get message:critical error creating backup file:could not read all configuration files.
RB is 1200, running smoothly for months. Mtik staff, have any idea how to solve this problem?

rextended · Thu May 21, 2015 8:36 pm

Today are one month after I install 6.28 on all my production devices.
No one single problem on 5GHz ac or n, BGP, VRRP, EoIP, PPtP, PPPoE, user-manager, DHCP server, DNS server & big lists of static entry, firewall & big address-list, Routing, etc., etc., etc.

(I do not use VLAN, IPv6, MPLS or OSFP)

Thanks MikroTik, finally one stable MRRU+EoIP version of the RouterOS

Thu May 21, 2015 10:44 pm

... and the excessive flash writings?

rextended · Fri May 22, 2015 11:19 am

... and the excessive flash writings?

I do not have this problem, on production I do not have any mipsbe devices.

The only mipsbe devices are AP or PTP and no one of these have any firewall rule.

Cha0s · Fri May 22, 2015 2:21 pm

So just because -you- are not affected by the bug, does not mean that 6.28 was all colors and rainbows.

Fri May 22, 2015 2:42 pm

This is it. I decided to skip 6.28. Hope the 6.29 will be at least without major bugs. I know that there will not be corrections I am waiting for, so maybe I will wait till 6.30 before I start my own update tests. Keeping 6.27 in production so far.

rextended · Sat May 23, 2015 10:00 am

So just because -you- are not affected by the bug, does not mean that 6.28 was all colors and rainbows.

And this provocative phrase that means?
And where I have writed "6.28 is all colors and rainbows"?
If you do not like positive recensions, simply do not read, or comment, those.

If you have time and patience to ignore english errors, read that:
http://forum.mikrotik.com/viewtopic.php ... 00#p479732
I think that explain some bugs...

james · Sat May 23, 2015 11:23 am

Since upgrading my home box seems very unstable, lots of spontaneous reboots and enough bad connections that I have power cycled it a few times.

[admin@catbus_home] /system> /system routerboard print
routerboard: yes
model: 751G-2HnD
serial-number: 2F8701252CD1
current-firmware: 3.17
upgrade-firmware: 3.22
[admin@catbus_home] /system> /system resource print
uptime: 58m57s
version: 6.28
build-time: Apr/15/2015 15:18:31
free-memory: 26.1MiB
total-memory: 64.0MiB
cpu: MIPS 24Kc V7.4
cpu-count: 1
cpu-frequency: 250MHz
cpu-load: 4%
free-hdd-space: 21.4MiB
total-hdd-space: 64.0MiB
write-sect-since-reboot: 3132
write-sect-total: 12762707
bad-blocks: 0%
architecture-name: mipsbe
board-name: RB751G-2HnD
platform: MikroTik
[admin@catbus_home] /system>

From the log

may/18 23:14:53 system,error,critical System rebooted because of kernel failure
may/18 23:14:54 system,error,critical Out of memory condition was detected
may/18 23:14:54 system,error,critical router was rebooted without proper shutdown

Cha0s · Sat May 23, 2015 2:59 pm

So just because -you- are not affected by the bug, does not mean that 6.28 was all colors and rainbows.
And this provocative phrase that means?

It means that Mikrotik 6.28 (or any other 6.x version) is FAR from stable.

I may have been a bit offensive to you but that was not my intention. I just used your comment to point out that Mikrotik is far from stable in a sarcastic way (hence the colors and rainbows).

rextended · Sat May 23, 2015 3:09 pm

I may have been a bit offensive to you but that was not my intention.

rextended · Sat May 23, 2015 3:43 pm

Since upgrading my home box seems very unstable, lots of spontaneous reboots and enough bad connections that I have power cycled it a few times.

[admin@catbus_home] /system> /system routerboard print
routerboard: yes
model: 751G-2HnD
serial-number: 2F8701252CD1
current-firmware: 3.17
upgrade-firmware: 3.22
[admin@catbus_home] /system> /system resource print
uptime: 58m57s
version: 6.28
build-time: Apr/15/2015 15:18:31
free-memory: 26.1MiB
total-memory: 64.0MiB
cpu: MIPS 24Kc V7.4
cpu-count: 1
cpu-frequency: 250MHz
cpu-load: 4%
free-hdd-space: 21.4MiB
total-hdd-space: 64.0MiB
write-sect-since-reboot: 3132
write-sect-total: 12762707
bad-blocks: 0%
architecture-name: mipsbe
board-name: RB751G-2HnD
platform: MikroTik
[admin@catbus_home] /system>

From the log

may/18 23:14:53 system,error,critical System rebooted because of kernel failure
may/18 23:14:54 system,error,critical Out of memory condition was detected
may/18 23:14:54 system,error,critical router was rebooted without proper shutdown

1) Update the bios (/sys rou up)
reboot
2) Download winbox from the gateway ip of routerboard
3) Export the config (/export file=config)
4) Save the config from routerboard to your pc
5) Reset the config of your routerboard (/system reset-configuration keep-users=no no-defaults=yes)
reboot
6) Do this command: /romon port remove [find]
7) open config.rsc with notepad and paste block by block on new terminal inside winbox untill you done.

AlexVB · Mon May 25, 2015 8:11 am

After upgrade from 6.7 to 6.28 we now get spontaneous hangs every few days on CCR1036-12G-4S.

rextended · Mon May 25, 2015 8:32 pm

After upgrade from 6.7 to 6.28 we now get spontaneous hangs every few days on CCR1036-12G-4S.

What method you use for install 6.7 (or the previous and what method use to go to 6.7)?

What method you use to upgrade to 6.28?

BIOS version of the router?

What package you have installed and what date and time have the packages?

AlexVB · Tue May 26, 2015 9:05 am

> What method you use for install 6.7 (or the previous and what method use to go to 6.7)?
Upgrade from prevoius, preinstalled version

> What method you use to upgrade to 6.28?
Upgrade package

> BIOS version of the router?

routerboard: yes
model: CCR1036-12G-4S
serial-number: ...
current-firmware: 3.10
upgrade-firmware: 3.22

What package you have installed and what date and time have the packages?
See attached image.

AlexVB · Tue May 26, 2015 9:06 am

I not mentioned that integrated watchdog does not detects the hangs, but If it is set to ping address then it detects hang but does not reboots router.

SCare · Tue May 26, 2015 1:06 pm

Problems with CAPsMAN with the hAP lite as APs, I've differents APs with the same configuration working together (mAP and RB751) and working ok with 6.28, hAP lite not working

Thu May 28, 2015 11:22 am

bajodel - Unfortunately on bridge interface you can only see traffic where bridge itself is in/out interface. Bridged traffic will not be visible on this interface.

elurex · Thu Jun 25, 2015 5:52 am

can wireless bridge and virtual ap working on sxt SA5 ac?

hrehm · Mon Jul 06, 2015 7:19 pm

nmaton,

please contact MikroTik support (support@mikrotik.com) with attached support output files from one of your routers.
We will try to find out what is the reason for your issues. We are sorry for any inconvenience.

I do have EXACTLY the same issue with v6.28.
CCR1036 running a Hotspot, Etherrnet "crashes" after a few days, while ROS is still accessible on seial.

trying to disable Ether1 (the only connected port) brings:

[admin@Wifi-Master] /interface> disable ether1
(5,0) WARNING: waiting for reply message: timed out
(5,0) Now cycle 220085031284627, timeout 40000000000 cycles
(5,0) Caller 0xfffffffc00042c58 with ICS 0
(5,0) Message type 0x10f, dest 0x3
(5,0) chan 4: 271
(5,0) chan 5: 271
(5,0) chan 19: 271
(5,0) chan 20: 271
(5,0) chan 21: 271
(5,0) chan 35: 271
(5,0) chan 36: 271
(5,0) chan 37: 271
(5,0) chan 51: 271
(5,0) chan 52: 271
(5,0) chan 53: 271
(5,0) chan 67: 271
(5,0) chan 68: 271
(5,0) chan 69: 271
(5,0) chan 83: 271
(5,0) chan 84: 271
(5,0) chan 85: 271
(3,1) WARNING: waiting for reply message: timed out
(3,1) Now cycle 220371650555530, timeout 40000000000 cycles
(3,1) Caller 0xfffffffc00037860 with ICS 0
(3,1) Message type 0x405, dest 0x3
(4,4) WARNING: waiting for reply message: timed out
(4,4) Now cycle 220440770250348, timeout 40000000000 cycles
(4,4) Caller 0xfffffffc00037860 with ICS 0
(4,4) Message type 0x405, dest 0x3

CherryNOD · Thu Jul 30, 2015 8:38 am

Tools -> RoMON, then check ENABLE, finally click OK

Who is online