RouterOS v6.6 released

Fri Nov 08, 2013 9:27 am

What's new in 6.6 (2013-Nov-07 13:04):

*) winbox - fixed problem where all previous session opened windows were read only;
*) certificate - no more 'reset-certificate-cache' and 'decrypt' commands,
private keys can be decrypted only on 'import', use 'decrypt' before upgrade if needed;
*) fixed arp-reply only with more than one ip address on interface;
*) fixed RB400 not to reboot by watchdog during micro-sd format;
*) web proxy - fix SPDY server push handling;
*) certificate - merged '/certificate ca issued', '/certificate scep client' and
'/certificate templates' into '/certificate';
*) console - :foreach command can iterate over keys and values in an array,
by specifying two counter variables, e.g.:
:foreach k,v in=[/system clock get] do={:put "$k is $v"};
*) added support for new Intel 10Gb ethernet cards (82599);
*) certificates - fixed certificate import;
*) wireless - fixed crash when dfs was enabled on pre-n wireless cards;
*) fixed port flapping on CCR;

Simply push "Check for Updates" in your router,
or download from http://www.mikrotik.com/download

Chupaka · Fri Nov 08, 2013 10:08 am

http://tracker.mikrotik.com: Torrent not registered

BobcatGuy · Fri Nov 08, 2013 10:11 am

I have a RB1100 AHX2 that has a port flapping issue on ETHER1 but not on ETHER4, I used ether 4 since it was the only available port.

Did the Port flapping issue get fixed for this board as well? Are they using the same chipset as the CCR?

The issue was with not being able to force 100 Mbit full duplex, and will only work on auto negotiate, and connect at 1 Gigabit.

Would like to know before I decide to upgrade, I am running 6.5 on there now.

honzam · Fri Nov 08, 2013 10:13 am

Where is the mentioned this change? http://forum.mikrotik.com/viewtopic.php ... 75#p394593
I don´t see it in changelog 6.6

ivanich · Fri Nov 08, 2013 10:38 am

MT are you kidding? Firmware bug on rb2011 is still here, what's wrong with you?

Fri Nov 08, 2013 10:38 am

Where is the mentioned this change? http://forum.mikrotik.com/viewtopic.php ... 75#p394593
I don´t see it in changelog 6.6

This is a fix for a problem, that was introduced in v6.6rc1, so it is not mentioned, since regular users upgrade from v6.5

Fri Nov 08, 2013 10:38 am

MT are you kidding? Firmware bug on rb2011 is still here, what's wrong with you?

Please clarify what firmware bug are you talking about ?

ivanich · Fri Nov 08, 2013 10:43 am

MT are you kidding? Firmware bug on rb2011 is still here, what's wrong with you?
Please clarify what firmware bug are you talking about ?

NAND corruption with firmware 3.10, and as a result it's allmost not possible to upload anything via ftp or webUI, so now i should find a windows computer with buggy netinstall and spent time to downgrade that piece of sh** to Ros6.4 and than to 3.0.8 firmware

ySgAIsvNBI · Fri Nov 08, 2013 11:26 am

I'm testing CCR-1036-12G-4S with latest ROSv6 but it cannot route more than 600-800Mbps with one simple queue entry (no firewall or mangle). This is very disappointing. Is this a known issue or maximum HW limit of CCR?

Setup:
Two bonding interfaces (802.3ad, layer 2 n 3 hash)
bonding1 - eth7, eth8
bonding2 - eth9, eth10

Traffic volume: 1.2Gbps down / 200Mbps up
Firewall/Mangle: None
Route: static routes less than 10

When I add a simple queue entry with max-limit 2000Mbps down with pfifo 50 packets, traffic drops to 600-700Mbps instanly from 1.2Gbps with dropped statistics incrementing several thousands per second. Theoretically it shouldn't drop any packet as the traffic is well below the limit.

I have tried queue tree but result was same. As soon as a default pfifo 50pkts queue with max limit 2000M was added, traffic dropped to 600Mbps. Various pfifo queue sizes from 50 to 100000 didn't change the result. I have tried applying queue tree entry to bonding interface & global and increased max limit to 4000M but result was still the same. Initially I thought this was due to load distribution between cores but none of CPU reached over 20%. Above setup works perfectly on x86 with ROSv5.

Anyone else experiencing same issue?

Fri Nov 08, 2013 11:55 am

I'm testing CCR-1036-12G-4S with latest ROSv6 but it cannot route more than 600-800Mbps with one simple queue entry (no firewall or mangle). This is very disappointing. Is this a known issue or maximum HW limit of CCR?

Setup:
Two bonding interfaces (802.3ad, layer 2 n 3 hash)
bonding1 - eth7, eth8
bonding2 - eth9, eth10

Traffic volume: 1.2Gbps down / 200Mbps up
Firewall/Mangle: None
Route: static routes less than 10

When I add a simple queue entry with max-limit 2000Mbps down with pfifo 50 packets, traffic drops to 600-700Mbps instanly from 1.2Gbps with dropped statistics incrementing several thousands per second. Theoretically it shouldn't drop any packet as the traffic is well below the limit.

I have tried queue tree but result was same. As soon as a default pfifo 50pkts queue with max limit 2000M was added, traffic dropped to 600Mbps. Various pfifo queue sizes from 50 to 100000 didn't change the result. I have tried applying queue tree entry to bonding interface & global and increased max limit to 4000M but result was still the same. Initially I thought this was due to load distribution between cores but none of CPU reached over 20%. Above setup works perfectly on x86 with ROSv5.

Anyone else experiencing same issue?

you have 36 cores and you wanna put everything in single simple queue -- really?

it was noted in RouterOS v6 presentation at the MUM Russia (and US if i'm not mistaken), that on CCR simple queues gets max performance if there are at least 32 simple queues

massive multi-core devices require different approach

ySgAIsvNBI · Fri Nov 08, 2013 1:15 pm

you have 36 cores and you wanna put everything in single simple queue -- really?

it was noted in RouterOS v6 presentation at the MUM Russia (and US if i'm not mistaken), that on CCR simple queues gets max performance if there are at least 32 simple queues

massive multi-core devices require different approach

You mean 32 parent queues or 32 child queues? My test had one parent and about 10 child queues but result was same.

Fri Nov 08, 2013 1:43 pm

you have 36 cores and you wanna put everything in single simple queue -- really?

it was noted in RouterOS v6 presentation at the MUM Russia (and US if i'm not mistaken), that on CCR simple queues gets max performance if there are at least 32 simple queues

massive multi-core devices require different approach
You mean 32 parent queues or 32 child queues? My test had one parent and about 10 child queues but result was same.

32 main level queues, one parent queue still gives single point of congestion

Chupaka · Fri Nov 08, 2013 2:45 pm

you have 36 cores and you wanna put everything in single simple queue -- really?

it was noted in RouterOS v6 presentation at the MUM Russia (and US if i'm not mistaken), that on CCR simple queues gets max performance if there are at least 32 simple queues

massive multi-core devices require different approach

so, currently there's no way to limit total amount of huge traffic on MikroTik devices?..

hapi · Fri Nov 08, 2013 3:00 pm

No, tested a few months ago. One queue dies around 70.000pps. Xeon to handle.

djmanu · Fri Nov 08, 2013 4:35 pm

i made the upgrade to stable 6.6 with my CCR1036-12G-4S crashed my network...!

www---CCR1---CCR2----Clients
CCR1=6.5
CCR2=6.6

on CCR1 /ip/arp
00:00:00:00:00 for CCR2

it is a BUG!!!
i have downgrade the CCR2 to 6.5 and all is fine now...
i use teamed lans and VLAN on CCRs

ySgAIsvNBI · Fri Nov 08, 2013 4:43 pm

you have 36 cores and you wanna put everything in single simple queue -- really?

it was noted in RouterOS v6 presentation at the MUM Russia (and US if i'm not mistaken), that on CCR simple queues gets max performance if there are at least 32 simple queues

massive multi-core devices require different approach
You mean 32 parent queues or 32 child queues? My test had one parent and about 10 child queues but result was same.
32 main level queues, one parent queue still gives single point of congestion

much appreciated!
but 32 main level queues?? hmm.. is this real? where are the benefits of HTB/QT?

makrovirus · Fri Nov 08, 2013 6:57 pm

the version 6.6 not working vlan on the bridge . if disable enable starts to work. After rebooting everything is repeated

patrickmkt · Fri Nov 08, 2013 6:58 pm

What's new in 6.6 (2013-Nov-07 13:04):

*) certificates - fixed certificate import;

Good I can import fine now, however, I can't change the internal name of the certificate anymore:
'couldn't change certificate xxx - certificate subject is read only'

Lakis · Fri Nov 08, 2013 8:25 pm

When AR93xx 30Mhz channel width support ?

skibi82 · Fri Nov 08, 2013 10:37 pm

CRS125 after updating the E353 modem dangled on one of devices. After upgrade on second device was missing in action E3131 modem .
Disconnect the modem from the device. Entry into System-> Resources hmm I see Ghost USB modem on WTF visible.

Factory reset. Waiting 1 min. Modem connection.
WFT No serial port.

Plug and unplug unplug plug plug plug unplug unplug restart
WTF no port??
Ten-minute break per quarter vodka.
To not fuck device through the window. plug unplug wait plug reboot. Bingo appeared ports.

Ehman · Fri Nov 08, 2013 10:51 pm

To not f**k device through the window

LOL!!

..thats like me and my RB1100AHx2 with ANY new version on, I'm running a netinstalled v6.0rc13 ...and I'm a very happy man now

..no more upgrades for me, thank you

ranpha · Sat Nov 09, 2013 5:43 am

RouterOS 6.6 has a fatal bug regarding IPv6.

I pull a /64 prefix from my ISP via DHCPv6-PD which will then put into the 'pool-ipv6' pool, then I advertise the 'pool-ipv6' on 'bridge-local' interface so that devices in the network can get IPv6 addresses via SLAAC. With routerOS 6.6, if I recycle the PPPoE connection, the DHCPv6-PD client will pull a new /64 from ISP, the 'pool-ipv6' pool will be updated to the new prefix, but the advertised prefix at the 'bridge-local' interface is still the old one from the previous connection.

Has to downgrade to 6.5 to fix this back.

bysard · Sat Nov 09, 2013 8:02 am

Reading this i'm still afraid (very discouraged) to upgrade our network to 6.x (200+ RB devices). We still use RoS 5.25 or 5.26. Because Rb2011xx's don't like RoS 5.x we just put them in a closet and have replaced them with different devices. It's much cheaper then waiting for a "rock-solid" solution from MT. I must admit I lost my faith in RB products. Either it's hardware, software, upgrades, something is always wrong. The network was 100% Mikrotik even 6 months ago, now it's more like 70% and dropping. We also completely replaced all wireless devices with UBNT. Our network is upgradable to RoS 7.x so i'm expecting very low MT percentage when the time comes.

RoS is version 6.6 and still this many bugs, so I'm especially surprised when I see an admin posting a sentence like "We didn't get any email to support, so we couldn't have fixed that" and he or they were posting in the same thread numerous times, so they were aware of the problem and could have forwarded it to the responsible person within the company - but they didn't! Just because the users themselves didn't send an email again describing the same problem to support, nothing happened?! This sounds like someone doesn't give a damn and it's not the users (since we are all beta-testers). I've sent many tickets to support over the years and lots of them were solved and I also posted their replies on the forum, so other users can potentially benefit from their answers. In my opinion this can also be done by admins, especially when so many users report the same problem over and over again.

br,
bysaRD

steen · Sat Nov 09, 2013 9:24 am

Hello Folks!

We upgraded relatively many devices this time: RB750, RB600, RB333, CCR1016, RB2011UAS-2HnD
Not yet any RB411.

So far everything seems fine.

Observation from previous release:
CCR: Ethernet port stopped passing traffic for no reason. Mitigation from mikrotik was new beta firmware and disabled LCD. So far working.
RB411: Ethernet port stopped passing traffic for no reason. Mitigation, restart of RB400 and basestation
After those mitigations the problems has not come back what we know.

Attila · Sat Nov 09, 2013 10:04 am

Dear Normis.!

Multicast routing (PIM-SM) same source to different interface still doesn't work in CCR1036.

Ticket#2013020566000411

The device cannot be used for multicast routing since available.

So we can not use in IPTV environment.
Please fix this bug!

Sincerely,

Attila

honzam · Sat Nov 09, 2013 10:20 am

Where is the mentioned this change? http://forum.mikrotik.com/viewtopic.php ... 75#p394593
I don´t see it in changelog 6.6
This is a fix for a problem, that was introduced in v6.6rc1, so it is not mentioned, since regular users upgrade from v6.5

As I understand that you have made some NV2 changes in 6.6rc1. In final 6.6 you have a problem fixed.
It is therefore logical that have changed the code NV2. Why is it not mentioned anywhere? On what is the changelog?

steen · Sat Nov 09, 2013 11:11 am

Hello Folks!

We upgraded relatively many devices this time: RB750, RB600, RB333, CCR1016, RB2011UAS-2HnD
Not yet any RB411.

So far everything seems fine.

Observation from previous release:
CCR: Ethernet port stopped passing traffic for no reason. Mitigation from mikrotik was new beta firmware and disabled LCD. So far working.
RB411: Ethernet port stopped passing traffic for no reason. Mitigation, restart of RB400 and basestation
After those mitigations the problems has not come back what we know.

Now I begin quote myself

All our network is now upgraded to RoS6.6
Very many RB411 is included now also RB450G and some else devices.

Everything went fine all working normally, wireless and wired networks is operating as usal.

npero · Sat Nov 09, 2013 1:56 pm

Nice release for killing x86 for now 3 router is dead and counting.

Erastus · Sat Nov 09, 2013 4:37 pm

Jip,

Mikrotik is going out of their way to kill 386 yet MY hardware & software does not compare to other suppliers.
I guess it is time to knock on Ubuntu door to get this software done as free software.
If I have to go to the slow devices I might as well spend the money and get them developing a AP
We need to replace a lot of 386 systems and then there is no MT product capable of proper hardware.
Even the R52Hn under high data transfer the final amplifier shorts and then over heat and and and ...

Tired of software bugs reintroduced and then they don't even answer tickets via email etc.
Or use a kernel that is working with PCie and or they are not willing to tell on what board they have tested the new RB14e and and and ..

I guess it is time to get a open source project supported by Centos / Ubuntu

Ehman · Sat Nov 09, 2013 5:20 pm

software bugs reintroduced

I'm still searching for the answer to that..

armandfumal · Sat Nov 09, 2013 6:12 pm

BUG REPORT
Ros 6.6
Routerboard: 3.10
Platerforme: RB1100AH2

Something has changed about Vlan because when I upgrade from 6.5 to 6.6, I loose Vlan to ethernet.
but I found a workaround to stay in 6.6

previous config in 6.5, I have a bridge, vlan30 created on the bridge, ethernet7 in the bridge, so vlan30 was goind thru the ethernet 7...
working ok.

but in 6.6, no more vlan30 behind the ethernet7

I changed:
created a second bridge, create a vlan30 on every ethernet interface that I need to have vl30 and put all vl30 interface in the new separate bridge and now the vl30 is going thru my ethernet port...

So It seems when you have vlan tagged on a bridge and the eternet port into the bridge, no more tagged vlan(coming from the bridge) is sended to ethernet...

that put down all vlan...

Armand

npero · Sat Nov 09, 2013 7:16 pm

Nice release for killing x86 for now 3 router is dead and counting.

One come back after 5 minutes reboot without proper shutdown and now working, two need to remove hdd reinstall and now working. Both x86 failed to update have some chipset on mainboard nForce 3, for that reason I replace one with Intel mainboard and waiting next update.

On Intel mainboard chipset P35 processor is E2160 but MT don't see both cores. Try disable multi-cpu after that router is dead, remove HDD insert in some old P4 with HT MT see virtual and real core, enable multi-cpu, back to router, don't see both core again disable but with monitor connected router is booting when booting loading some service like Linux distribution never seen this on MT, two beep and froze.
Again remove hdd back to old PC enable multi cpu, back reset configuration don't see both core no more time to experiment.
All network is on 6.6 now.

jcem · Sat Nov 09, 2013 9:39 pm

BUG REPORT
Ros 6.6
Routerboard: 3.10
Platerforme: RB1100AH2

Something has changed about Vlan because when I upgrade from 6.5 to 6.6, I loose Vlan to ethernet.
but I found a workaround to stay in 6.6

previous config in 6.5, I have a bridge, vlan30 created on the bridge, ethernet7 in the bridge, so vlan30 was goind thru the ethernet 7...
working ok.

but in 6.6, no more vlan30 behind the ethernet7

I changed:
created a second bridge, create a vlan30 on every ethernet interface that I need to have vl30 and put all vl30 interface in the new separate bridge and now the vl30 is going thru my ethernet port...

So It seems when you have vlan tagged on a bridge and the eternet port into the bridge, no more tagged vlan(coming from the bridge) is sended to ethernet...

that put down all vlan...

Armand

More problems with ROS6.6 and vlan

1. After upgrade from 5.26 to 6.6, vlan change interface or become ”unknown" on some RB's

2. ROS 6.6 has killed one RB2011 and I can se it reboots every minute. It’s out on an island,
so I can’t get to it other then on a ferry on monday - not good
- I know this RB is running Hotspot on vlan and I guess that is what’s causing kernel failure

3. Several other RB's I had to downgrade to 6.5 since vlan no longer pass traffic

MT - Please pull the upgrade from the download before more people loose there equipment

MT - Please put more in change log so people know what to test for before upgrade remote RB’s

RGDS
Jörgen

Petzl · Sat Nov 09, 2013 9:40 pm

Scheduler Bug also not fixed in 6.6

http://forum.mikrotik.com/viewtopic.php?f=2&t=77876

Note this is caused by different system language than english

ojsa · Sat Nov 09, 2013 9:55 pm

BUG REPORT
Ros 6.6
Routerboard: 3.10
Platerforme: RB1100AH2

Something has changed about Vlan because when I upgrade from 6.5 to 6.6, I loose Vlan to ethernet.
but I found a workaround to stay in 6.6

previous config in 6.5, I have a bridge, vlan30 created on the bridge, ethernet7 in the bridge, so vlan30 was goind thru the ethernet 7...
working ok.

but in 6.6, no more vlan30 behind the ethernet7

I changed:
created a second bridge, create a vlan30 on every ethernet interface that I need to have vl30 and put all vl30 interface in the new separate bridge and now the vl30 is going thru my ethernet port...

So It seems when you have vlan tagged on a bridge and the eternet port into the bridge, no more tagged vlan(coming from the bridge) is sended to ethernet...

that put down all vlan...

Armand

Seem to experience the same problem on RB433GL with ROS 6.6.

ojsa · Sat Nov 09, 2013 10:38 pm

BUG REPORT
Ros 6.6
Routerboard: 3.10
Platerforme: RB1100AH2

Something has changed about Vlan because when I upgrade from 6.5 to 6.6, I loose Vlan to ethernet.
but I found a workaround to stay in 6.6

previous config in 6.5, I have a bridge, vlan30 created on the bridge, ethernet7 in the bridge, so vlan30 was goind thru the ethernet 7...
working ok.

but in 6.6, no more vlan30 behind the ethernet7

I changed:
created a second bridge, create a vlan30 on every ethernet interface that I need to have vl30 and put all vl30 interface in the new separate bridge and now the vl30 is going thru my ethernet port...

So It seems when you have vlan tagged on a bridge and the eternet port into the bridge, no more tagged vlan(coming from the bridge) is sended to ethernet...

that put down all vlan...

Armand
Seem to experience the same problem on RB433GL with ROS 6.6.

Same problem with the ROS 6.7RC1 compiled at 17:37 at 08.nov.

Sat Nov 09, 2013 11:21 pm

ntp don't work correct

Sat Nov 09, 2013 11:47 pm

NTP still don't work, routerboard 2011 UAS-RM, omnitik, rb 751G-2HnD

RafGan · Sun Nov 10, 2013 12:00 am

the version 6.6 not working vlan on the bridge . if disable enable starts to work. After rebooting everything is repeated

In my case vlans are death on bonding interface. Disable/enable doesn't help - only VID change and back, turns to live again.
V 6.5 doesn't have this serious bug (!!!).

greek · Sun Nov 10, 2013 1:20 am

the version 6.6 not working vlan on the bridge . if disable enable starts to work. After rebooting everything is repeated
In my case vlans are death on bonding interface. Disable/enable doesn't help - only VID change and back, turns to live again.
V 6.5 doesn't have this serious bug (!!!).

I confirm it!

I has the same bug. Vlan on bondings stoped work after upgrade on 6.6 (CCR1026)

This is a disaster for my big netwok. Network idle two hours.

samsung172 · Sun Nov 10, 2013 3:51 am

I confirm vlan bug in 6.6.

shinobi · Sun Nov 10, 2013 8:05 am

I confirm vlan bug in 6.6.

me too

nurmia · Sun Nov 10, 2013 10:47 am

very fine news.i have finding it.thanks for great post.

bax · Sun Nov 10, 2013 12:42 pm

My RB 433 - upgrade normal ... but my x86 on intel atom is dead ... stay on loading

onnoossendrijver · Sun Nov 10, 2013 1:04 pm

My atom upgraded fine.. Intel DN2800MT board.

szastan · Sun Nov 10, 2013 1:37 pm

the version 6.6 not working vlan on the bridge . if disable enable starts to work. After rebooting everything is repeated

the same thing happened to me today on RB2011. This is unacceptable. I was trying to move to RB as device for business client, but I can't take the risk of loosing my clients. Moving back to L2 switches.

keraia · Sun Nov 10, 2013 2:12 pm

It seems that everybody has different problems, even with the same hardware.
I have a couple of RB435Gs for a private link. I upgraded from 5.26 to 6.0->6.1->6.2->6.3->6.4->6.5 and finally 6.6, whenever each version became available.
Apart from the known (and fixed) problem of winbox panels requiring a reopening for proper operation (occurred around 6.4-6.5), the only problem I had was with the graphs and in particular the wireless LAN interface graphs, where each version of RouterOS kept and displayed a different portion of history graphs. In particular, 6.5 refused to update the "Weekly" graph!. Now in 6.6 all wlan graphs started from scratch. The other graphs seem to work fine though.
But these are minor problems. All basic router operations functioned without a hitch at each upgrade: no configuration transfer problems, no port flapping, no link loss, no DHCP problems, no DNS problems, nothing.
Well, maybe I am lucky.

tron · Sun Nov 10, 2013 3:38 pm

You should do more practical tests before you release new RoS version, Mikrotik.

I will use v6.5, v6.6 isn't good choice for me. :/ Thanks for your comments, guys.

incoming429 · Sun Nov 10, 2013 9:03 pm

Upgraded RB532 from 5.25 my IP pool got lost, I'm not able to create a new one and 'ip-pool' is causing 100% cpu usage:

miszak · Mon Nov 11, 2013 3:39 am

RB493G; on ethert 1 and 6, 9 can not see any movement, tx / rx status stopped for some values, but the web all values are correct. All ports are connect to "switch1".

andrijco · Mon Nov 11, 2013 6:34 am

Hi, I recently tried to update my x86 routers from 5.22 to 6.6

and after all I need to roolback to 5.23 =(

reason:

on all my x86 routers I have installed Intel PRO/1000 PT Dual Port Server Adapter's
and driver in system do something strange

1. If manual Ethernet speed is setup to 1Gbps, real speed is 10Mbps with no connectivity
2. If Auto mode is used - ports flaps constantly 2-3 times on minute with random connectivity

bug reproduced on 5 routers with empty configuration

P.S. ROS 6.6 tilegx works great

Mon Nov 11, 2013 9:10 am

MT are you kidding? Firmware bug on rb2011 is still here, what's wrong with you?
Please clarify what firmware bug are you talking about ?
NAND corruption with firmware 3.10, and as a result it's allmost not possible to upload anything via ftp or webUI, so now i should find a windows computer with buggy netinstall and spent time to downgrade Ros6.4 and than to 3.0.8 firmware

I am not aware of such problems. Could you please tell me your MikroTik support ticket number, so I can check the status of this problem?

Mon Nov 11, 2013 9:12 am

CRS125 after updating the E353 modem dangled on one of devices. After upgrade on second device was missing in action E3131 modem .
Disconnect the modem from the device. Entry into System-> Resources hmm I see Ghost USB modem on WTF visible.

Factory reset. Waiting 1 min. Modem connection.
WFT No serial port.

The CRS has a serial port on it's back side.

Mon Nov 11, 2013 10:07 am

RB493G; on ethert 9 can not see any movement, tx / rx status stopped for some values, but the web all values are correct.

Please send us your supout.rif file to support@mikrotik.com

Mon Nov 11, 2013 10:38 am

Upgraded RB532 from 5.25 my IP pool got lost, I'm not able to create a new one and 'ip-pool' is causing 100% cpu usage:

Please post your configuration, or email support with more details. How was the pool configured, and where was it used?

Mon Nov 11, 2013 10:52 am

NTP still don't work, routerboard 2011 UAS-RM, omnitik, rb 751G-2HnD

Weird, i have ge-upgrade with netinstall and upgrade packages 6.6 now is everything fine but i see no ntp server settings.

Mon Nov 11, 2013 10:58 am

NTP still don't work, routerboard 2011 UAS-RM, omnitik, rb 751G-2HnD
Weird, i have ge-upgrade with netinstall and upgrade packages 6.6 now is everything fine but i see no ntp server settings.

NTP server is not installed by default. you must manually install the NTP package to have the server.

prague · Mon Nov 11, 2013 11:43 am

the version 6.6 not working vlan on the bridge . if disable enable starts to work. After rebooting everything is repeated
In my case vlans are death on bonding interface. Disable/enable doesn't help - only VID change and back, turns to live again.
V 6.5 doesn't have this serious bug (!!!).

I have the same issue! 6.5 is more stable..

Mon Nov 11, 2013 11:44 am

We have confirmed a VLAN problem. It is fixed in v6.7

Neilson · Mon Nov 11, 2013 11:57 am

@normis

Is the VLAN issues fixed in the 6.7RC1 from 8/11/2013? or is it fixed now and a new build coming soon?

Regards
Alexander

Mon Nov 11, 2013 11:58 am

@normis

Is the VLAN issues fixed in the 6.7RC1 from 8/11/2013? or is it fixed now and a new build coming soon?

Regards
Alexander

coming soon

JanezFord · Mon Nov 11, 2013 12:24 pm

We have confirmed a VLAN problem. It is fixed in v6.7

This is very very very serious problem in basic functionality of a router ... how can this happen? dont you preform some basic functionality tests before releasing software?

JF

Mon Nov 11, 2013 12:25 pm

Temporary fix for VLAN issue - provide static MAC address for the bridge it is on.

MrYan · Mon Nov 11, 2013 12:41 pm

Upgraded a router to 6.6 and hit a problem with bridging Ethernet (BCP) over SSTP:

10:39:19 sstp,info <sstp-0>: waiting for call...
10:39:19 sstp,info sstp-in1: authenticated
10:39:19 sstp,info sstp-in1: connected
10:39:19 sstp,info,account door logged in, 0.0.0.0
10:39:19 sstp,error could not add bridge port: sstp-in1 is not allowed to be put in bridge (6)
10:39:19 sstp,info,account door logged out, 0 72 76 6 7
10:39:19 sstp,info sstp-in1: terminating... - could not add bridge port: sstp-in1 is not allowed to be put in bridge (6)
10:39:19 sstp,info sstp-in1: disconnected

This was working on 6.5 but breaks on 6.6 - is there some new config I need to apply or is this a bug?

Matt.

theprism · Mon Nov 11, 2013 1:53 pm

normis, is there any ETA for Windows <-> Mikroitk ipsec/l2tp fix - v6.7, v7.1, 8.9, v10.1?
See:
v6.5 http://forum.mikrotik.com/viewtopic.php ... 50#p391817
v6.3 http://forum.mikrotik.com/viewtopic.php ... 50#p385736
v6.1 http://forum.mikrotik.com/viewtopic.php ... 50#p373740

Mon Nov 11, 2013 2:06 pm

normis, is there any ETA for Windows <-> Mikroitk ipsec/l2tp fix - v6.7, v7.1, 8.9, v10.1?
See:
v6.5 http://forum.mikrotik.com/viewtopic.php ... 50#p391817
v6.3 http://forum.mikrotik.com/viewtopic.php ... 50#p385736
v6.1 http://forum.mikrotik.com/viewtopic.php ... 50#p373740

Have you contacted MikroTik support? What is the ticket number, I could check the status of the issue reported.

efaden · Mon Nov 11, 2013 2:09 pm

Please clarify what firmware bug are you talking about ?
NAND corruption with firmware 3.10, and as a result it's allmost not possible to upload anything via ftp or webUI, so now i should find a windows computer with buggy netinstall and spent time to downgrade Ros6.4 and than to 3.0.8 firmware
I am not aware of such problems. Could you please tell me your MikroTik support ticket number, so I can check the status of this problem?

I'd like to know this myself before I update my devices.

theprism · Mon Nov 11, 2013 2:20 pm

normis, is there any ETA for Windows <-> Mikroitk ipsec/l2tp fix - v6.7, v7.1, 8.9, v10.1?
See:
v6.5 http://forum.mikrotik.com/viewtopic.php ... 50#p391817
v6.3 http://forum.mikrotik.com/viewtopic.php ... 50#p385736
v6.1 http://forum.mikrotik.com/viewtopic.php ... 50#p373740
Have you contacted MikroTik support? What is the ticket number, I could check the status of your issue.

Well, it's not only my issue, see previous posts for other releases and you'll find more people complaining.
So, as you can see in my (and others) previous posts - NO, I haven't contacted mikrotik support and I won't, so I have no ticket number. It's your job to test your releases before publishing them. If something is broken (THAT WAS WORKING) by a new release - you have to fix it with the information you get from any sources and not only "support".

You must understand that I won't spend time in upgrading my devices just for your tests and downgrading back to have my PROD environment back online.

Regards,
T.P.

Mon Nov 11, 2013 2:21 pm

normis, is there any ETA for Windows <-> Mikroitk ipsec/l2tp fix - v6.7, v7.1, 8.9, v10.1?
See:
v6.5 http://forum.mikrotik.com/viewtopic.php ... 50#p391817
v6.3 http://forum.mikrotik.com/viewtopic.php ... 50#p385736
v6.1 http://forum.mikrotik.com/viewtopic.php ... 50#p373740
Have you contacted MikroTik support? What is the ticket number, I could check the status of your issue.
Well, it's not only my issue, see previous posts for other releases and you'll find more people complaining.
So, as you can see in my (and others) previous posts - NO, I haven't contacted mikrotik support and I won't, so I have no ticket number. It's your job to test your releases before publishing them. If something is broken (THAT WAS WORKING) by a new release - you have to fix it with the information you get from any sources and not only "support".

You must understand that I won't spend time in upgrading my devices just for your tests and downgrading back to have my PROD environment back online.

Regards,
T.P.

I was asking, because those people who did contact support, turned out to have different mistakes in the configuration, and their tickets were successfully closed. If you still seeing IPSec problems, you should contact support. We currently have no known bugs with IPsec, including the issues reported in mentioned topics.

nurmia · Mon Nov 11, 2013 2:34 pm

thank you share this.

Ehman · Mon Nov 11, 2013 2:52 pm

Upgraded a router to 6.6 and hit a problem with bridging Ethernet (BCP) over SSTP:

10:39:19 sstp,info <sstp-0>: waiting for call...
10:39:19 sstp,info sstp-in1: authenticated
10:39:19 sstp,info sstp-in1: connected
10:39:19 sstp,info,account door logged in, 0.0.0.0
10:39:19 sstp,error could not add bridge port: sstp-in1 is not allowed to be put in bridge (6)
10:39:19 sstp,info,account door logged out, 0 72 76 6 7
10:39:19 sstp,info sstp-in1: terminating... - could not add bridge port: sstp-in1 is not allowed to be put in bridge (6)
10:39:19 sstp,info sstp-in1: disconnected

This was working on 6.5 but breaks on 6.6 - is there some new config I need to apply or is this a bug?

Matt.

I've ALSO had the exact same issue and some others with my 1100AHx2 after upgrading to v6.5 BUG .... V6.6rc1 BUG , and decided that I didn't have time for this bs...so I've downgraded to my trusty 3 random reboots per day version... 6.0rc13

Chupaka · Mon Nov 11, 2013 5:30 pm

Temporary fix for VLAN issue - provide static MAC address for the bridge it is on.

and what about bonding?..

Rudios · Mon Nov 11, 2013 5:51 pm

Upgraded a router to 6.6 and hit a problem with bridging Ethernet (BCP) over SSTP:

10:39:19 sstp,info <sstp-0>: waiting for call...
10:39:19 sstp,info sstp-in1: authenticated
10:39:19 sstp,info sstp-in1: connected
10:39:19 sstp,info,account door logged in, 0.0.0.0
10:39:19 sstp,error could not add bridge port: sstp-in1 is not allowed to be put in bridge (6)
10:39:19 sstp,info,account door logged out, 0 72 76 6 7
10:39:19 sstp,info sstp-in1: terminating... - could not add bridge port: sstp-in1 is not allowed to be put in bridge (6)
10:39:19 sstp,info sstp-in1: disconnected

This was working on 6.5 but breaks on 6.6 - is there some new config I need to apply or is this a bug?

Matt.

I am experiencing the same problems with pptp tunnels.
Bcp is completely broken in 6.6
Besides this BCP is bad in V6.x anyhow.
DHCP requests are not working properly, while they do in 5.26

steen · Mon Nov 11, 2013 6:55 pm

Hello Folks!

A minore problem is found...

In RB333, reproduced in all RB411, RB600, RB450G and RB750.
A problem, static dns entry was added for a device with fictive name A and ip fictive adress n.n.n.n
Then some time later, a new name was added fictive name B with same ip as A.
After some time, static entry A was removed, and so it was.

But in the dynamic cached dns records, it is still there and we can not remove it.
Yet after several weeks it was there in the cache. And it is not anywhere else in our networks DNS:es or anywhere else.

We had to reboot the RB333 to get rid of that dynamic entry.

Anyone who have seen this from before ?

Stillhard · Tue Nov 12, 2013 6:26 am

Hi, normis

Continue from here,

ros code

/tool user-manager customer pr
Flags: X - disabled 
 0   login="xxxxxx" password="xxxxxx" backup-allowed=yes currency="Rupiah" time-zone=+07:00 permissions=owner signup-allowed=no paypal-allowed=no paypal-secure-response=no 
     paypal-accept-pending=no

1.jpg

2.jpg

As you can see with profile Starts = NOW, the time in Users UI is not equal to the actual time.

Tue Nov 12, 2013 7:49 am

so, currently there's no way to limit total amount of huge traffic on MikroTik devices?..

Queue tree - Interface HTB is the fastest one, HTB Global requires little bit more processing.

unm · Tue Nov 12, 2013 8:48 am

so, currently there's no way to limit total amount of huge traffic on MikroTik devices?..

Queue tree - Interface HTB is the fastest one, HTB Global requires little bit more processing.

Big problem on OS v6.6. HTB crash router with big traffic, but all ok on 5.26 all. I'll send all support.rif, export config and screenshot ti support@mikrotik.com. But no result.
Screen attached

ySgAIsvNBI · Tue Nov 12, 2013 9:24 am

so, currently there's no way to limit total amount of huge traffic on MikroTik devices?..

Queue tree - Interface HTB is the fastest one, HTB Global requires little bit more processing.

regardless of whether htb is on interface or global, CCR cannot police/shape more than 600-800Mbps per queue including its child queues
there is no alternative approach because only one top level queue can attach to an interface or global in HTB

timberwolf · Tue Nov 12, 2013 10:39 am

so, currently there's no way to limit total amount of huge traffic on MikroTik devices?..

Queue tree - Interface HTB is the fastest one, HTB Global requires little bit more processing.
regardless of whether htb is on interface or global, CCR cannot police/shape more than 600-800Mbps per queue including its child queues
there is no alternative approach because only one top level queue can attach to an interface or global in HTB

The Tilera cores aren't that powerfull as single instances. To achieve higher throughput a new queuing strategy which makes use of multiple cores would be needed.

RafGan · Tue Nov 12, 2013 10:43 am

Temporary fix for VLAN issue - provide static MAC address for the bridge it is on.

Doesn't help in bonding - made before upgrade to v6.6

Tue Nov 12, 2013 10:53 am

so, currently there's no way to limit total amount of huge traffic on MikroTik devices?..

Queue tree - Interface HTB is the fastest one, HTB Global requires little bit more processing.
regardless of whether htb is on interface or global, CCR cannot police/shape more than 600-800Mbps per queue including its child queues
there is no alternative approach because only one top level queue can attach to an interface or global in HTB

The screenshot is showing an x86 machine. We have received unm's Supout.rif file, and we are looking into it

unm · Tue Nov 12, 2013 11:06 am

The screenshot is showing an x86 machine. We have received unm's Supout.rif file, and we are looking into it

Thanks very much, problem is still active for as. This router used for PPPoE server (appr 1500 clients, 250Mb traffic max). Queue standard(dynamic pppoe queue simple) and queue tree use(firewall traffic mark and queue tree), thas all.

Tue Nov 12, 2013 11:09 am

The screenshot is showing an x86 machine. We have received unm's Supout.rif file, and we are looking into it
Thanks very much, problem is still active for as. This router used for PPPoE server (appr 1500 clients, 250Mb traffic max). Queue standard(dynamic pppoe queue simple) and queue tree use(firewall traffic mark and queue tree), thas all.

Support has answered your ticket. The issue is fixed in v6.6 already.

ajimenez · Tue Nov 12, 2013 12:04 pm

Hi, I recently tried to update my x86 routers from 5.22 to 6.6

and after all I need to roolback to 5.23 =(

reason:

on all my x86 routers I have installed Intel PRO/1000 PT Dual Port Server Adapter's
and driver in system do something strange

1. If manual Ethernet speed is setup to 1Gbps, real speed is 10Mbps with no connectivity
2. If Auto mode is used - ports flaps constantly 2-3 times on minute with random connectivity

bug reproduced on 5 routers with empty configuration

P.S. ROS 6.6 tilegx works great

Hi,

I have also reproduced issue 1. I tried to upgrade one of the my x96 router (5.26) to v6.6 but I observed very high latencies. Then I tried to change the ethernet speed with the same result than adrijco. I solved the problem swaping the application of the network cards (Driver compatibility?)
Also, I have observed how sometimes the router hang up when I changed the ethernet speed. Winbox connection freezed and the system wasn't able to reboot. I have to disconnect the router.

samsung172 · Tue Nov 12, 2013 12:05 pm

Issues found in 6.6:

When using ipv6 over pppoe, my NAS got 50% CPU to dhcp (x86). Profile showing dhcp using 50% CPU. usualy the box use 6-9% CPU.

Upgrade make all packages activated, eg wireless package on rb1100

CA certificate error (I dont use CA)

Something happend to my l2tp "terminator". Clients did not get a sucsesfull login. Downgraded to 6.5.

Tue Nov 12, 2013 1:36 pm

Anyone with NTP synchronisation issues, could you please email support with your supout.rif file ? Thanks!

dnyl · Tue Nov 12, 2013 3:14 pm

Where is my posts? Why mikrotik deleted my post, what contains the real things about mikrotiks update habit?
I wrote to support, but no answer.

Ticket#2013111166000632

Tue Nov 12, 2013 3:17 pm

Where is my posts? Why mikrotik deleted my post, what contains the real things about mikrotiks update habit?
I wrote to support, but no answer.

Ticket#2013111166000632

Please follow our forum posting guidelines, and we will not have to remove your posts:
http://forum.mikrotik.com/faq.php

This is a community forum, and your personal requests towards MikroTik management, should be sent to our E-mail, not to our forum users.

nurmia · Tue Nov 12, 2013 6:06 pm

thanks for sharing in valuable info.

maxiscool · Tue Nov 12, 2013 7:08 pm

Was posting my issue here (http://forum.mikrotik.com/viewtopic.php?f=13&t=78764), care to give some help if this is some VLAN issue on the FW?

nurmia · Wed Nov 13, 2013 6:34 am

thank you for your post.

Wed Nov 13, 2013 8:36 am

What's new in 6.6 (2013-Nov-07 13:04):

*) certificates - fixed certificate import;

Good I can import fine now, however, I can't change the internal name of the certificate anymore:
'couldn't change certificate xxx - certificate subject is read only'

Did you upgrade to the final release, the one from our webpage? We can't repeat the problem, we can change names just fine. Maybe you are running a pre-release v6.6?

maxiscool · Wed Nov 13, 2013 8:59 am

Welcome. May I know is this a bug that going to be fix on next release?

Wed Nov 13, 2013 9:13 am

Welcome. May I know is this a bug that going to be fix on next release?

VLANs have problems in v6.6, this is known, and v6.7 will have this fixed. We will release the new version today or tomorrow.

Stillhard · Wed Nov 13, 2013 9:27 am

Normis, dont forget about this too: http://forum.mikrotik.com/viewtopic.php ... 83#p395250

Wed Nov 13, 2013 9:32 am

Normis, dont forget about this too: http://forum.mikrotik.com/viewtopic.php ... 83#p395250

As I already asked you in the other topic, please check if your timezone is also correct in the "/system clock" menu. Your RouterOS system time should also match the same settings.

Stillhard · Wed Nov 13, 2013 10:37 am

Normis, dont forget about this too: http://forum.mikrotik.com/viewtopic.php ... 83#p395250
As I already asked you in the other topic, please check if your timezone is also correct in the "/system clock" menu. Your RouterOS system time should also match the same settings.

As i said many times, it is correctly set always...

ros code

/system clock> pr
            time: 15:36:18
            date: nov/13/2013
  time-zone-name: Asia/Jakarta
      gmt-offset: +07:00

armandfumal · Wed Nov 13, 2013 7:35 pm

What's new in 6.7rc1 (2013-Nov-11 15:04):

*) fixed hairpin nat on bridge with use-ip-firewall=yes;
*) fixed vlan on bridge after reboot having 00:00:00:00:00:00 mac address;

littlebill · Thu Nov 14, 2013 3:04 am

RB2011 6.6

Ntp client shows stopped in WinBox even though enabled

CLI indicates its synced and shows date/time correctly

apparently this required a reboot to clear it up

littlebill · Thu Nov 14, 2013 3:17 am

RB2011 6.6

DNS Static change bug

removing a static dns name does not remove it from cache even after a flush, only way to clear is a reboot

Thu Nov 14, 2013 7:46 am

RB2011 6.6

Ntp client shows stopped in WinBox even though enabled

CLI indicates its synced and shows date/time correctly

apparently this required a reboot to clear it up

Are you using SNTP or the NTP package? Can you make a supout.rif file and send to support ?

ATROX · Thu Nov 14, 2013 9:28 am

Bug found.
If i connected over SSTP in my Enterprise LAN, and begining download file over 400 MByte - my connection is lost. And i can not connection over SSTP - system messages: wrong login or password. After 10-20 minutes is OK, i can connection over SSTP.
Fix please.
Over PPTP connection - no errors, no connection lost.
OS: Routeboard v6.5-v6.6
CCR1016-16
Firmware: v3.10

Thu Nov 14, 2013 1:04 pm

RB2011 6.6

DNS Static change bug

removing a static dns name does not remove it from cache even after a flush, only way to clear is a reboot

I just did some extensive testing, and could not repeat the issue. Here is what I did, please tell me how your test was different:

1. nslookup mikrotik.com returns Address: 159.148.147.196 in my PC
2. added a new static entry in the router, mikrotik.com address 127.0.0.1
3. nslookup mikrotik.com now returns address 127.0.0.1
4. removed this static enry and flushed the cache in router
5. nslookup in the PC again returns original Address: 159.148.147.196

Did you do something else?

littlebill · Thu Nov 14, 2013 5:57 pm

RB2011 6.6

DNS Static change bug

removing a static dns name does not remove it from cache even after a flush, only way to clear is a reboot
I just did some extensive testing, and could not repeat the issue. Here is what I did, please tell me how your test was different:

1. nslookup mikrotik.com returns Address: 159.148.147.196 in my PC
2. added a new static entry in the router, mikrotik.com address 127.0.0.1
3. nslookup mikrotik.com now returns address 127.0.0.1
4. removed this static enry and flushed the cache in router
5. nslookup in the PC again returns original Address: 159.148.147.196

Did you do something else?

i can't repeat it right now. hmm it was bugged even in 5.26, then entry would stay until you rebooted it, it would leave the static entry page but persist in the cache. i will try to follow closer next time

xds · Fri Nov 15, 2013 1:16 pm

Normis Hi , Could you help me with the following (sorry that I write in this thread, there is a vivid discussion

), and so:
I use RB2011UAS-2HnD-IN (v6.6)we have two ISP are connected via first Ether1-(1ISP) - fixed internet, after which use our network to the Internet, another provider are connected via microUSB - 3G Modem (Huawei E1550) - (2ISP) - through this second provider should be administration of the router (using changeip.com), I did everything seems fine, but after binding DNS to IP, connections with the router but occurs in about 2-5 minutes and then disappears when PPP connections off and then turn on change under IP, again from outside connections is with the router and then disappears, it seems to me that the problem is as I tuned the rules in the router itself but a week couldn understand

If someone else wants to help with this issue, I would be happy!
In advance thank you!

maxiscool · Fri Nov 15, 2013 5:20 pm

Welcome. May I know is this a bug that going to be fix on next release?
VLANs have problems in v6.6, this is known, and v6.7 will have this fixed. We will release the new version today or tomorrow.

Is 6.7 going to release anytime soon? Looking forward to get my networking file transfer issue fix.

duppeh · Fri Nov 15, 2013 5:41 pm

can anyone else confirm that port flapping still exist? mine does

rini · Fri Nov 15, 2013 6:57 pm

I just upgrade from V6.3 to V6.6

Very high CPU usage. Mikrotik rb1100ahX2 Cpu was 15-20%, now goes 50-60%.
I have to downgrade

kobuki · Fri Nov 15, 2013 9:17 pm

Certificate export is not working. RB2011LS-IN, Windows 7 SP1 x64 running Winbox. When I press Export in the certificate details window, Winbox exits in an instant and all windows settings since its last start are lost.

Also, cannot rename the certificate, it says "certificate subject is read only!" - it has nothing to do with the cert subject, I just want to change its internal name in ROS.

Otherwise on this router I haven't experienced anything out of the ordinary. Upgrade went without a hitch (from 5.19).

siuslawbb · Fri Nov 15, 2013 9:48 pm

confirmed vlan bug for us as well. From what we can tell, mainly on bridge interfaces for us. major issue for us.

patrickmkt · Fri Nov 15, 2013 11:31 pm

Also, cannot rename the certificate, it says "certificate subject is read only!" - it has nothing to do with the cert subject, I just want to change its internal name in ROS.

I already signaled this one earlier. However if you really need to change it, you can still do it in the terminal mode. You will be able to edit the name there.

littlebill · Sun Nov 17, 2013 3:09 am

Also, cannot rename the certificate, it says "certificate subject is read only!" - it has nothing to do with the cert subject, I just want to change its internal name in ROS.
I already signaled this one earlier. However if you really need to change it, you can still do it in the terminal mode. You will be able to edit the name there.

i have this issue with certifcates as well. seems winbox thinks its 365 regardless of the actual date

http://forum.mikrotik.com/viewtopic.php ... 47#p396047

steen · Sun Nov 17, 2013 8:36 am

RB2011 6.6

DNS Static change bug

removing a static dns name does not remove it from cache even after a flush, only way to clear is a reboot
I just did some extensive testing, and could not repeat the issue. Here is what I did, please tell me how your test was different:

1. nslookup mikrotik.com returns Address: 159.148.147.196 in my PC
2. added a new static entry in the router, mikrotik.com address 127.0.0.1
3. nslookup mikrotik.com now returns address 127.0.0.1
4. removed this static enry and flushed the cache in router
5. nslookup in the PC again returns original Address: 159.148.147.196

Did you do something else?

i can't repeat it right now. hmm it was bugged even in 5.26, then entry would stay until you rebooted it, it would leave the static entry page but persist in the cache. i will try to follow closer next time

Hello Folks!

I can confirm this as well, removed static entries remains in cache even after flush, reboot is the only way to sort it out, which is bad for production devices.

LinFor · Sun Nov 17, 2013 10:20 am

RB2011 6.6
DNS Static change bug
removing a static dns name does not remove it from cache even after a flush, only way to clear is a reboot

I have the same issue, but only with "RegExp" dns entries.

Sun Nov 17, 2013 12:44 pm

I just did some extensive testing, and could not repeat the issue. Here is what I did, please tell me how your test was different:

1. nslookup mikrotik.com returns Address: 159.148.147.196 in my PC
2. added a new static entry in the router, mikrotik.com address 127.0.0.1
3. nslookup mikrotik.com now returns address 127.0.0.1
4. removed this static enry and flushed the cache in router
5. nslookup in the PC again returns original Address: 159.148.147.196

Did you do something else?

i can't repeat it right now. hmm it was bugged even in 5.26, then entry would stay until you rebooted it, it would leave the static entry page but persist in the cache. i will try to follow closer next time
Hello Folks!

I can confirm this as well, removed static entries remains in cache even after flush, reboot is the only way to sort it out, which is bad for production devices.

How do you check? If you try to ping the host your static entry points to, don't forget to flush your Windows DNS cache ('ipconfig /flushdns' from the elevated command prompt). When using nslookup this should not be required, since nslookup contacts the DNS server directly, thus bypassing OS cache. It is also easily explainable why rebooting your Mikrotik device fixes that for you- provided your PC is connected to your Routerboard directly, rebooting the Routerboard causes the network link to go down and then up again, and that flap forces the OS DNS cache flush.

bysard · Sun Nov 17, 2013 12:51 pm

[/quote]
How do you check? If you try to ping the host your static entry points to, don't forget to flush your Windows DNS cache ('ipconfig /flushdns' from the elevated command prompt). When using nslookup this should not be required, since nslookup contacts the DNS server directly, thus bypassing OS cache. It is also easily explainable why rebooting your Mikrotik device fixes that for you- provided your PC is connected to your Routerboard directly, rebooting the Routerboard causes the network link to go down and then up again, and that flap forces the OS DNS cache flush.[/quote]

+1 I can't reproduce this problem myself. Someone must have done something wrong. This works for me.

kobuki · Sun Nov 17, 2013 1:03 pm

i have this issue with certifcates as well. seems winbox thinks its 365 regardless of the actual date

I can also confirm this. Simply forgot to report in my previous post.

zap71 · Sun Nov 17, 2013 2:01 pm

2. ROS 6.6 has killed one RB2011 and I can se it reboots every minute. It’s out on an island,
so I can’t get to it other then on a ferry on monday - not good
- I know this RB is running Hotspot on vlan and I guess that is what’s causing kernel failure

I have exactly this behaviour as well and posted it in a separate topic here: http://forum.mikrotik.com/viewtopic.php?f=2&t=78978.

Zap.

littlebill · Mon Nov 18, 2013 7:25 am

2011 6.6

master ports show link status's even though there is no cable plugged in?

Erastus · Mon Nov 18, 2013 8:06 am

There is still a kernel issue with ver 6.6 it does not work on the new chip sets of Intel ( H61MXP) yet if you load Atheros HAL system on centos it creates an ap on centos and works well.
System does not crash with data transfer.
I am busy getting "openhal" converted to centos to see if it also works.
The main problem with ROS 386X is the kernel is not talking to the newer chip sets.

Still waiting for a response when this will be fixed.

iDen · Mon Nov 18, 2013 8:25 am

CCR-1016
still port-flapping
in more obvious way.
before 6.6 ports were going down and that's it
now it flaps ports many times before restart by watchdog.
first week on 6.6 was very stable I thought it is final holy aid....
no...

Erastus · Mon Nov 18, 2013 8:34 am

We get the to and then we get some packet loss on it

kobuki · Mon Nov 18, 2013 8:22 pm

On RB450G, when I change the MAC of ether1 to match the one required by my ISP, no stats are displayed. "Overall Stats", "Rx Stats", "Tx Stats" windows are empty. Traffic graphs are OK. This error is present since 5.23 as far as can remember, but for all 6.x versions I tried on this router, for sure.

sovaby · Mon Nov 18, 2013 8:52 pm

RouterOS v6.6 winbox from wine
Wrong time fonts
http://minus.com/lbaGLS8nkdLJfj
http://minus.com/lRL8SMPpv0O91

wrong time zone for Europe / Minsk +2
Correct +3

KoDAk · Mon Nov 18, 2013 8:57 pm

Write please List of products
that good to be use RouterOS v6.6+

and what List of products better stay on 5.26

i see that 750 \ 450 and non wo-fi 2011 beter stay on 5.26!

Ivoshiee · Tue Nov 19, 2013 12:24 am

Write please List of products
that good to be use RouterOS v6.6+

and what List of products better stay on 5.26

i see that 750 \ 450 and non wo-fi 2011 beter stay on 5.26!

I saw somewhere that devices with multi core should stay away from the v6 as well.

Erastus · Tue Nov 19, 2013 6:31 am

My Personal opinion is that Ver 6 is a beta version.
To get an answer on the successful implementation is a little far fetch.
In short if you need a working system with some bugs 5.26 is good if you need to get involved with the beta testing of version 6.X then version 6.X is the one you should use.

bysard · Tue Nov 19, 2013 8:25 am

My Personal opinion is that Ver 6 is a beta version.
To get an answer on the successful implementation is a little far fetch.
In short if you need a working system with some bugs 5.26 is good if you need to get involved with the beta testing of version 6.X then version 6.X is the one you should use.

I could agree except 5.25 still has OSPF log errors.

Erastus · Tue Nov 19, 2013 1:26 pm

thinks for the information . We do not use any of the routing mpls , ospf .... or any other feature of the Router OS.
It is simply a ap and a link. We learned that OSPF gave so much problems in the past that we simply installed another suppliers router.
We now have Madwifi installed on Centos and it works with Intell chip set H61M... yet Router OS on the same 386 MB crashes.....

bysard · Tue Nov 19, 2013 3:13 pm

"We learned that OSPF gave so much problems in the past that we simply installed another suppliers router."

We learned the same thing but we needed to keep Mikrotik devices for the time being, so we still use RIP which is working as it should. Tnx for the advice.

honzam · Tue Nov 19, 2013 6:53 pm

in winbox is missing LCD menu on rb2011UiAS-RM

in terminal works

tomaskir · Tue Nov 19, 2013 7:30 pm

"We learned that OSPF gave so much problems in the past that we simply installed another suppliers router."

We learned the same thing but we needed to keep Mikrotik devices for the time being, so we still use RIP which is working as it should. Tnx for the advice.

We have large multi-type multi-area OSPF networks with MikroTik and all works. We peer with our customers with OSPF for redundancy, it all works.
In one particular datacenter, we even peer with the data center with OSPF in a separate OSPF instance, which we then redistribute into our internal OSPF instance, and it all works.

What OSPF issues are you talking about? Have you made sure it was not your config that was wrong? Did you report the issues to support?

bysard · Wed Nov 20, 2013 8:43 am

"We learned that OSPF gave so much problems in the past that we simply installed another suppliers router."

We learned the same thing but we needed to keep Mikrotik devices for the time being, so we still use RIP which is working as it should. Tnx for the advice.
We have large multi-type multi-area OSPF networks with MikroTik and all works. We peer with our customers with OSPF for redundancy, it all works.
In one particular datacenter, we even peer with the data center with OSPF in a separate OSPF instance, which we then redistribute into our internal OSPF instance, and it all works.

What OSPF issues are you talking about? Have you made sure it was not your config that was wrong? Did you report the issues to support?

Yes, I have sent a ticket to support and got answer that it will be solved in 6.x and it was. This is one example thread http://forum.mikrotik.com/viewtopic.php?f=14&t=72231.

timberwolf · Wed Nov 20, 2013 1:51 pm

A new routing engine was announced for v6, but this has been posponed till v7, if I recall this correctly.

villi · Wed Nov 20, 2013 11:39 pm

RouterOS v6.6 winbox from wine
Wrong time fonts
http://minus.com/lbaGLS8nkdLJfj
http://minus.com/lRL8SMPpv0O91

wrong time zone for Europe / Minsk +2
Correct +3

Add to Wineskin Advanced -> Custom Commands: -> export LC_ALL="en_US.UTF-8"

lekozs · Thu Nov 21, 2013 10:52 pm

Take a RB800.
Update it to ROS 6.6.
Update the BootLoader (fw) to 3.10
Put either a Ubiquity SR-71-15 or a DBII F50-N Pro or a Mikrotik R52-nM or a Mikrotik R5-SHPn mini-PCI card into whatever slot you like. (put just one card at the time!)
Turn on the RB800.
Connect with Winbox, and...: there is NO mini-PCI card present under the Interfaces menu.
... now comes the tricky part...: DO NOT turn off the RB800, just do a simple reboot... and, voila... the card appears!

Then try this:... turn it off completely (cut the power) Wait for 30 seconds... Turn it on... and, again, there is no card in the system...!

NOTE: this "trick" was performed ONLY with the cards I mentioned above and ONLY with RB800! Non "n" cards are working flawlessly.
I tried this with 4 different RB800s!
I even did a complete ROS 6.6 Netinstall... but to no avail...

And then, the solution to this problem...:
Downgrade to ROS 6.4 (I did this with Netinstall)
Downgrade the BootLoader to 3.07 (This is the important part!)
Upgrade back to ROS 6.6
... now, put one of the aforementioned cards into whatever slot you like... and they are appearing even during the first cold start.
They aren't vanishing after a power cycle either...

scampbell · Fri Nov 22, 2013 3:40 am

CCR-1016
still port-flapping
in more obvious way.
before 6.6 ports were going down and that's it
now it flaps ports many times before restart by watchdog.
first week on 6.6 was very stable I thought it is final holy aid....
no...

We are also seeing port flap on CCR1036 and RoS 6.6, FW 3.10

SwissWISP · Fri Nov 22, 2013 4:02 am

CCR-1016
still port-flapping
in more obvious way.
before 6.6 ports were going down and that's it
now it flaps ports many times before restart by watchdog.
first week on 6.6 was very stable I thought it is final holy aid....
no...
We are also seeing port flap on CCR1036 and RoS 6.6, FW 3.10

+1
Our CCR1016 does it too. (6.6, FW 3.10)
This needs to be fixed ASAP!!

iDen · Fri Nov 22, 2013 8:18 am

yeah
it looks like this on my site.
before it was like in most screens only 1-2 port flaps. now it's like this already for 2 weeks.

sovaby · Fri Nov 22, 2013 11:49 am

RouterOS v6.6 winbox from wine
Wrong time fonts

Fixed
create launch shortcut with

env WINEPREFIX="~/.wine" LANG="en_US" wine "~/.wine/drive_c/Program Files/winbox/winbox.exe"

tomaskir · Fri Nov 22, 2013 2:30 pm

Upgraded a router to 6.6 and hit a problem with bridging Ethernet (BCP) over SSTP:

I've ALSO had the exact same issue

I am experiencing the same problems with pptp tunnels.
Bcp is completely broken in 6.6

Can any of you guys share your PPP profile, PPP secret, SSTP server and bridge setup exports?

I have tried to recreate this issue, but for me, in a test enviroment, SSTP works with BCP.

ununium · Fri Nov 22, 2013 5:25 pm

yeah
it looks like this on my site.
before it was like in most screens only 1-2 port flaps. now it's like this already for 2 weeks.

What parameters did you use to make watchdog reboot the ccr while experiencing the port flap?

Mine never seems to activate.

sovaby · Fri Nov 22, 2013 6:27 pm

Add to Wineskin Advanced -> Custom Commands: -> export LC_ALL="en_US.UTF-8"

Linux does not have Wineskin this MAK feature

dorijan · Fri Nov 22, 2013 10:10 pm

Hi to all..
I just want to report a minor bug in 6.6 with SXT 5nD r2
Spectral scan is not working correctly if HT TX and RX chain0 is not selected, and only chain1 is selected...

w32pamela · Sun Nov 24, 2013 7:04 pm

Groove 52HPn running default Mikrotik CPE setup with v6.6

Country resets to "no_country_set" when accessed via Winbox. No problem if using WebFig.

fronczek · Sun Nov 24, 2013 10:19 pm

Hello,
RouterOS 6.6 on powerpc 'kills' DHCP server. Be careful. Don't install it on powerpc platform...
'Offering DHCP lease without success' case.

Mon Nov 25, 2013 9:39 am

Hello,
RouterOS 6.6 on powerpc 'kills' DHCP server. Be careful. Don't install it on powerpc platform...
'Offering DHCP lease without success' case.

Which device, and which interface did you see this on?

Rudios · Mon Nov 25, 2013 9:47 am

Upgraded a router to 6.6 and hit a problem with bridging Ethernet (BCP) over SSTP:

I've ALSO had the exact same issue

I am experiencing the same problems with pptp tunnels.
Bcp is completely broken in 6.6
Can any of you guys share your PPP profile, PPP secret, SSTP server and bridge setup exports?

I have tried to recreate this issue, but for me, in a test enviroment, SSTP works with BCP.

Just out the top of my head this is what I've done.
On both routers created a profile with bridge configured.
Both PPtP Client and Server configured to use the created profile. Secret is as default as possible, just username and password.

tomaskir · Mon Nov 25, 2013 11:12 am

Just out the top of my head this is what I've done.
On both routers created a profile with bridge configured.
Both PPtP Client and Server configured to use the created profile. Secret is as default as possible, just username and password.

I managed to replicate the problem with PPTP and SSTP, submitting a support ticket.

iDen · Mon Nov 25, 2013 11:43 am

yeah
it looks like this on my site.
before it was like in most screens only 1-2 port flaps. now it's like this already for 2 weeks.

What parameters did you use to make watchdog reboot the ccr while experiencing the port flap?

Mine never seems to activate.

standard params... just entering ip address of central point in network, cisco...which is directly connected to this CCR via 50cm patch cord

Rudios · Mon Nov 25, 2013 1:19 pm

Just out the top of my head this is what I've done.
On both routers created a profile with bridge configured.
Both PPtP Client and Server configured to use the created profile. Secret is as default as possible, just username and password.
I managed to replicate the problem with PPTP and SSTP, submitting a support ticket.

Could you please keep me posted with the reaction from MikroTik Support.

Adam84 · Mon Nov 25, 2013 4:28 pm

UPS probe still (since RoS 6.2) doesn't work.

UPS.jpg

Mon Nov 25, 2013 4:29 pm

UPS probe still (since RoS 6.2) doesn't work.
UPS.jpg

why do you have such curious UPS model and Serial Number? Did you know that RouterOS only supports APC UPS with the Smart protocol?

Adam84 · Mon Nov 25, 2013 5:15 pm

Cause it's simple and it does the job. It used to work till version RoS 6.1. After that it's like I've posted.

Here is link (in polish, but u can always use google translate;)):
http://www.inetech.pl/?info=products&id=4

EDIT:
Here is screen from properly working version at one of previous RoS:

UPS.jpg

RouteRite · Mon Nov 25, 2013 7:06 pm

Did you know that RouterOS only supports APC UPS with the Smart protocol?

FYI -- I have had good luck with Cyberpower UPS's (http://www.cyberpowersystems.com) -- they emulate APC Smart protocol and are less expensive. I have yet to use them with Mikrotik, but have had success on other Linux systems that use the apcupsd daemon.

scampbell · Tue Nov 26, 2013 9:46 pm

yeah
it looks like this on my site.
before it was like in most screens only 1-2 port flaps. now it's like this already for 2 weeks.

I had two CCR's doing this port flap. In both cases all active ports would flap.

I checked further and found on both units there was at least one device connected that was NOT set to Auto-Negotiate while all CCR Ports WERE set to Auto-Negotiate.

Have now set the appropriate ports to fixed 100Mbps/Full Duplex. So far it has been running over 12 hours and no flapping.......

Why this affects all ports instead of just the one incorrectly set port I cannot say but I will give it time and see if it helps or is just a coincidence.

iDen · Wed Nov 27, 2013 10:23 am

I had two CCR's doing this port flap. In both cases all active ports would flap.

I checked further and found on both units there was at least one device connected that was NOT set to Auto-Negotiate while all CCR Ports WERE set to Auto-Negotiate.

Have now set the appropriate ports to fixed 100Mbps/Full Duplex. So far it has been running over 12 hours and no flapping.......

Why this affects all ports instead of just the one incorrectly set port I cannot say but I will give it time and see if it helps or is just a coincidence.

Right now I have almost 3 days uptime on this CCR.
Haven't changed or done anything.
Version still 6.6...
Mikrotik sent me 6.7 but i'm not going to fu*** with this dev version untill it's official release...
This point is too critical and i just can't restart it anytime i wish to do this.
1 port connected to cisco, 2 ports connected to other mikrotiks.
all of them have auto-negotiation

scampbell · Wed Nov 27, 2013 10:47 pm

I had two CCR's doing this port flap. In both cases all active ports would flap.

I checked further and found on both units there was at least one device connected that was NOT set to Auto-Negotiate while all CCR Ports WERE set to Auto-Negotiate.

Have now set the appropriate ports to fixed 100Mbps/Full Duplex. So far it has been running over 12 hours and no flapping.......

Why this affects all ports instead of just the one incorrectly set port I cannot say but I will give it time and see if it helps or is just a coincidence.

Day two and still no Port Flapping .......

If this keeps on working then the bug is that all ports flap if one port is not correctly set to match its client's device (auto-negotiation=disabled in my case).

armandfumal · Wed Nov 27, 2013 10:55 pm

I had two CCR's doing this port flap. In both cases all active ports would flap.

I checked further and found on both units there was at least one device connected that was NOT set to Auto-Negotiate while all CCR Ports WERE set to Auto-Negotiate.

Have now set the appropriate ports to fixed 100Mbps/Full Duplex. So far it has been running over 12 hours and no flapping.......

Why this affects all ports instead of just the one incorrectly set port I cannot say but I will give it time and see if it helps or is just a coincidence.
Day two and still no Port Flapping .......

If this keeps on working then the bug is that all ports flap if one port is not correctly set to match its client's device (auto-negotiation=disabled in my case).

Is "all ports" are in a same bridge ?

madboy007 · Fri Nov 29, 2013 2:50 pm

wowo,cool~

Zito · Sun Dec 01, 2013 1:43 am

I'm confirm ports are flapping also in RB2011. Directly connected to DSL modems and slaves into bond interface.
Also confirm that vlans onto this bond are unusable after reboot. Needs to be reenable.

Disabling auto-negotiation not helps when interface are bond slaves.

Ivoshiee · Sun Dec 01, 2013 11:53 am

Disabling auto-negotiation not helps...

I recall as reading the 1Gbps spec as mandating the use of auto-negotiation and you can not have standard way for the 1Gbps operation without it.

amorphous · Sun Dec 01, 2013 12:00 pm

Hi, normis

Continue from here,
ros code
/tool user-manager customer pr
Flags: X - disabled 
 0   login="xxxxxx" password="xxxxxx" backup-allowed=yes currency="Rupiah" time-zone=+07:00 permissions=owner signup-allowed=no paypal-allowed=no paypal-secure-response=no 
     paypal-accept-pending=no
1.jpg

2.jpg
As you can see with profile Starts = NOW, the time in Users UI is not equal to the actual time.

this is exactly the same issue happened on my rb951ui-2hnd... i hope you have a fix on this on the next release..

ySgAIsvNBI · Mon Dec 02, 2013 1:53 am

64-bit counter of in/out bytes statistics for bonding and vrrp interfaces do not work on ROS v6.6 x86 but works properly on tile arch.
please fix this bug asap

Zito · Mon Dec 02, 2013 11:35 am

RouterOS 6.5 has also flapping ports issue and high CPU usage. Any v6.x are working correctly?