v6.24 RC

Mon Dec 22, 2014 2:01 pm

We are currently fixing last issues with 6.24, for those that had confirmed bug reports, please check if the issue is resolved. Thank you for helping. Some reported issues may not be addressed in this version, but please do write and report details

We have also included a new NTP package that includes all NTP upgrades until this day, if you use the NTP package for the server functionality, please upgrade to this release

We would appreciate all comments regarding the biggest changes - NTP and Queue Tree

We plan to make a full release of v6.24 soon.

Download:
http://www.mikrotik.com/download/share/ ... latest.zip
http://www.mikrotik.com/download#rc

What's new in 6.24rc10 (2014-Dec-22 13:43):

*) ntp - fixed vulnerabilities;
*) web proxy - fix problem when dscp was not set when ipv6 was enabled;
*) fixed problem where some of ethernet cards do not work on x86;
*) improved CCR ethernet driver (less dropped packets);
*) improved queue tree parent=global performance (especially on SMP systems and CCRs);
*) eoip/eoipv6/gre/gre6/ipip/ipipv6/6to4 tunnels have improved per core balancing on CCRs;
*) fixed tx for 6to4 tunnels with unspecified dst address;
*) fixed vrrp - could sometimes not work properly because of advertising bad set of ip addresses;

alexburke · Mon Dec 22, 2014 5:14 pm

I updated to 6.24rc10 (build 2014-12-22 13:43:03) using the all-in-one routeros-mipsbe-6.24rc10.npk on an RB2011UiAS-2HnD. Upon doing so and rebooting, the ntp and multicast packages were gone from the router, disabling some important functionality. I had to manually fetch and unpack all_files_6.24rc10.zip, and upload the two missing packages and reboot again, to fix it.

Your software development process needs a complete overhaul — you regularly regress on important patches and fixes, and updating RouterOS is always a roll of the dice as this is not the first time installed packages have vanished (not just becoming disabled, but totally gone) during an upgrade. This is not only the case for RC releases, but full production releases as well.

Mon Dec 22, 2014 5:45 pm

alexburke, if you mean routeros-mipsbe-6.24.npk package then it does not include the multicast and ntp pacages, they need to be uploaded toghether with bundle npk.

Mon Dec 22, 2014 7:04 pm

Don't use combined package. Use your own set extracted from all packages zip file.

jd603 · Tue Dec 23, 2014 4:44 am

So.. sorry if this is the wrong place for a "feature" request. ...but isn't it as easy as compiling the x86 linux kernel with PAE support in order to allow 4GB+ RAM? Or would that break old x86 hardware? I've compiled for PAE before and it was simple and apps just worked. Seems like a trivial improvement.

Sob · Tue Dec 23, 2014 6:00 am

No changes in 6to4 routes since rc4? Doesn't seem so, still "no route to host" for anything outside 2002::/16.

[Ticket#2014120466000688]

mihaimikrotik · Tue Dec 23, 2014 6:20 pm

Need changelog confirmation on 6.24 for CCR regarding PPTP fix from 6.23.

On 6.23 when using PPTP, after a while the user connecting to the CCR cannot access the LAN behind the CCR. I have already tried using proxy-arp on the LAN interface to no avail. The only thing that seems to work is using ARP on the LAN interface and rebooting the CCR.

Jeroen1000 · Tue Dec 23, 2014 7:11 pm

Are the NTP fixes related to the recently discovered vulnerabilities? http://www.ubuntu.com/usn/usn-2449-1/

Could you please tell me what version ROS is using? IS it version 4.2.8

armandfumal · Tue Dec 23, 2014 8:28 pm

upgrade to 6.24 release successfully. no issue found at all.
Using GRE, EoIP tunnel, OSPF...

noyo · Tue Dec 23, 2014 8:42 pm

Is the AC is developed? Because that's what it is, it's a disaster, CCQ up and down, speed up and down

Tue Dec 23, 2014 8:53 pm

Yes, wireless ac updates are verry welcome

ochm · Tue Dec 23, 2014 9:29 pm

The Quick set shows latest version 6.24, but at http://www.mikrotik.com/download is latest 6.23 or 6.23.1, and 6.24rc only.
Is the 6.24 really latest non RC version?

ivimail · Wed Dec 24, 2014 12:22 am

on winbox there aren't any reference to indicate that's a RC version

Neilson · Wed Dec 24, 2014 8:59 am

However if you change the link address to get the 6.24 torrent file then you find it isn't being seeded.

It almost feels like they found a show stopper and forgot to pull it from the auto update system.

Regards
Alexander

rpingar · Wed Dec 24, 2014 9:13 am

the new CCR ethernet driver seems working fine because we don't have more drops on VLAN using only-hardware-queue

We had a crash on our PPPOE-Concentrator when huge number of pppoe clients connect/disconnect, sent supout: Tcket#2014122466000113]

We still don't see a good balance of the cores using VPLS / Bridge / PPPOE server:
[Ticket#2014122366000393]

hope to receive soon some answer about these tickets.

regards
Ros

ufm · Wed Dec 24, 2014 10:06 am

We had a crash on our PPPOE-Concentrator when huge number of pppoe clients connect/disconnect, sent supout: Tcket#2014122466000113]

We still don't see a good balance of the cores using VPLS / Bridge / PPPOE server:
[Ticket#2014122366000393]

We not had crash, but have extremally CPU usage in massive pppoe disconnection.

And, also, we have no reply from support (Ticket#2014122166000217). Only letter from robot. Third day. :-/

mars · Wed Dec 24, 2014 11:00 am

wireless-fp package still has a lot of packet loss in 6.23 hope it is fixed

rajibfci · Wed Dec 24, 2014 1:39 pm

IPv6 is new version of internet protocol.

Majklik · Wed Dec 24, 2014 4:15 pm

I have very strange VRRP result in this release ( 6.24 build-time: Dec/23/2014 13:38:45).

I upgraded for test two routers from ROS6.20 (RB450G and RB800) and result is:
VRRPv2 without authorization - both routers become master,
VRRPv2 with AH authorization - this works OK but slave router reports: received packet from 10.51.110.12 AH authentication failed
VRRPv3/IPv4 - both routers becomes master
VRRPv3/IPv6 with no global IPv6 assigned on the vrrp interface - works OK but this is useless configuration sometimes
VRRPv3/IPv6 with two global IPv6 addressess on the vrrp interface - both router report: "received packet from fe80::ff:fe00:X misconfigured IP addresses" and after few router switching both routers became master.

Maggiore81 · Wed Dec 24, 2014 6:17 pm

6.20 is the latest stable for me.
from 6.20 upwards I get very heavy lag connecting to a RB...

about 10 secs to login via winbox.
with 6.20 everything is OK

I have been able to reproduce the problem. The problem of "lag" appears on the RB that is doing NAT MASQUERADE.

For example in my case the RB in my office is a 951, if >6.20 the problem appears.

On the RB that does routing I have no issues.

bajodel · Wed Dec 24, 2014 8:29 pm

Webfig design skin customization problems still there on 6.24 (present in 6.23 and 6.22, not in 6.20 and 6.19)

http://picpaste.com/pics/webfig-skin-de ... 445790.png

linux42 · Wed Dec 24, 2014 9:18 pm

system,error,critical System rebooted because of kernel failure 2014-12-24 21:07
system,error,critical router was rebooted without proper shutdown, probably kernel failure 2014-12-24 21:07

He reboot ccr1036 14 minutes intervals. Do not tested this software !!!! Are we cadaver? !!!!! We live in very big trouble ...

((((

linux42 · Wed Dec 24, 2014 9:27 pm

350 users online devices are constantly reboot. I do not downgrade in an emergency such nonsense.

Ferrograph · Thu Dec 25, 2014 1:01 am

Guys,

Upgraded a 951 from 6.17 and some of the config got lost:

# all firewall config
# Interface names
# all wireless
# bridges lost

And a few other odd things like the Wireless SSID was the same as 'Identity', system password was also lost.

deanMKD1 · Thu Dec 25, 2014 1:48 am

Please dont post buggy versions on RouterOS anymore. With every new release if fix something, the rest is buggy.. Please make one or two month on new version ,then test carrefully, and finally post as stable version.

honzam · Thu Dec 25, 2014 6:03 pm

Dear MK team. Please unify your opinion on the new versions of ROS.
On download page is still 6.23

On forum is from Tuesday writed so latest version is 6.24

Auto upgrade - lastest version is 6.24.

Take in that order.
Have a nice Christmas

blagynchy · Fri Dec 26, 2014 1:27 am

From 6.x and 3.1x(firmware) interface leds are GONE only green light is on but no blinking (with cable connected)
In interfaces list->ethernet and tryes to "Power cycle" some interface and returns me and error:
Couldn't perform action - object doesn't exist (4) << these is from drop menu on any int. on button power cycle shows only Inteface: unkown
In System->Routerboard->USB Power Reset again returns me an error : Couldn't perform action - ERROR: USB power reset implemented on this hardware(6) flash drive are recognized and mounted successfully... it si not from hardware!
is it routerboard 751u-2hand; before starting updating before updating the firmware all of these was working!!!
I've tried all possible versions from 6.0 to beta 6.24.rc4, like as i do with firmware i think first as working was 2.9x or 2.7x don't remember.. and last was 3.21!
Soft or hard reset didn't help! i lose 4 days of trying to understand why these is happening..
If anyone can tell me to what verion to rollback i will never update it or buy again RB/MKT!

ooh and yes, the LEDs are dont broken, they flash/blinking before boot, and after boot it up all interface leds is on when have a cable connected, NO blinking(flash); then i plugout the interface cable led is still on for about 2-3 sec and the led turns to off.

[admin@xxxxxxxx] > system leds prinFlags: X - disabled, * - default
# TYPE INTERFACE LEDS
0 * interface-activity wlan1 wlan-led

is it same as you shows us here in wiki :

[admin@MikroTik] /system leds> print
Flags: X - disabled
# TYPE INTERFACE LEDS
0 wireless-signal-strength led1
led2
led3
led4
led5
1 interface-activity ether1 user-led

yeah it was, before!

ffernandes · Fri Dec 26, 2014 2:55 am

using the 6.24 in a ccr and i'm seeing uplink port going up...down...
anyone else?

http://prntscr.com/5kxm6o

cmoegele · Fri Dec 26, 2014 1:06 pm

Friends you should take much more care on your CCR customers! Since 6.20 there is no real stable ROS available,...

camlost · Fri Dec 26, 2014 3:16 pm

Friends you should take much more care on your CCR customers! Since 6.20 there is no real stable ROS available,...

Have no problems with 6.22 on four devices
~500 Mbit/s traffic
~1000 pppoe client

Clauu · Fri Dec 26, 2014 3:18 pm

Logs are saved on a usb stick, at every reboot the stick is recognized either as disk2 or disk1 and never with the same name and as a result logs are often not saved because in system-logging-actions the path must be manually changed, I'm really tired of such bugs

WirelessRudy · Fri Dec 26, 2014 4:12 pm

Is 6.24 already in place? I just updated some routers and found they have been auto updated to 6.24.
Can we have the definite changelog please?

rpingar · Fri Dec 26, 2014 5:43 pm

the new CCR ethernet driver seems working fine because we don't have more drops on VLAN using only-hardware-queue

We had a crash on our PPPOE-Concentrator when huge number of pppoe clients connect/disconnect, sent supout: Tcket#2014122466000113]

We still don't see a good balance of the cores using VPLS / Bridge / PPPOE server:
[Ticket#2014122366000393]

hope to receive soon some answer about these tickets.

regards
Ros

my idea is that VPLS and Simple QUEUE are not yes well optimized about CCR manycores.

Any words from MT about it????

Regards
Ros

Fri Dec 26, 2014 10:31 pm

Yes, v6.24 has been released, please trust the auto-updater. It gets the new version first, before the download page.

ALEJANDROgordon · Fri Dec 26, 2014 10:46 pm

i want to congratulate because this review is more effective using queue tree, but i also have problems in some models of ccr router, the logs shows kerner failure.
i have problems too whit eoip tunnel, i think the synchronization is the problem, because downgrade the router to v6.14 and i don't have any problems whit the tunnel.
sorry for my terrible english.

mars · Fri Dec 26, 2014 11:56 pm

Yes, v6.24 has been released, please trust the auto-updater. It gets the new version first, before the download page.

if 6.24 has been released why don't you put the torrent on the download page
not everybody uses your auto updater

lelmus · Sat Dec 27, 2014 5:05 am

Hi Everyone...

Used the auto-updater today on my CCR (loving it) to get 6.24 (from 6.23) and one strange thing happened, my QOS wasn't throttling the download, but upload throttling was fine.

I checked and it turned out that under query tree the download was using global parent with no-mark, and the no-mark it didn't like anymore. The kernel even panicked and reset while I was playing with its parameters. The upload uses global with a marked name and that was working fine.

So, I fixed the problem by creating a marked name in mangle for the download and replaced no-mark with a marked name, problem solved.

Maybe I was using the no-mark wrong in the first place? Seemed self explanatory in my opinion.

boscolee · Sat Dec 27, 2014 5:22 pm

After upgraded to 6.24. The RB951F-2HnD is switched to Bridge mode in the Quick set page.

Even, I have click it back to Router mode, and apply the configuration. Bridge mode is still keep in the Quick Set.

The IP address from WAN cannot be shown after set the "Automatic" for the Address Acquisition.

Seems, it is a big bug which affect all the RB951F-2HnD in router mode.

Please correct this bug.

bbmax1 · Sun Dec 28, 2014 11:42 am

hello ..
i just want to know the most useful config about the th guard interval and the adapive noice immunity..
by using RB912 , Metal

cdemers · Sun Dec 28, 2014 1:56 pm

Ipv6 local addresses do not seem to be working right on this release. Simply reset configuration, with Ipv6 package enabled, and winbox cannot see the router Ipv6 local address.

Sent from my Nexus 5 using Tapatalk

XoCinek · Sun Dec 28, 2014 4:35 pm

Hi.

Have something change in IP/ROUTE in 6.24?
On several routers with RB750GL we are using RFC3021 /31 addressing. And it works fine till 6.24

Every of that rtr has two WAN vlans, one main and second to other provideer, usually used as backup link (almost never used). When main vlan is gone (ex. cable broke somewhere, switch is gone, etc.) router automatically switch all traffic to second vlan. Using for that distance, and check gateway.

Earlier it works fine but now rtr always thinks that broken link is OK. Only pull out cable set nexthop unreachable.

ip route nexthop print
0 address=10.1.2.2 gw-state=reachable forwarding-nexthop="" interface="" scope=10 check-gateway=icmp gw-check-ok=no
1 address=10.6.152.1 gw-state=reachable forwarding-nexthop="" interface="" scope=10 check-gateway=icmp gw-check-ok=yes

nexthop "0" have gw-check-ok=no (missing answer on ping, correct) but still gw-state=reachable

It works with standar /30 addressing. Understand that it "may" think that nexthop is ok because it really is ... network address, but why he "knows" that gw-check is not ok and gw-state still reachable? Any clue on that?

philippetev · Mon Dec 29, 2014 12:11 am

6.24 on RB951G-2HnD. The NTP client, set to unicast mode, doesn't work (again). Wasn't working in 6.22 and 6.23 too.

Mon Dec 29, 2014 7:24 am

6.24 on RB951G-2HnD. The NTP client, set to unicast mode, doesn't work (again). Wasn't working in 6.22 and 6.23 too.

Do you use the NTP package, or the built-in SNTP client? What is the error message? Make a supout.rif file and email it to support, so we can check what could be wrong.

Mon Dec 29, 2014 10:02 am

system,error,critical System rebooted because of kernel failure 2014-12-24 21:07
system,error,critical router was rebooted without proper shutdown, probably kernel failure 2014-12-24 21:07

He reboot ccr1036 14 minutes intervals. Do not tested this software !!!! Are we cadaver? !!!!! We live in very big trouble ...

((((

did you manage to capture the console output, or make supout.rif file? if yes, please email to support

linux42 · Mon Dec 29, 2014 1:10 pm

ok. email send. continues to reboot periodically.

jeffbrl · Tue Dec 30, 2014 4:17 am

Ipv6 local addresses do not seem to be working right on this release. Simply reset configuration, with Ipv6 package enabled, and winbox cannot see the router Ipv6 local address.

Sent from my Nexus 5 using Tapatalk

cdemers is correct. IPv6 is broken in this release. Link local addresses aren't working. Statically assigned IPv6 addresses are marked with the Invalid flag. I returned to 6.18. I haven't tested versions between 6.18 and 6.24 so I don't know if they are functional with respect to IPv6.

Majklik · Tue Dec 30, 2014 3:13 pm

IPv6 is broken in this release. Link local addresses aren't working. Statically assigned IPv6 addresses are marked with the Invalid flag. I returned to 6.18. I haven't tested versions between 6.18 and 6.24 so I don't know if they are functional with respect to IPv6.

There is a long time problem with the link local IPv6 addresses (I'm seeing it from ROS 5.2x times). Sometimes after a router start is not assigned the link local IPv6 address to a bridge or VRRP interface. When this happen then a global IPv6 address do not works on affected interface too (is marked as invelid). Simple workaround is to disable/enable once or twice affected interface then is the link local address added in correct way and all works fine. This problem do not occur on every start but only sometimes and on the different interfaces....
When this happen then this missing IPv6 address is displayed in the Winbox with the interface "unknown". From command line is output a bit different. For example ROS6.20 (RB1100AH) and ROS6.24 (RB2011) in this case show:

/ipv6 address print where interface="*FFFFFFFF"
Flags: X - disabled, I - invalid, D - dynamic, G - global, L - link-local 
 #    ADDRESS                                     FROM-POOL INTERFACE   ADVERTISE
 0 DL fe80::ff:fe00:6/64                                    *FFFFFFFF   no       
 1 DL fe80::ff:fe00:1/64                                    *FFFFFFFF   no

Workaround:

/interface bridge 
disable 1;enable 1
disable 2;enable 2

A corrected result:

/ipv6 address print where address~"fe80::ff:fe00:./64" 
Flags: X - disabled, I - invalid, D - dynamic, G - global, L - link-local 
 #    ADDRESS                                     FROM-POOL INTERFACE                                      ADVERTISE
 0 DL fe80::ff:fe00:6/64                                    *FFFFFFFF                                      no       
 1 DL fe80::ff:fe00:1/64                                    *FFFFFFFF                                      no       
 2 DL fe80::ff:fe00:6/64                                    bridge-guests                                  no       
 3 DL fe80::ff:fe00:1/64                                    bridge-lan                                     no

The bridges bridge-lan and bridge-guests are definec with the admin MAC addressess 02:00:00:00:00:01 and 02:00:00:00:00:06.

draguzet · Wed Dec 31, 2014 12:11 am

We have BIG Wireless problems on 2.4 band:

1. Channels 2412, 2417, 2422, 2427 are missing on regulatory domain and manual-tx-power, country are Croatia ? WTF ?
2. Cusomers with R52Hn cards have low throughput, very low ( below 1 Mbps), on same location we have Backup AP on 6.22 and throughput are 30-40 Mbps on same place, same channel, same clients ?

We are mailing support with detailed information (suppout, screenshots, etc) and last answer are "use duplex antennas" ??

antsh · Wed Dec 31, 2014 7:34 am

The no-mark parameter is not working for me either. Reboots router as well when trying to change the parameter. Please fix if possible!

Fri Jan 02, 2015 10:22 am

We have BIG Wireless problems on 2.4 band:

1. Channels 2412, 2417, 2422, 2427 are missing on regulatory domain and manual-tx-power, country are Croatia ? WTF ?
2. Cusomers with R52Hn cards have low throughput, very low ( below 1 Mbps), on same location we have Backup AP on 6.22 and throughput are 30-40 Mbps on same place, same channel, same clients ?

We are mailing support with detailed information (suppout, screenshots, etc) and last answer are "use duplex antennas" ??

1. Most likely you have enable the 20/40mhz with eC option and that is why you don't see those channels. Set the channel-width options to Ce and you will see those channels. Look here for more info on the 20/40mhz operations rules:
http://en.wikipedia.org/wiki/IEEE_802.1 ... 4.C2.A0GHz
2. If you downgrade the AP from v6.24 to v6.22 everything works fine?
The configuration for both Main and Backup AP are the same?

jeffbrl · Sat Jan 03, 2015 9:37 pm

Majklik - Thanks for providing a workaround. I guess it must be coincidence that I encountered the defect in 6.24 only.

IPv6 is broken in this release. Link local addresses aren't working. Statically assigned IPv6 addresses are marked with the Invalid flag. I returned to 6.18. I haven't tested versions between 6.18 and 6.24 so I don't know if they are functional with respect to IPv6.
There is a long time problem with the link local IPv6 addresses (I'm seeing it from ROS 5.2x times). Sometimes after a router start is not assigned the link local IPv6 address to a bridge or VRRP interface. When this happen then a global IPv6 address do not works on affected interface too (is marked as invelid). Simple workaround is to disable/enable once or twice affected interface then is the link local address added in correct way and all works fine. This problem do not occur on every start but only sometimes and on the different interfaces....
When this happen then this missing IPv6 address is displayed in the Winbox with the interface "unknown". From command line is output a bit different. For example ROS6.20 (RB1100AH) and ROS6.24 (RB2011) in this case show:
Code: Select all
/ipv6 address print where interface="*FFFFFFFF"
Flags: X - disabled, I - invalid, D - dynamic, G - global, L - link-local 
 #    ADDRESS                                     FROM-POOL INTERFACE   ADVERTISE
 0 DL fe80::ff:fe00:6/64                                    *FFFFFFFF   no       
 1 DL fe80::ff:fe00:1/64                                    *FFFFFFFF   no   
Workaround:
Code: Select all
/interface bridge 
disable 1;enable 1
disable 2;enable 2
A corrected result:
Code: Select all
/ipv6 address print where address~"fe80::ff:fe00:./64" 
Flags: X - disabled, I - invalid, D - dynamic, G - global, L - link-local 
 #    ADDRESS                                     FROM-POOL INTERFACE                                      ADVERTISE
 0 DL fe80::ff:fe00:6/64                                    *FFFFFFFF                                      no       
 1 DL fe80::ff:fe00:1/64                                    *FFFFFFFF                                      no       
 2 DL fe80::ff:fe00:6/64                                    bridge-guests                                  no       
 3 DL fe80::ff:fe00:1/64                                    bridge-lan                                     no
The bridges bridge-lan and bridge-guests are definec with the admin MAC addressess 02:00:00:00:00:01 and 02:00:00:00:00:06.

Who is online